Compare commits

..

129 Commits

Author SHA1 Message Date
Dr-Noob
c718d83868 [v0.96] Add Rocket Lake uarch detection as suggested by #68 2021-04-09 11:26:34 +02:00
Dr-Noob
32b035f1a2 [v0.96] Fix typo as noticed by #70 2021-04-09 10:50:10 +02:00
Dr-Noob
8bb65e0cc0 [v0.96] Fix compilation issue in Windows 2021-04-09 09:25:48 +02:00
Dr-Noob
b699fdc3f2 [v0.96] Update README and fix typo 2021-04-08 11:13:21 +02:00
Dr-Noob
812ee0acc6 [v0.96] Add PREFIX to Makefile and uninstall target, as requested by many users 2021-04-08 09:36:21 +02:00
Dr-Noob
7916e8cbb4 [v0.96] Replace "Simplistic" by "Simple" in description 2021-04-07 20:06:24 +02:00
Dr-Noob
41dbb22a20 [v0.96] Merge branch bugfix3 to include an ARM SoC fix 2021-04-07 19:55:58 +02:00
Dr-Noob
bb9fb17ec8 [v0.96] Tracking issue #54 2021-04-07 16:25:31 +02:00
Dr-Noob
b3ed3e9240 [v0.96] Add 32bit support in Makefile 2021-04-07 14:52:52 +02:00
Dr-Noob
2879876500 [v0.96] Dont treat unknown unified cache as a bug, since there are some processors with eDRAM which supports this level, like #41 2021-04-07 11:17:43 +02:00
Dr-Noob
c7cc8be712 [v0.96] Use lower verbosity for some errors found in cpuid 2021-04-07 10:37:17 +02:00
Dr-Noob
654d2e27e1 [v0.96] Replace Makefile var names (use C names instead of C++) 2021-04-07 10:14:20 +02:00
Dr-Noob
09cbb8874b [v0.96] Fix previous mistake: use level 0xB to check if the level is supported or not, according to Intel docs 2021-04-06 20:37:49 +02:00
Dr-Noob
fe95ca3e10 [v0.96] Fix bug where ebx returned 0 in apic.c when CPU max level >= 0xB but CPU does not support x2apic 2021-04-06 16:56:26 +02:00
Dr-Noob
ec2ad4fef6 [v0.96] Do not consider missing frequency file as a bug 2021-04-06 16:35:12 +02:00
Dr-Noob
d56f7ffd14 [v0.96] Fix segfault when invalid cache size is found 2021-04-06 12:56:44 +02:00
Dr-Noob
4900c10eb3 [v0.96] Remove space between badges in README 2021-03-31 15:33:15 +02:00
Dr-Noob
9f0dc85bd8 [v0.96] Small modifications to README 2021-03-31 15:31:46 +02:00
Dr-Noob
36aeba0e73 [v0.96] Improve README with badges and ToC 2021-03-31 15:15:01 +02:00
Dr-Noob
ca7091bc5e [v0.96] Add short options. Improve --help flag. Update man page 2021-03-31 12:40:19 +02:00
Dr-Noob
8abbd8f69f [v0.96] Fix AMD ASCII art. Add third digit in frequency output 2021-03-31 11:03:54 +02:00
Dr-Noob
2f61ebd35a [0.95] Update README table 2021-01-14 19:30:36 +01:00
Dr-Noob
e21ca95da8 [0.95][x86] Merge bugfix branch, which adds macOS support to cpufetch 2021-01-14 18:46:41 +01:00
Dr-Noob
3bac5cbfe2 [v0.95][x86] Print CPUID 0x1 EAX register with debug flag 2021-01-10 22:03:25 +01:00
Dr-Noob
04f0bfcbde [v0.94][x86] Add uarch detection for Intel families derived from Kaby Lake 2021-01-10 22:01:17 +01:00
Dr-Noob
697a921042 [X86] Avoid checking /sys directory in macOS to find frequency 2021-01-10 09:01:50 +01:00
Dr-Noob
3b624f3025 [X86] Fix bug where unknown hypervisor caused a segfault. This should solve issue #38 2021-01-05 17:37:26 +01:00
Dr-Noob
2494b56a49 [X86] Fix compilation error in MacOS. MacOS does not provide any method to bin threads to cores, so its pretty hard to get apic ids. This first approach is really dark and I hope I can improve it in the future 2021-01-05 17:35:57 +01:00
Dr-Noob
797c708f2d [v0.94][x86] Consider missing frequency file in x86_64 as a bug if no hypervisor is present. Took this idea from issue #37. Add if hypervisor is present to debug mode to prevent more confusions in the future 2020-12-29 00:09:01 +01:00
Dr-Noob
56a1da3428 [v0.94] Do not consider missing frequency file in x86_64 as a bug. Fix typos 2020-12-26 08:52:14 +01:00
Dr-Noob
1ef6daf943 Fix pictures links and add ARM pictures 2020-12-25 18:39:58 +01:00
Dr-Noob
cc20fff6ea Fix grammar and typos 2020-12-25 17:50:10 +01:00
Dr-Noob
c7d1165a94 Update documentation 2020-12-25 15:42:33 +01:00
Dr-Noob
fce0bbf012 [v0.94][ARM] Do not print cache sizes 2020-12-08 18:33:42 +01:00
Dr-Noob
2939d5b352 [v0.94] When SoC string is not matched against any SoC is not considered a bug anymore 2020-12-08 18:04:13 +01:00
Dr-Noob
7e532d57a6 [v0.94] Add reminder in Qualcomm SoCs, that need more time to work properly 2020-12-08 18:01:26 +01:00
Dr-Noob
a96d95eba1 [v0.94] Reviewed MTK SoCs 2020-12-08 15:56:13 +01:00
Dr-Noob
5b9a9e90d0 [v0.94] Fix compilation issues 2020-12-05 11:56:40 +01:00
Dr-Noob
f68c81395b [v0.94] Reviewed snapdragon 6XX, 7XX and 8XX (some SoC models needs more work) 2020-12-05 11:10:15 +01:00
Dr-Noob
4971774ce4 [v0.94] Reviewed snapdragon 2XX and 4XX 2020-12-04 20:23:48 +01:00
Dr-Noob
ea5d07504c [v0.94] Fix snapdragon SoC models 2020-12-04 15:41:33 +01:00
Dr-Noob
c111eb9a41 [v0.94] SoC detection stores the exact SoC model 2020-12-04 10:11:07 +01:00
Dr-Noob
01e22b8090 [v0.94] Fix compilation issues 2020-12-01 16:13:38 +01:00
Dr-Noob
b1f3196e0d [v0.94] Refactor CPU features in a separate struct. Remove x86 debug functions 2020-12-01 12:16:12 +01:00
Dr-Noob
d04d535807 [v0.94][ARM] Add CPU feature detection, such as NEON, AES, SHA... Take into account NEON capabilities for PP computation 2020-12-01 11:40:09 +01:00
Dr-Noob
35c2aa7e6f [v0.94][ARM] Add Kirin and Broadcom ASCII arts. Fix Broadcom SoCs detection 2020-11-28 13:51:20 +01:00
Dr-Noob
fd898331f8 [v0.93][ARM] Tune Exynos ASCII art algorithm 2020-11-28 12:39:25 +01:00
Dr-Noob
532a65e35d [v0.93][ARM] Refactoring in ARM printer 2020-11-28 12:31:08 +01:00
Dr-Noob
1d58b1808d [v0.93][ARM] Support printing ASCII on chips that have lots of attributes (if they have more than 2 different CPUs). Fix all ASCII arts (they were missing the last empty character) 2020-11-28 12:24:13 +01:00
Dr-Noob
7d7af00e68 Add Samsung Exynos ASCII art. Add experimental exynos printer algorithm 2020-11-28 11:09:25 +01:00
Dr-Noob
84fb38a507 [v0.93][ARM] Add Kirin and Broadcom SoC detection. Add function for special SoC strings detection 2020-11-28 10:02:25 +01:00
Dr-Noob
ccfcab88d3 [v0.92][ARM] Add MediaTek ASCII art 2020-11-26 20:33:35 +01:00
Dr-Noob
e5d5e5ef92 [v0.92][Refactoring] Refactor Intel and AMD printer code into algorithms 2020-11-26 18:45:33 +01:00
Dr-Noob
ff92107d7c [v0.92] Fix previous commit (I forgot to add the file) 2020-11-26 18:35:29 +01:00
Dr-Noob
b9e96baf91 [v0.92] Add printer algorithms. This approach allows more flexibility in the printer. Improve Snapdragon ASCII 2020-11-26 18:30:39 +01:00
Dr-Noob
c62a63f539 [v0.92][ARM] Print ACII art based on SoC instead of CPU vendor. Add snapdragon ASCII art. Refactor printer to save some lines of code 2020-11-26 16:42:36 +01:00
Dr-Noob
89e5e30e53 [v0.91][X86][BUGFIX] Fix annoying compilation issue 2020-11-26 13:05:14 +01:00
Dr-Noob
7c2463eb8f [v0.91][ARM] New experimental way of matching SoC names 2020-11-26 12:41:41 +01:00
Dr-Noob
b4b693a11e [v0.91][ARM][BUGFIX] Do not print error message if frequency file does not exist in ARM. Fix arm udev file (I uploaded my debug udev by mistake) 2020-11-26 12:15:07 +01:00
Dr-Noob
ff032efb28 [v0.91][ARM] Fix for snapdragon chips reporting name in lowercase 2020-11-26 11:51:06 +01:00
Dr-Noob
da94c7ee18 [v0.91][ARM] Add Exynos SoCs 2020-11-26 11:30:14 +01:00
Dr-Noob
4c36e4c5e5 [v0.91][ARM] Print manufacturing process from SoC information 2020-11-26 10:52:44 +01:00
Dr-Noob
20dbc3be27 [v0.90][ARM] Print bug message if SoC string is found but not detected. This is maybe too verbose, but I would like to increase the support for more ARM SoC, and I hope this message is useful for this purpose 2020-11-26 10:43:42 +01:00
Dr-Noob
53b4ff5793 [v0.90][ARM] Add MediaTek SoCs 2020-11-26 09:53:53 +01:00
Dr-Noob
4f1dd82bba [v0.90][ARM] Add most of Qualcomm SoCs 2020-11-25 11:34:49 +01:00
Dr-Noob
37e849978a [v0.90][ARM] Print right number of cores and frequency for each CPU 2020-11-24 16:17:53 +01:00
Dr-Noob
233565c052 [v0.89][ARM][BUGFIX] Fix uarch detection for Kryo 260 / 280 2020-11-24 16:16:22 +01:00
Dr-Noob
5d168f5707 [v0.89][ARM][BUGFIX] Fix obscure bug where cpufetch failed if cpuinfo did not start with "processor" string 2020-11-24 15:20:07 +01:00
Dr-Noob
0bc978564e [v0.89][ARM][BUGFIX] Two CPUs are equal when Main ID Register AND frequency are equal, not just when Main ID register are equal 2020-11-24 13:21:27 +01:00
Dr-Noob
fbea497740 [v0.89] Change freq from int64 to int32, which fixes a compilation issue. Fix Makefile in Windows 2020-11-24 12:52:42 +01:00
Dr-Noob
8645b54b58 [v0.89][ARM] Separate udev for ARM in a different file, since there are many functions that are ARM only. Refactoring of file reading code 2020-11-24 12:32:38 +01:00
Dr-Noob
220dd02abd [v0.89][ARM] Add very basic SoC parsing from cpuinfo/android strings. Only detects two SoCs, but allows debugging 2020-11-24 11:44:24 +01:00
Dr-Noob
685e78f2b7 [v0.88][ARM][BUGFIX] Fix bugs in single CPU SoCs and small bugs related to strings 2020-11-24 10:27:33 +01:00
Dr-Noob
71d660d7b1 [v0.88][ARM] Fetch raw soc name from cpuinfo or android (if supported) 2020-11-24 10:03:21 +01:00
Dr-Noob
bb05a4d577 [v0.88][ARM][BUGFIX] Fix some compilation issues 2020-11-23 18:45:42 +01:00
Dr-Noob
eaa86522a4 [v0.88][ARM] Add very basic SoC detection (Android only) 2020-11-23 18:30:00 +01:00
Dr-Noob
8927048c95 [v0.88] Reorganize files and documentation. Remove ascii directory (it was useless) 2020-11-23 16:15:46 +01:00
Dr-Noob
093222d533 [v0.88][DOC] Update documentation and divide it in two files, one for each architecture 2020-11-23 16:11:54 +01:00
Dr-Noob
716706d0a7 [v0.88][ARM][BUGFIX] Fetch number of cores from /sys/devices/system/cpu/present file, instead of /proc/cpuinfo. Pay attention to cases where frequency and/or MIDR can not be fetched from cpuinfo. This happens when the CPU has offline cores 2020-11-22 15:18:15 +01:00
Dr-Noob
fcb2c716db [v0.87][FREQ] Frequency in udev is now fetched as a per core basis. Before this commit, freq was always fetched from core 0. This allows ARM do detect the max frequency of each of the cores (which may or may not be the same in all of them) 2020-11-22 10:23:02 +01:00
Dr-Noob
0875c4d425 [v0.87][ARM] cpuInfo now holds all the structs (freq, cache, etc), instead of having them separated. This allows ARM to represent a single CPU, because from its pointer, it is able to access the specific frequency, cache, etc 2020-11-22 09:57:50 +01:00
Dr-Noob
f5ec566577 [v0.87][DOC] Explain commit #42ade63 in documentation. Add Android information in README 2020-11-22 09:20:35 +01:00
Dr-Noob
5b0cbd622f [v0.87][BUGFIX] Add checks to detect wrong ASCII arts. Fix ARM ASCII art 2020-11-21 19:21:19 +01:00
Dr-Noob
0b9f0e860c [v0.87][PRINTER] ASCII art attributes are no longer fixed in a given position, and are not fixed in length (max length is now 100). Added different algorithms for printing ASCII art for X86 and ARM, which allows ARM to show each CPU inside a SoC 2020-11-21 19:04:57 +01:00
Dr-Noob
50931ee94d [v0.86][BUGFIX] Fix print format for hex values 2020-11-21 16:45:51 +01:00
Dr-Noob
42ade63746 [v0.86][BUGFIX] Add old AMD CPUs cache fix to master branch 2020-11-21 16:41:08 +01:00
Dr-Noob
e4a4e13d56 [v0.82][BUGFIX] Using 0x80000006 in new AMD CPUs outputs wrong L3 size since it reports the full size instead the size of a single L3. Use old method just when is necessary 2020-11-21 16:37:51 +01:00
Dr-Noob
7d707916fb [v0.86][OPTIONS] Replace levels option with debug option, which does the same on x86, but also exists on ARM, which prints MIDR registers (need work to be properly implemented) 2020-11-18 23:41:42 +01:00
Dr-Noob
c44a646cd1 [v0.85][ARM] Add SoC field in ARM and remove CPU Name field, which is only valid in x86. Fix Makefile for some strict compilers 2020-11-18 23:22:26 +01:00
Dr-Noob
8c11cb2422 [v0.84][ARM] Add ISA field in ARM. ISA depends on uarch, not on specific CPU. Fill all the missing data in uarch.c 2020-11-14 11:11:32 +01:00
Dr-Noob
cb78f18de1 [v0.84][BUGFIX] Fix more compilation issues 2020-11-10 22:46:39 +01:00
Dr-Noob
07f3f26ff6 [v0.84][BUGFIX] Fix more compilation issues 2020-11-10 22:45:14 +01:00
Dr-Noob
27aabb35be [v0.84][BUGFIX] Fix compilation issues 2020-11-10 18:51:13 +01:00
Dr-Noob
904cb46765 [v0.84][ARM] Add lots of new microarch detection 2020-11-10 18:50:32 +01:00
Dr-Noob
3aa13269b7 [v0.83][ARM] Add basic support for microarchitecture and CPU name detection. Need to add the remaining models 2020-11-08 16:49:01 +01:00
Dr-Noob
b978ddc83d [v0.82][BUGFIX] Issue #33: Use 0x80000006 for cache fetching in AMD, instead of 0x8000001D. This means that a different approach in Intel and AMD CPUs 2020-11-07 10:48:48 +01:00
Dr-Noob
16abfa7022 [v0.82][README] Update README to show ARM experimental support 2020-11-06 11:10:30 +01:00
Dr-Noob
9c8e169592 [v0.82][ARM][ASCII][Refactoring] ARM ascii changes. Remove the assumption that all sockets are equal in a ARM based SoC. Little more support for ARM processors. Add ARM color style 2020-11-06 10:06:13 +01:00
Dr-Noob
4f1722ead6 [v0.81][ARM][Refactoring] Refactoring and very basic ARM support 2020-11-05 13:44:46 +01:00
Dr-Noob
f4f68287aa [v0.8][Refactoring] Refactoring ARM code and source code tree 2020-11-05 11:01:46 +01:00
Dr-Noob
1fad4fd10b [v0.8][ARM] Building support in ARM 2020-11-05 09:28:41 +01:00
Dr-Noob
5cc9038f3d Fix peak performance in KNL 2020-10-20 21:13:04 +02:00
Dr-Noob
f992d0122f Merge remote-tracking branch 'Wunkolo/feat/windows-colors' into master 2020-10-20 20:49:29 +02:00
Dr-Noob
ac86be2d7a Fix bug in Windows where specifying a style while using a terminal that supports color does not enable the color support, so colors do not show correctly 2020-10-20 20:43:14 +02:00
Wunkolo
c158cab005 Update readme with new Windows terminal support 2020-10-17 18:46:53 -07:00
Wunkolo
9867754d08 Implement VT100 escape-code detection for Windows
Latest versions of windows have support for the parsing VT100 escape
code sequences, allowing for terminal colors similar to Linux.

https://docs.microsoft.com/en-us/windows/console/console-virtual-terminal-sequences#screen-colors

Here I have it get the console mode, set the
`ENABLE_VIRTUAL_TERMINAL_PROCESSING` flag, and then grab the console
mode again to both verify that VT100 escape sequences are supported and
that it is enabled after setting it to determine if the printer should
allow for fancy-color mode.
2020-10-17 18:46:20 -07:00
Dr-Noob
5119ece0dd Refactoring defines to enums 2020-10-14 10:55:46 +02:00
Dr-Noob
e37c7d9ae0 Basic support for virtual machines 2020-10-11 23:27:19 +02:00
Dr-Noob
aa5f0a8b88 Add install option in Makefile 2020-09-21 13:07:18 +02:00
Dr-Noob
075e4399f8 Update description 2020-09-05 10:51:16 +02:00
Dr-Noob
3dedb0bbc3 Add programming documentation 2020-09-05 09:46:37 +02:00
Dr-Noob
60bc02185d Small fix 2020-09-01 20:48:37 +02:00
Dr-Noob
ae752bac77 Add images 2020-09-01 20:45:52 +02:00
Dr-Noob
500ccfa871 Stable version 0.7 heavily tested in many different CPUs 2020-09-01 20:44:48 +02:00
Dr-Noob
877833db0a Dont fetch if smt is enabled if its not supported (AMD). Dont guess cache topology, fetch it from CPUID (AMD) 2020-09-01 13:08:44 +02:00
Dr-Noob
5cca6df218 Fix memory leaks. Add debug message when microarch is unknown 2020-09-01 11:32:08 +02:00
Dr-Noob
de8952b4ea Fix bug which caused you couldnt use --version. Change --style to be more user friendly. Update --help 2020-09-01 11:00:11 +02:00
Dr-Noob
1f80566f63 New info to be displayed (uarch and process) instead of other info (sha, aes, sse) 2020-09-01 09:37:53 +02:00
Dr-Noob
ab1416563c Fix PP in Ice Lake 2020-08-31 18:27:32 +02:00
Dr-Noob
1a9c0546f2 Add support for detecting AMD microarch 2020-08-31 15:56:21 +02:00
Dr-Noob
35efdd8f2c Fix #26. Guess number of VPUs according to microarchitecture 2020-08-31 14:04:41 +02:00
Dr-Noob
5148962fa3 Add code to detect CPU microarchitecture (Intel only, at the moment) 2020-08-31 13:18:25 +02:00
Dr-Noob
d998acdcdf Fix #25: Compute PP taking into account the number of sockets 2020-08-31 09:33:39 +02:00
Dr-Noob
81a45628f0 Code refactoring. Forgot to add verbose option to help 2020-08-30 13:55:37 +02:00
Dr-Noob
4f98a5bccf Refactor previous commit 2020-08-30 12:42:38 +02:00
Dr-Noob
dae0f678ad Fix #23. I tried fetching the cache topology in AMD but could not find a proper way, so the code fallback to two commits ago. cpufetch has to guess cache sizes except L3, which can be fetched. Since I have been trying many approaches and stuff, the code needs to be refactored 2020-08-30 12:12:25 +02:00
Dr-Noob
69cc08759a Fix #21 and #22: Obtain the number of caches of every level instead of guessing them. It is done by fetching cache topology from apic. It works, but it needs a big refactoring. Moreover, it currently works only on Intel CPUs, so this breaks the cache in AMD. 2020-08-29 21:51:14 +02:00
Dr-Noob
d8dad29a57 Fix SMT bug in AMD. I would like to improve it, since Intel can use APIC with 0x1 and 0xB (extended) while AMD does with 0x1 and extended seems to be 0x1E. Add support to detect more than one L3 cache. This is not a very elegant solution, since we still assume that we have the same number of caches as caches in a given level. To fix it, cpufetch should know how many caches are in a given level (hint, Linux knows using shared_cpu_map) 2020-08-29 15:42:56 +02:00
Dr-Noob
e08b60b1c8 Project stopped until I have time to continue 2020-07-12 19:08:38 +02:00
55 changed files with 5790 additions and 2156 deletions

View File

@@ -1,34 +1,63 @@
CXX=gcc
CC=gcc
CXXFLAGS=-Wall -Wextra -Werror -pedantic -fstack-protector-all -pedantic -std=c99
SANITY_FLAGS=-Wfloat-equal -Wshadow -Wpointer-arith -Wstrict-overflow=5 -Wformat=2
CFLAGS=-Wall -Wextra -Werror -pedantic -fstack-protector-all -pedantic -std=c99
SANITY_FLAGS=-Wfloat-equal -Wshadow -Wpointer-arith
SRC_DIR=src/
SOURCE=$(SRC_DIR)main.c $(SRC_DIR)cpuid.c $(SRC_DIR)apic.c $(SRC_DIR)cpuid_asm.c $(SRC_DIR)printer.c $(SRC_DIR)args.c $(SRC_DIR)global.c
HEADERS=$(SRC_DIR)cpuid.h $(SRC_DIR)apic.h $(SRC_DIR)cpuid_asm.h $(SRC_DIR)printer.h $(SRC_DIR)ascii.h $(SRC_DIR)args.h $(SRC_DIR)global.h
PREFIX ?= /usr
SRC_COMMON=src/common/
COMMON_SRC = $(SRC_COMMON)main.c $(SRC_COMMON)cpu.c $(SRC_COMMON)udev.c $(SRC_COMMON)printer.c $(SRC_COMMON)args.c $(SRC_COMMON)global.c
COMMON_HDR = $(SRC_COMMON)ascii.h $(SRC_COMMON)cpu.h $(SRC_COMMON)udev.h $(SRC_COMMON)printer.h $(SRC_COMMON)args.h $(SRC_COMMON)global.h
ifneq ($(OS),Windows_NT)
SOURCE += $(SRC_DIR)udev.c
HEADERS += $(SRC_DIR)udev.h
OUTPUT=cpufetch
arch := $(shell uname -m)
ifeq ($(arch), $(filter $(arch), x86_64 i686))
SRC_DIR=src/x86/
SOURCE += $(COMMON_SRC) $(SRC_DIR)cpuid.c $(SRC_DIR)apic.c $(SRC_DIR)cpuid_asm.c $(SRC_DIR)uarch.c
HEADERS += $(COMMON_HDR) $(SRC_DIR)cpuid.h $(SRC_DIR)apic.h $(SRC_DIR)cpuid_asm.h $(SRC_DIR)uarch.h
CFLAGS += -DARCH_X86
else
SRC_DIR=src/arm/
SOURCE += $(COMMON_SRC) $(SRC_DIR)midr.c $(SRC_DIR)uarch.c $(SRC_DIR)soc.c $(SRC_DIR)udev.c
HEADERS += $(COMMON_HDR) $(SRC_DIR)midr.h $(SRC_DIR)uarch.h $(SRC_DIR)soc.h $(SRC_DIR)udev.c $(SRC_DIR)socs.h
CFLAGS += -DARCH_ARM -Wno-unused-parameter
endif
OUTPUT=cpufetch
else
# Assume x86_64
SRC_DIR=src/x86/
SOURCE += $(COMMON_SRC) $(SRC_DIR)cpuid.c $(SRC_DIR)apic.c $(SRC_DIR)cpuid_asm.c $(SRC_DIR)uarch.c
HEADERS += $(COMMON_HDR) $(SRC_DIR)cpuid.h $(SRC_DIR)apic.h $(SRC_DIR)cpuid_asm.h $(SRC_DIR)uarch.h
CFLAGS += -DARCH_X86
SANITY_FLAGS += -Wno-pedantic-ms-format
OUTPUT=cpufetch.exe
endif
all: $(OUTPUT)
debug: CXXFLAGS += -g -O0
debug: $(OUTPUT)
debug: CFLAGS += -g -O0
debug: $(OUTPUT)
release: CXXFLAGS += -static -O3
release: CFLAGS += -static -O3
release: $(OUTPUT)
$(OUTPUT): Makefile $(SOURCE) $(HEADERS)
$(CXX) $(CXXFLAGS) $(SANITY_FLAGS) $(SOURCE) -o $(OUTPUT)
$(CC) $(CFLAGS) $(SANITY_FLAGS) $(SOURCE) -o $(OUTPUT)
run:
run: $(OUTPUT)
./$(OUTPUT)
clean:
@rm $(OUTPUT)
install: $(OUTPUT)
install -Dm755 "cpufetch" "$(PREFIX)/bin/cpufetch"
install -Dm644 "LICENSE" "$(PREFIX)/share/licenses/cpufetch-git/LICENSE"
install -Dm644 "cpufetch.8" "$(PREFIX)/share/man/man8/cpufetch.8.gz"
uninstall:
rm -f "$(PREFIX)/bin/cpufetch"
rm -f "$(PREFIX)/share/licenses/cpufetch-git/LICENSE"
rm -f "$(PREFIX)/share/man/man8/cpufetch.8.gz"

124
README.md
View File

@@ -1,12 +1,54 @@
# cpufetch
<p align="center"><img width=50% src="./pictures/cpufetch.png"></p>
Prints a fancy summary of the CPU with some advanced information
<div align="center">
### Platforms
This tool works on both 64 only and under Linux because of its [implementation details](#implementation). AMD support is not guaranteed so information may not be correct
![GitHub tag (latest by date)](https://img.shields.io/github/v/tag/Dr-Noob/cpufetch?label=cpufetch)
![GitHub Repo stars](https://img.shields.io/github/stars/Dr-Noob/cpufetch?color=4CC61F)
![GitHub issues](https://img.shields.io/github/issues/Dr-Noob/cpufetch)
![Contributions welcome](https://img.shields.io/badge/contributions-welcome-orange.svg)
[![License](https://img.shields.io/badge/license-MIT-blue.svg)](https://opensource.org/licenses/MIT)
### Usage and installation
<h4 align="center">Simple yet fancy CPU architecture fetching tool</h4>
&nbsp;
![cpu1](pictures/i9.png)
</div>
# Table of contents
<!-- UPDATE with: doctoc --notitle README.md -->
<!-- START doctoc generated TOC please keep comment here to allow auto update -->
<!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE -->
- [1. Support](#1-support)
- [2. Installation](#2-installation)
- [2.1 Building from source](#21-building-from-source)
- [2.2 Linux](#22-linux)
- [2.2 Windows](#22-windows)
- [2.3 macOS](#23-macos)
- [2.4 Android](#24-android)
- [3. Examples](#3-examples)
- [3.1 x86_64 CPUs](#31-x86_64-cpus)
- [3.2 ARM CPUs](#32-arm-cpus)
- [4. Colors and style](#4-colors-and-style)
- [5. Implementation](#5-implementation)
- [6. Bugs or improvements](#6-bugs-or-improvements)
<!-- END doctoc generated TOC please keep comment here to allow auto update -->
# 1. Support
cpufetch supports x86, x86_64 (Intel and AMD) and ARM.
| Platform | x86_64 | ARM | Notes |
|:---------:|:------------------------:|:-------------------:|:-----------------:|
| GNU/Linux | :heavy_check_mark: | :heavy_check_mark: | Best support |
| Windows | :heavy_check_mark: | :x: | Some information may be missing. <br> Colors will be used if supported |
| Android | :heavy_exclamation_mark: | :heavy_check_mark: | Some information may be missing. <br> Not tested under x86_64 |
| macOS | :heavy_check_mark: | :x: | Some information may be missing. <br> Apple M1 support may be added <br> in the future (see [#47](https://github.com/Dr-Noob/cpufetch/issues/47))|
# 2. Installation
## 2.1 Building from source
Just clone the repo and use `make` to compile it
```
@@ -16,38 +58,58 @@ make
./cpufetch
```
### Example
The Makefile is designed to work on Linux, Windows and macOS.
This is the output of `cpufetch` in a i7-4790K
## 2.2 Linux
There is a cpufetch package available in Arch Linux ([cpufetch-git](https://aur.archlinux.org/packages/cpufetch-git)). If you are in another distribution, you can build `cpufetch` from source.
![Example](/preview.png)
## 2.2 Windows
In the [releases](https://github.com/Dr-Noob/cpufetch/releases) section you will find some cpufetch executables compiled for Windows. Just download and run it from Windows CMD. You can also build `cpufetch` from source.
### Output
## 2.3 macOS
You need to build `cpufetch` from source.
Output is detailed as follows:
## 2.4 Android
1. Install `termux` app (terminal emulator)
2. Run `pkg install -y git make clang` inside termux.
3. Build from source normally:
- git clone https://github.com/Dr-Noob/cpufetch
- cd cpufetch
- make
- ./cpufetch
| Field | Description | Possible Values |
|:----------:|:-----------------------:|:-----------------:|
| Name | Name of the CPU | Any valid CPU name |
| Frequency | Max frequency of the CPU(in GHz) | X.XX(GHz or MHz)
| N.Cores | Number of cores the CPU has. If CPU supports `Hyperthreading` or similar, this will show cores and threads separately | X(cores)X(threads)
| AVX | Type of AVX supported by the CPU or None. AVX instructions allows the CPU to vectorize the code with a witdh of 256 bits in single precision(or 512bits if AVX512 is supported) | AVX,AVX2,AVX512,None
| SSE | Same as AVX, but SSE family are 128bits witdh | SSE, SSE2, SSE3, SSSE3, SSE4a, SSE4_1, SSE4_2,None |
| FMA | Does this CPU support FMA(Fused Multiply Add)?This instruction allows the CPU to multiply and add a value on the same clock cycle | FMA3,FMA4,None |
| AES | Does this CPU support AES? This instruction is allows the CPU to make AES cypher efficiently | Yes or No |
| SHA | Does this CPU support SHA? This instruction is allows the CPU to make SHA hashing efficiently | Yes or No |
| L1 Size | Size(in bytes) of the L1 cache, separated in data and instructions | XXB(Data)XXB(instructions) |
| L2 Size | Size(in bytes) of the L2 cache(both are unified) | XXXKB or None |
| L3 Size | Same as L3 | XXXXKB or None |
| Peak FLOPS | Max FLOPS(Floating Point Operation Per Second) this CPU could theoretical achieve. This is calculated by: `N.Cores*Freq*2(Because 2 functional units)*2(If has FMA)*VectorWidth` | XXX.XX (G/T)FLOPs |
# 3. Examples
Here are more examples of how `cpufetch` looks on different CPUs.
`cpufetch` also prints a simple ascii art of the manufacturer logo.
## 3.1 x86_64 CPUs
### Implementation
![cpu2](pictures/epyc.png)
`cpufetch` makes use of two techniques to fetch data:
* __cpuid__: CPU name, number of threads per core and instructions features are fetched via _cpuid_. See [this](http://www.sandpile.org/x86/cpuid.htm) and [Intel Processor Identification and the CPUID Instruction](https://www.scss.tcd.ie/~jones/CS4021/processor-identification-cpuid-instruction-note.pdf) for more information.
* __udev__: Cache and frequency are fetched via _udev_, by looking at specific files from `/sys/devices/system/cpu`
![cpu3](pictures/cascade_lake.png)
### Bugs or improvements
Feel free to open a issue on the repo to report a issue or propose any improvement in the tool
## 3.2 ARM CPUs
![cpu4](pictures/exynos.png)
![cpu5](pictures/snapdragon.png)
# 4. Colors and style
By default, `cpufetch` will print the CPU art with the system colorscheme. However, you can always set a custom color scheme, either
specifying Intel or AMD, or specifying the colors in RGB format:
```
./cpufetch --color intel (default color for Intel)
./cpufetch --color amd (default color for AMD)
./cpufetch --color 239,90,45:210,200,200:100,200,45:0,200,200 (example)
```
In the case of setting the colors using RGB, 4 colors must be given in with the format: ``[R,G,B:R,G,B:R,G,B:R,G,B]``. These colors correspond to CPU art color (2 colors) and for the text colors (following 2). Thus, you can customize all the colors.
# 5. Implementation
See [cpufetch programming documentation](https://github.com/Dr-Noob/cpufetch/blob/master/doc/README.md).
# 6. Bugs or improvements
There are many open issues in github (see [issues](https://github.com/Dr-Noob/cpufetch/issues)). Feel free to open a new one report an issue or propose any improvement in `cpufetch`
I would like to thank [Gonzalocl](https://github.com/Gonzalocl) and [OdnetninI](https://github.com/OdnetninI) for their help, running `cpufetch` in many different CPUs they have access to, which makes it easier to debug and check the correctness of `cpufetch`.

View File

@@ -1,21 +0,0 @@
@@@@ @@@% @@@. @@@@@@@@( .############
@@@@@@ @@@@@ %@@@@. @@@ @@@@ .((((((####
@@@ @@@ @@@/@@@@@@&@@. @@@ %@@% # ####
@@@, (@@# @@@ @@@/ @@@. @@@ .@@@ ### ####
,@@@@@@@@@@, @@@ @@@. @@@ @@@ #### ####
@@@ @@@ @@@ @@@. @@@@@@@@@& #######/ .##

View File

@@ -1,19 +0,0 @@
################
####### #######
#### ####
### ####
### ###
### ### ###
# ### ### ###
## ### ######### ###### ###### ### ###
## ### ### ### ### #### #### ### ###
## ### ### ### ### ### ### ### ###
## ### ### ### ### ########## ### ####
## ### ### ### ### ### ### #####
## ## ### ### ##### ######### ## ###
###
(###
#### ####
##### ##########
########## ################
###############################

View File

@@ -1,33 +1,64 @@
.TH man 8 "22 Jun 2018" "0.32" "cpufetch man page"
.\" DO NOT MODIFY THIS FILE! It was generated by help2man 1.48.2.
.TH CPUFETCH "1" "March 2021" "cpufetch v0.96 (x86_64 build)" "User Commands"
.SH NAME
cpufetch \- Prints a fancy summary of the CPU with some advanced information
cpufetch \- manual page for cpufetch v0.96 (x86_64 build)
.SH SYNOPSIS
cpufetch [--help] [--style STYLE]
.B cpufetch
[\fI\,OPTION\/\fR]...
.SH DESCRIPTION
cpufetch will print CPU information, for which will query cpuid instructions and udev directories on Linux. It should display:
.IP \[bu] 2
Name
.IP \[bu]
Frequency
.IP \[bu]
Number of cores(Physical and Logical)
.IP \[bu]
AVX,SSE,FMA,AES and SHA support
.IP \[bu]
L1,L2 and L3 size
.IP \[bu]
Theoretical peak flops
Simple yet fancy CPU architecture fetching tool
.SH OPTIONS
.TP
\fB\-\-help\fR
Prints help
\fB\-c\fR, \fB\-\-color\fR
Set the color scheme (by default, cpufetch uses the system color scheme)
.TP
\fB\-\-version\fR
Prints cpufetch version
\fB\-s\fR, \fB\-\-style\fR
Set the style of CPU art
.TP
\fB\-\-style\fR
Specify the color style of ascii logo
.SH BUGS
No known bugs. AMD CPUs may not be fully supported
.SH AUTHOR
\fB\-d\fR, \fB\-\-debug\fR
Prints CPU model and cpuid levels (debug purposes)
.TP
\fB\-v\fR, \fB\-\-verbose\fR
Prints extra information (if available) about how cpufetch tried fetching information
.TP
\fB\-h\fR, \fB\-\-help\fR
Prints this help and exit
.TP
\fB\-V\fR, \fB\-\-version\fR
Prints cpufetch version and exit
.SH "COLORS:"
.TP
* "intel":
Use Intel default color scheme
.TP
* "amd":
Use AMD default color scheme
.TP
* "arm":
Use ARM default color scheme
.TP
* custom:
If color argument do not match "intel", "amd" or "arm", a custom scheme can be specified.
4 colors must be given in RGB with the format: R,G,B:R,G,B:...The first 2 colors are the CPU art color and the next 2 colors are the text colors
.SH "STYLES:"
.TP
* "fancy":
Default style
.TP
* "retro":
Old cpufetch style
.TP
* "legacy":
Fallback style for terminals that do not support colors
.SH "EXAMPLES:"
Run cpufetch with Intel color scheme:
\&./cpufetch \fB\-\-color\fR intel
Run cpufetch with a custom color scheme:
\&./cpufetch \fB\-\-color\fR 239,90,45:210,200,200:100,200,45:0,200,200
.SH "BUGS:"
Report bugs to https://github.com/Dr\-Noob/cpufetch/issues
.SH "NOTE:"
Peak performance information is NOT accurate. cpufetch computes peak performance using the max frequency. However, to properly compute peak performance, you need to know the frequency of the CPU running AVX code, which is not be fetched by cpufetch since it depends on each specific CPU. For peak performance measurement see: https://github.com/Dr\-Noob/peakperf
.SH "AUTHOR:"
Dr-Noob (https://github.com/Dr-Noob)

60
doc/DOCUMENTATION_ARM.md Normal file
View File

@@ -0,0 +1,60 @@
### 2. Why ARM cpufetch works on Linux based systems?
CPUID instructions (present in x86 architectures) [[1](#references)] allow user level applications to obtain information about the CPU. In ARM architectures, there are many registers (MIDR [[2](#references)], CCSIDR [[3](#references)]) that provide information about the CPU too. However, those registers can only be read from privilege mode (PL1 or higher). Therefore, any user level tool which can actually read information about the running CPU must use the operating system to do so. cpufetch uses some Linux kernel features (see the remaining sections). Therefore, cpufetch in ARM processors is limited to Linux kernel based systems, such as __GNU/Linux__ and __Android__
### 3. How to get CPU microarchitecture?
__Involved code: [get_midr_from_cpuinfo (udev.c)](https://github.com/Dr-Noob/cpufetch/blob/master/src/arm/udev.c), [midr.c](https://github.com/Dr-Noob/cpufetch/blob/master/src/arm/midr.c)__
Microarchitecture information is acquired from the Main ID Register (MIDR) [[2](#references)]. Currently, cpufetch rebuilds this register using `/proc/cpuinfo` file. While this file does not contain the value of the register per se, it contains the following fields:
- `CPU implementer`,
- `CPU architecture`
- `CPU variant`
- `CPU part`
- `CPU revision`
The MIDR register can be built with this information. Another posible approach is to read MIDR directly from `/sys/devices/system/cpu/cpu*/regs/identification/midr_el1`
With the MIDR available, the approach is the same as the one used in x86_64 architectures. cpufetch has a file that acts like a database that tries to match the MIDR register with the specific CPU microarchitecture.
### 4. How to get CPU topology?
__Involved code: [get_ncores_from_cpuinfo (udev.c)](https://github.com/Dr-Noob/cpufetch/blob/master/src/arm/udev.c), [midr.c](https://github.com/Dr-Noob/cpufetch/blob/master/src/arm/midr.c)__
ARM provides a new interesting architecture feature: big.LITTLE architectures [[4](#references)]. An ARM CPU can be organized like a typical x86_64 CPU, where all cores share the same microarchitecture. However, ARM big.LITTLE architecture breaks this schema. In a big.LITTLE CPU, two or more CPUs microarchitecture live in the same chip.
This means that cpufetch can't just read which microarchitecture is the first core and assume that the rest of them shares the same microarchitecture. To get the CPU topology, cpufetch first reads the number of CPU cores. This can be obtained from `/sys/devices/system/cpu/present`
Then, for each core, cpufetch reads the MIDR and also the frequency (see section 5). Then, cpufetch assumes that two cores are different when their MIDR are different. This idea allows cpufetch to detect big.LITTLE architectures, and to know how many cores of each architecture the running CPU has.
### 5. How to get the frequency?
Frequency is read directly from `/sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_max_freq`
### 6. How to get system on chip model?
__Involved code: [soc.c](https://github.com/Dr-Noob/cpufetch/blob/master/src/arm/soc.c)__
System on chip (SoC) model is obtained using the same idea as the microarchitecture. First, SoC string is read. Then, the string has to be matched against a database-like function (__parse_soc_from_string__). The SoC string of the running CPU can be obtained using two different approaches:
- Using `/proc/cpuinfo`. This is the first thing to try. Linux kernel usually provides the string under the `Hardware` keyword. However, the Linux kernel may be unable to provide this information, or this string may not be found in the database-like function.
- Using Android properties: This only works on Android systems. Android properties can be read using `__system_property_get` function. cpufetch tries to read two properties:
- `ro.mediatek.platform`
- `ro.product.board`
If any string is returned, cpufetch tries to find a match in the database (using the same database as in the case of `/proc/cpuinfo`).
The expected strings have to be hardcoded. I found two ways of knowing which string should correspond to which SoC:
- Searching on the internet. Manufacturers __usually__ provide this information. For example, Qualcomm usually publishes the chip name along with other characteristics (under the `Part` or `Part number` keyword [[6](#references)]).
- "Hunting" for the strings. For example, finding smartphones with a given SoC and manually reading the `/proc/cpuinfo` or the `build.prop` file. A very good resource to do this is the SpecDevice webpage [[7](#references)]).
### 7. How to get cache size and topology?
ARM architecture supports reading the cache information via some registers (for example, the CCSIDR register [[3](#references)]). As mentioned earlier, user level applications are not able to read these registers directly. The remaining option is to ask the operating system for this information. However, at the moment, the __Linux kernel does not provide cache information__. Therefore, cpufetch does not print any cache information on ARM CPUs at the moment. There are, however, other approaches to be explored:
- Read the registers in kernel mode. This can be accomplished by running a kernel module [[4](#references)]. Unfortunately, running a custom kernel module is tricky, and sometimes impossible to do reasonably (for example, in Android devices). In any case, my decision is to run cpufetch on user level only.
- Hardcode the cache information for each SoC: Sometimes, manufacturers publish technical information about the chips, where cache topology and size are shown. This method is impractical, since this kind of information is very hard (or impossible) to find online, and the number of SoC is huge.
#### References
- [1] [cpufetch x86_64 documentation](https://github.com/Dr-Noob/cpufetch/blob/master/doc/DOCUMENTATION_X86.md)
- [2] [Main ID Register](https://developer.arm.com/documentation/ddi0433/c/system-control/register-descriptions/main-id-register)
- [3] [Cache size ID Register](https://developer.arm.com/documentation/100403/0200/register-descriptions/aarch32-system-registers/ccsidr--cache-size-id-register)
- [4] [How to get the size of the CPU cache in Linux](https://stackoverflow.com/a/63474811/9975463)
- [5] [ARM big.LITTLE](https://en.wikipedia.org/wiki/ARM_big.LITTLE)
- [6] [Snapdragon 855+ Mobile Platform](https://www.qualcomm.com/products/snapdragon-855-plus-mobile-platform)
- [7] [SpecDevice](http://specdevice.com/unmoderated.php)

120
doc/DOCUMENTATION_X86.md Normal file
View File

@@ -0,0 +1,120 @@
### 2. Why differences between Intel and AMD?
There are many different CPUID leaves [[1](#references)]. In some cases, a given leaf does the same thing in Intel and AMD processors, but in the majority of them, they don't. For example, leaf 0x4 gives you the caches information, but in AMD is a reserved (invalid) leaf! In the case of AMD, is more common to fetch information using extended levels than using the standard levels (the other way around with Intel).
### 3. How to get the frequency?
__Involved code: [get_frequency_info (cpuid.c)](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/cpuid.c)__
CPUID leaf 0x16 is used.
If the CPU does not support supports such level:
- Linux: cpufetch will try to obtain this information using `/sys` filesystem in Linux. I think that Linux knows the frequency using model specific registers (MSRs), which you can't read at the user level.
- Windows: cpufetch can't obtain CPU frequency. This means that peak performance can't be computed because the frequency is needed to compute it.
### 4. How to get cache sizes?
__Involved code: [get_cache_info (cpuid.c)](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/cpuid.c)__
- Intel: CPUID leaf 0x4 is used (using __get_cache_info_general__). If the CPU does not support it, cpufetch can't get this information.
- AMD: Extended CPUID leaf 0x1D is used (using __get_cache_info_general__). If the CPU does not support this level, cpufetch uses a fallback method, which uses extended leaves 0x5 and 0x6. This fallback method uses __get_cache_info_amd_fallback__.
### 5. How to get CPU microarchitecture?
__Involved code: [get_cpu_uarch (cpuid.c)](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/cpuid.c), [get_uarch_from_cpuid (uarch.c)](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/uarch.c)__
CPUID leaf 0x1 is used. From there, we get:
- Model
- Extended Model
- Family
- Extended Family
- Stepping
Knowing this information, we can distinguish any CPU microarchitecture. Inside __uarch.c__ there is a function that behaves like a database or a lookup table. The function of this database is to find a match between the information obtained from 0x1 and what kind of microarchitecture the current CPU is. I got the data using and adapting the code from Todd Allen's cpuid program [[5](#references)]. Knowing the microarchitecture, we can obtain the manufacturing process (or technology, the size in nm of the transistors).
### 6. How to get CPU topology?
__Involved code: [cpuid.h](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/cpuid.h), [get_topology_info (cpuid.c)](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/cpuid.c), [apic.c](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/apic.c)__
cpufetch aims to support the most complex systems, so it supports multi-socket CPUs and detailed SMT (Intel HyperThreading) information. The CPU topology is stored in the following struct:
```
struct topology {
int64_t total_cores;
uint32_t physical_cores;
uint32_t logical_cores;
uint32_t smt_available;
uint32_t smt_supported;
uint32_t sockets;
};
```
This structure needs a bit of explanation, to know what are we looking for:
- `physical_cores`: Number of physical cores. In a multi socket system, this field stores the number of cores for just one socket.
- `logical_cores`: Number of logical cores. In a multi socket system, this field stores the number of logical cores for just one socket.
- `total_cores`: Total number of logical cores. In a multi socket system, this field stores the number of logical cores for the entire system.
- `sockets`: How many sockets the system has.
- `smt_supported`: Stores if SMT (or Intel HT) is supported in the CPU, storing the number of threads per core. So, if `smt_supported == 1`, it means that there is 1 thread per core, and SMT is not supported. If SMT is supported, then `smt_supported >= 1`. Note this field tells if the CPU if supports it, but not if SMT is activated or not.
- `smt_available`: The same idea as `smt_supported`, but it stores if SMT is available. If SMT is not supported, then `smt_available` is always `1`. The differentiation between supported and available lets cpufetch distinguish when a CPU has SMT capabilities, but are disabled (probably in the BIOS).
Let's give two CPU examples and the values that `struct topology` would have in these CPUs.
- Example 1: Dual Socket Intel Xeon 6248:
```
total_cores = 80
physical_cores = 20
logical_cores = 40
smt_available = 2
smt_supported = 2
sockets = 2
```
- Example 2: Intel Core i7-4790K with SMT disabled in BIOS:
```
total_cores = 8
physical_cores = 4
logical_cores = 8
smt_available = 1
smt_supported = 2
sockets = 1
```
Now that we know what data are we looking for, let's see how we get it:
- __Intel__: The methodology used is explained in the Intel webpage [[2](#references)]. Intel explains how to do it and also gives an example source code. I used it and modified it to fit cpufetch style. The core of this methodology is the usage of the APIC id, so the code is inside __apic.c__.
- __AMD__: Intel's algorithm using APIC does not work for AMD. To get the same information in AMD, I used the reference from OSdev [[3](#references)] and also ideas from lscpu [[4](#references)]. This uses:
- CPUID extended leaf 0x8: Fill `logical_cores`
- CPUID extended leaf 0x1E: Fill `smt_supported`
- CPUID standard leaf 0x1 (APIC): Fill `smt_available`
If any of these levels are not supported, these fields are just guessed. For example, if we are not able to know if SMT is supported, we guess it is not. With all of these data, we can calculate the rest of the fields:
```
physical_cores = logical_cores / smt_available;
if(topo->smt_supported > 1)
sockets = total_cores / smt_supported / physical_cores; // Idea borrowed from lscpu
else
sockets = total_cores / physical_cores;
```
### 7. How to get cache topology?
__Involved code: [get_cache_topology_amd (cpuid.c)](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/cpuid.c), [apic.c](https://github.com/Dr-Noob/cpufetch/blob/master/src/x86/apic.c)__
The topology of a cache gives us information about how many caches we have at a given level. It usually follows the rule of:
- L1: The same as the number of cores (one L1i and one L1d per core).
- L2: If L2 is the last level cache, one L2. If not, the same as the number of cores (one L2 per core).
- L3: One L3 cache per socket (shared among all cores).
These assumptions are generally (but not always) true. For example, for the AMD Zen generation, or the Intel Xeon Phi KNL. Thus, cpufetch does not assume the topology but obtains it instead.
- __Intel__: The idea is similar to the mentioned in CPU topology [[2](#references)](it also covers how to get cache topology using APIC id).
- __AMD__: Again, we have to look for another path for AMD. This time, the way to do it is easier and (I think) more solid and future proof. The idea is to use extended CPUID leaf 0x1D. If the CPU does not support it, we can still guess the topology of the caches (as mentioned earlier). If it does, CPUID can give us how many cores shares a given level of cache. So, if we have the number of cores, we can guess how many caches are there for any given level (see __get_cache_topology_amd__).
#### References
- [1] [sandpile CPUID webpage](https://www.sandpile.org/x86/cpuid.htm)
- [2] [CPU topology and cache topology: Intel](https://software.intel.com/content/www/us/en/develop/articles/intel-64-architecture-processor-topology-enumeration.html)
- [3] [CPU topology: AMD](https://wiki.osdev.org/Detecting_CPU_Topology_(80x86))
- [4] [lscpu](https://github.com/karelzak/util-linux/blob/master/sys-utils/lscpu.c)
- [5] [Todd Allen's cpuid](http://www.etallen.com/cpuid.html)
- [6] [AMD specific CPUID specification](https://www.amd.com/system/files/TechDocs/25481.pdf)
- [7] [Intel vs AMD CPU Architectural Differences: Chips and Chiplets](https://c.mi.com/thread-2585048-1-0.html)
In addition to all these resources, I found it very interesting to search in the Linux kernel source code (for example, the directory [`arch/x86/kernel/cpu/`](https://elixir.bootlin.com/linux/latest/source/arch/x86/kernel/cpu)), because sometimes you can find ideas that cannot be found anywhere else!

59
doc/README.md Normal file
View File

@@ -0,0 +1,59 @@
# cpufetch programming documentation (v0.94)
This documentation explains how cpufetch works internally and all the design decisions I made. This document intends to be useful for me in the future, for everyone interested in the project, and for anyone who is trying to obtain any specific information from the CPU. In this way, this can be used as a manual or a page that collects interesting material in this area.
### 1. Basics
cpufetch works for __x86_64__ (Intel and AMD) and __ARM__ CPUs. However, cpufetch is expected to work better on x86_64, because the codebase is older and has been tested much more than the ARM version. Other kinds of x86_64 CPU are not supported (I don't think supporting other CPUs may pay off). Depending on the architecture, cpufetch choose certain files to be compiled. A summarized tree of the source code of cpufetch is shown below.
```
cpufetch/
├── doc
│   ├── DOCUMENTATION_ARM.md
│   ├── DOCUMENTATION_X86.md
│   └── README.md
├── Makefile
├── README.md
└── src/
├── arm/
│   ├── midr.c
│   ├── midr.h
│   └── other files ...
├── common/
│   └── common files ...
└── x86/
├── cpuid.c
├── cpuid.h
└── other files ...
```
Source code is divided into three directories:
- `common/`: Source code shared between x86 and ARM
- `arm/`: ARM dependant source code
- `x86/`: x86_64 dependant source code
##### 1.1 Basics (x86_64)
In x86, __cpufetch works using the CPUID instruction__. It is called directly using assembly (see `src/x86/cpuid_asm.c`). To understand how CPUID works, see [DOCUMENTATION_X86.md](https://github.com/Dr-Noob/cpufetch/blob/master/doc/DOCUMENTATION_X86.md).
At the beginning of execution, cpufetch needs to know the max standard CPUID level and max CPUID extended level supported in the running CPU. We also need to know if the x86 CPU is Intel or AMD because sometimes, the way to obtain the information depends on the manufacturer. This information will be stored in:
```
struct cpuInfo {
...
VENDOR cpu_vendor;
uint32_t maxLevels;
uint32_t maxExtendedLevels;
...
};
```
To use any CPUID leaf, cpufetch always needs to check that it is supported in the current CPU.
##### 1.2 Basics (ARM)
In ARM, __cpufetch works using the MIDR register and Linux filesystem__. MIDR (Main ID Register) is read from `/proc/cpuinfo`. It allows the detection of the microarchitecture of the cores. Furthermore, Linux filesystem `/sys/devices/system/cpu/` is used to fetch the number of cores and other information. This is the main reason to explain __why `cpufetch` only works on Linux kernel based systems.__
##### 1.3 Documentation organization
The rest of the documentation is divided into x86 and ARM architectures since each one needs different implementations:
- [DOCUMENTATION_X86.md](https://github.com/Dr-Noob/cpufetch/blob/master/doc/DOCUMENTATION_X86.md)
- [DOCUMENTATION_ARM.md](https://github.com/Dr-Noob/cpufetch/blob/master/doc/DOCUMENTATION_ARM.md)

BIN
pictures/cascade_lake.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 42 KiB

BIN
pictures/cpufetch.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 76 KiB

BIN
pictures/epyc.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 33 KiB

BIN
pictures/exynos.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 44 KiB

BIN
pictures/i9.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 34 KiB

BIN
pictures/snapdragon.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 43 KiB

Binary file not shown.

Before

Width:  |  Height:  |  Size: 32 KiB

View File

@@ -1,260 +0,0 @@
#ifdef _WIN32
#include <windows.h>
#else
#define _GNU_SOURCE
#include <sched.h>
#endif
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#include "apic.h"
#include "cpuid_asm.h"
#include "global.h"
/*
* bit_scan_reverse and create_mask code taken from:
* https://software.intel.com/content/www/us/en/develop/articles/intel-64-architecture-processor-topology-enumeration.html
*/
unsigned char bit_scan_reverse(uint32_t* index, uint64_t mask) {
for(uint64_t i = (8 * sizeof(uint64_t)); i > 0; i--) {
if((mask & (1LL << (i-1))) != 0) {
*index = (uint64_t) (i-1);
break;
}
}
return (unsigned char) (mask != 0);
}
uint32_t create_mask(uint32_t num_entries, uint32_t *mask_width) {
uint32_t i = 0;
uint64_t k = 0;
// NearestPo2(numEntries) is the nearest power of 2 integer that is not less than numEntries
// The most significant bit of (numEntries * 2 -1) matches the above definition
k = (uint64_t)(num_entries) * 2 -1;
if (bit_scan_reverse(&i, k) == 0) {
if (mask_width) *mask_width = 0;
return 0;
}
if (mask_width) *mask_width = i;
if (i == 31) return (uint32_t ) -1;
return (1 << i) -1;
}
uint32_t get_apic_id(bool x2apic_id) {
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
if(x2apic_id) {
eax = 0x0000000B;
cpuid(&eax, &ebx, &ecx, &edx);
return edx;
}
else {
eax = 0x00000001;
cpuid(&eax, &ebx, &ecx, &edx);
return (ebx >> 24);
}
}
bool bind_to_cpu(int cpu_id) {
#ifdef _WIN32
HANDLE process = GetCurrentProcess();
DWORD_PTR processAffinityMask = 1 << cpu_id;
return SetProcessAffinityMask(process, processAffinityMask);
#else
cpu_set_t currentCPU;
CPU_ZERO(&currentCPU);
CPU_SET(cpu_id, &currentCPU);
if (sched_setaffinity (0, sizeof(currentCPU), &currentCPU) == -1) {
perror("sched_setaffinity");
return false;
}
return true;
#endif
}
bool fill_topo_masks_apic(struct topology** topo) {
uint32_t eax = 0x00000001;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
uint32_t core_plus_smt_id_max_cnt;
uint32_t core_id_max_cnt;
uint32_t smt_id_per_core_max_cnt;
cpuid(&eax, &ebx, &ecx, &edx);
core_plus_smt_id_max_cnt = (ebx >> 16) & 0xFF;
eax = 0x00000004;
ecx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
core_id_max_cnt = (eax >> 26) + 1;
smt_id_per_core_max_cnt = core_plus_smt_id_max_cnt / core_id_max_cnt;
(*topo)->apic->smt_mask = create_mask(smt_id_per_core_max_cnt, &((*topo)->apic->smt_mask_width));
(*topo)->apic->core_mask = create_mask(core_id_max_cnt,&((*topo)->apic->pkg_mask_shift));
(*topo)->apic->pkg_mask_shift += (*topo)->apic->smt_mask_width;
(*topo)->apic->core_mask <<= (*topo)->apic->smt_mask_width;
(*topo)->apic->pkg_mask = (-1) ^ ((*topo)->apic->core_mask | (*topo)->apic->smt_mask);
return true;
}
bool fill_topo_masks_x2apic(struct topology** topo) {
int32_t level_type;
int32_t level_shift;
int32_t coreplus_smt_mask = 0;
bool level2 = false;
bool level1 = false;
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
uint32_t i = 0;
while(true) {
eax = 0x0000000B;
ecx = i;
cpuid(&eax, &ebx, &ecx, &edx);
if(ebx == 0) break;
level_type = (ecx >> 8) & 0xFF;
level_shift = eax & 0xFFF;
switch(level_type) {
case 1: // SMT
(*topo)->apic->smt_mask = ~(0xFFFFFFFF << level_shift);
(*topo)->apic->smt_mask_width = level_shift;
(*topo)->smt_supported = ebx & 0xFFFF;
level1 = true;
break;
case 2: // Core
coreplus_smt_mask = ~(0xFFFFFFFF << level_shift);
(*topo)->apic->pkg_mask_shift = level_shift;
(*topo)->apic->pkg_mask = (-1) ^ coreplus_smt_mask;
level2 = true;
break;
default:
printErr("Found invalid level when querying topology: %d", level_type);
break;
}
i++; // sublevel to query
}
if (level1 && level2) {
(*topo)->apic->core_mask = coreplus_smt_mask ^ (*topo)->apic->smt_mask;
}
else if (!level2 && level1) {
(*topo)->apic->core_mask = 0;
(*topo)->apic->pkg_mask_shift = (*topo)->apic->smt_mask_width;
(*topo)->apic->pkg_mask = (-1) ^ (*topo)->apic->smt_mask;
}
else {
printErr("SMT level was not found when querying topology");
return false;
}
return true;
}
bool build_topo_from_apic(uint32_t* apic_pkg, uint32_t* apic_smt, struct topology** topo) {
uint32_t sockets[64];
uint32_t smt[64];
memset(sockets, 0, sizeof(uint32_t) * 64);
memset(smt, 0, sizeof(uint32_t) * 64);
for(int i=0; i < (*topo)->total_cores; i++) {
sockets[apic_pkg[i]] = 1;
smt[apic_smt[i]] = 1;
}
for(int i=0; i < 64; i++) {
if(sockets[i] != 0)
(*topo)->sockets++;
if(smt[i] != 0)
(*topo)->smt_available++;
}
(*topo)->logical_cores = (*topo)->total_cores / (*topo)->sockets;
(*topo)->physical_cores = (*topo)->logical_cores / (*topo)->smt_available;
return true;
}
bool get_topology_from_apic(uint32_t cpuid_max_levels, struct topology** topo) {
uint32_t apic_id;
uint32_t* apic_pkg = malloc(sizeof(uint32_t) * (*topo)->total_cores);
uint32_t* apic_core = malloc(sizeof(uint32_t) * (*topo)->total_cores);
uint32_t* apic_smt = malloc(sizeof(uint32_t) * (*topo)->total_cores);
bool x2apic_id = cpuid_max_levels >= 0x0000000B;
if(x2apic_id) {
if(!fill_topo_masks_x2apic(topo))
return false;
}
else {
if(!fill_topo_masks_apic(topo))
return false;
}
for(int i=0; i < (*topo)->total_cores; i++) {
if(!bind_to_cpu(i)) {
printErr("Failed binding to CPU %d", i);
return false;
}
apic_id = get_apic_id(x2apic_id);
apic_pkg[i] = (apic_id & (*topo)->apic->pkg_mask) >> (*topo)->apic->pkg_mask_shift;
apic_core[i] = (apic_id & (*topo)->apic->core_mask) >> (*topo)->apic->smt_mask_width;
apic_smt[i] = apic_id & (*topo)->apic->smt_mask;
}
/* DEBUG
for(int i=0; i < (*topo)->total_cores; i++)
printf("[%2d] 0x%.8X\n", i, apic_pkg[i]);
printf("\n");
for(int i=0; i < (*topo)->total_cores; i++)
printf("[%2d] 0x%.8X\n", i, apic_core[i]);
printf("\n");
for(int i=0; i < (*topo)->total_cores; i++)
printf("[%2d] 0x%.8X\n", i, apic_smt[i]);*/
bool ret = build_topo_from_apic(apic_pkg, apic_smt, topo);
// Assumption: If we cant get smt_available, we assume it is equal to smt_supported...
if(!x2apic_id) (*topo)->smt_supported = (*topo)->smt_available;
return ret;
}
// Used by AMD
uint32_t is_smt_enabled(struct topology* topo) {
uint32_t id;
for(int i = 0; i < topo->total_cores; i++) {
if(!bind_to_cpu(i)) {
printErr("Failed binding to CPU %d", i);
return false;
}
id = get_apic_id(true) & 1; // get the last bit
if(id == 1) return 2; // We assume there isn't any AMD CPU with more than 2th per core
}
return 1;
}

View File

@@ -1,230 +0,0 @@
#include <getopt.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include "args.h"
#include "global.h"
#define ARG_STR_STYLE "style"
#define ARG_STR_COLOR "color"
#define ARG_STR_HELP "help"
#define ARG_STR_LEVELS "levels"
#define ARG_STR_VERBOSE "verbose"
#define ARG_STR_VERSION "version"
#define ARG_CHAR_STYLE 's'
#define ARG_CHAR_COLOR 'c'
#define ARG_CHAR_HELP 'h'
#define ARG_CHAR_LEVELS 'l'
#define ARG_CHAR_VERBOSE 'v'
#define ARG_CHAR_VERSION 'v'
#define STYLE_STR_1 "fancy"
#define STYLE_STR_2 "retro"
#define STYLE_STR_3 "legacy"
struct args_struct {
bool levels_flag;
bool help_flag;
bool verbose_flag;
bool version_flag;
STYLE style;
struct colors* colors;
};
static const char* SYTLES_STR_LIST[STYLES_COUNT] = { STYLE_STR_1, STYLE_STR_2, STYLE_STR_3 };
static struct args_struct args;
STYLE get_style() {
return args.style;
}
struct colors* get_colors() {
return args.colors;
}
bool show_help() {
return args.help_flag;
}
bool show_version() {
return args.version_flag;
}
bool show_levels() {
return args.levels_flag;
}
bool verbose_enabled() {
return args.verbose_flag;
}
STYLE parse_style(char* style) {
int i = 0;
while(i != STYLES_COUNT && strcmp(SYTLES_STR_LIST[i],style) != 0)
i++;
if(i == STYLES_COUNT)
return STYLE_INVALID;
return i;
}
void free_colors_struct(struct colors* cs) {
free(cs->c1);
free(cs->c2);
free(cs->c3);
free(cs->c4);
free(cs);
}
bool parse_color(char* optarg, struct colors** cs) {
*cs = malloc(sizeof(struct colors));
(*cs)->c1 = malloc(sizeof(struct color));
(*cs)->c2 = malloc(sizeof(struct color));
(*cs)->c3 = malloc(sizeof(struct color));
(*cs)->c4 = malloc(sizeof(struct color));
struct color** c1 = &((*cs)->c1);
struct color** c2 = &((*cs)->c2);
struct color** c3 = &((*cs)->c3);
struct color** c4 = &((*cs)->c4);
int32_t ret;
ret = sscanf(optarg, "%d,%d,%d:%d,%d,%d:%d,%d,%d:%d,%d,%d",
&(*c1)->R, &(*c1)->G, &(*c1)->B,
&(*c2)->R, &(*c2)->G, &(*c2)->B,
&(*c3)->R, &(*c3)->G, &(*c3)->B,
&(*c4)->R, &(*c4)->G, &(*c4)->B);
if(ret != 12) {
printErr("Expected to read 12 values for color but read %d", ret);
return false;
}
//TODO: Refactor c1->R c2->R ... to c[i]->R
if((*c1)->R < 0 || (*c1)->R > 255) {
printErr("Red in color 1 is invalid. Must be in range (0, 255)");
return false;
}
if((*c1)->G < 0 || (*c1)->G > 255) {
printErr("Green in color 1 is invalid. Must be in range (0, 255)");
return false;
}
if((*c1)->B < 0 || (*c1)->B > 255) {
printErr("Blue in color 1 is invalid. Must be in range (0, 255)");
return false;
}
if((*c2)->R < 0 || (*c2)->R > 255) {
printErr("Red in color 2 is invalid. Must be in range (0, 255)");
return false;
}
if((*c2)->G < 0 || (*c2)->G > 255) {
printErr("Green in color 2 is invalid. Must be in range (0, 255)");
return false;
}
if((*c2)->B < 0 || (*c2)->B > 255) {
printErr("Blue in color 2 is invalid. Must be in range (0, 255)");
return false;
}
return true;
}
bool parse_args(int argc, char* argv[]) {
int c;
int option_index = 0;
opterr = 0;
bool color_flag = false;
args.levels_flag = false;
args.verbose_flag = false;
args.help_flag = false;
args.style = STYLE_EMPTY;
args.colors = NULL;
static struct option long_options[] = {
{ARG_STR_STYLE, required_argument, 0, ARG_CHAR_STYLE },
{ARG_STR_COLOR, required_argument, 0, ARG_CHAR_COLOR },
{ARG_STR_HELP, no_argument, 0, ARG_CHAR_HELP },
{ARG_STR_LEVELS, no_argument, 0, ARG_CHAR_LEVELS },
{ARG_STR_VERBOSE, no_argument, 0, ARG_CHAR_VERBOSE },
{ARG_STR_VERSION, no_argument, 0, ARG_CHAR_VERSION },
{0, 0, 0, 0}
};
c = getopt_long(argc, argv, "", long_options, &option_index);
while (c != -1) {
if(c == ARG_CHAR_COLOR) {
if(color_flag) {
printErr("Color option specified more than once");
return false;
}
color_flag = true;
if(!parse_color(optarg, &args.colors)) {
printErr("Color parsing failed");
return false;
}
}
else if(c == ARG_CHAR_STYLE) {
if(args.style != STYLE_EMPTY) {
printErr("Style option specified more than once");
return false;
}
args.style = parse_style(optarg);
if(args.style == STYLE_INVALID) {
printErr("Invalid style '%s'",optarg);
return false;
}
}
else if(c == ARG_CHAR_HELP) {
if(args.help_flag) {
printErr("Help option specified more than once");
return false;
}
args.help_flag = true;
}
else if(c == ARG_CHAR_VERBOSE) {
if(args.verbose_flag) {
printErr("Verbose option specified more than once");
return false;
}
args.verbose_flag = true;
}
else if(c == ARG_CHAR_LEVELS) {
if(args.levels_flag) {
printErr("Levels option specified more than once");
return false;
}
args.levels_flag = true;
}
else if (c == ARG_CHAR_VERSION) {
if(args.version_flag) {
printErr("Version option specified more than once");
return false;
}
args.version_flag = true;
}
else if(c == '?') {
printWarn("Invalid options");
args.help_flag = true;
break;
}
else
printBug("Bug at line number %d in file %s", __LINE__, __FILE__);
option_index = 0;
c = getopt_long(argc, argv,"",long_options, &option_index);
}
if (optind < argc) {
printWarn("Invalid options");
args.help_flag = true;
}
if((args.help_flag + args.version_flag + color_flag) > 1) {
printWarn("You should specify just one option");
args.help_flag = true;
}
return true;
}

330
src/arm/midr.c Normal file
View File

@@ -0,0 +1,330 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <stdbool.h>
#include <errno.h>
#include <sys/auxv.h>
#include <asm/hwcap.h>
#include "../common/global.h"
#include "udev.h"
#include "midr.h"
#include "uarch.h"
#include "soc.h"
#define STRING_UNKNOWN "Unknown"
void init_cache_struct(struct cache* cach) {
cach->L1i = malloc(sizeof(struct cach));
cach->L1d = malloc(sizeof(struct cach));
cach->L2 = malloc(sizeof(struct cach));
cach->L3 = malloc(sizeof(struct cach));
cach->cach_arr = malloc(sizeof(struct cach*) * 4);
cach->cach_arr[0] = cach->L1i;
cach->cach_arr[1] = cach->L1d;
cach->cach_arr[2] = cach->L2;
cach->cach_arr[3] = cach->L3;
cach->max_cache_level = 0;
cach->L1i->exists = false;
cach->L1d->exists = false;
cach->L2->exists = false;
cach->L3->exists = false;
}
struct cache* get_cache_info(struct cpuInfo* cpu) {
struct cache* cach = malloc(sizeof(struct cache));
init_cache_struct(cach);
cach->max_cache_level = 2;
for(int i=0; i < cach->max_cache_level + 1; i++) {
cach->cach_arr[i]->exists = true;
cach->cach_arr[i]->num_caches = 1;
cach->cach_arr[i]->size = 0;
}
return cach;
}
struct frequency* get_frequency_info(uint32_t core) {
struct frequency* freq = malloc(sizeof(struct frequency));
freq->base = UNKNOWN_FREQ;
freq->max = get_max_freq_from_file(core, false);
return freq;
}
struct topology* get_topology_info(struct cpuInfo* cpu, struct cache* cach, uint32_t* midr_array, int socket_idx, int ncores) {
struct topology* topo = malloc(sizeof(struct topology));
topo->cach = cach;
topo->total_cores = 0;
int sockets_seen = 0;
int first_core_idx = 0;
int currrent_core_idx = 0;
int cores_in_socket = 0;
while(socket_idx + 1 > sockets_seen) {
if(midr_array[first_core_idx] == midr_array[currrent_core_idx] && currrent_core_idx < ncores) {
currrent_core_idx++;
cores_in_socket++;
}
else {
topo->total_cores = cores_in_socket;
cores_in_socket = 0;
first_core_idx = currrent_core_idx;
sockets_seen++;
}
}
return topo;
}
bool cores_are_equal(int c1pos, int c2pos, uint32_t* midr_array, int32_t* freq_array) {
return midr_array[c1pos] == midr_array[c2pos] && freq_array[c1pos] == freq_array[c2pos];
}
uint32_t fill_ids_from_midr(uint32_t* midr_array, int32_t* freq_array, uint32_t* ids_array, int len) {
uint32_t latest_id = 0;
bool found;
ids_array[0] = latest_id;
for (int i = 1; i < len; i++) {
int j = 0;
found = false;
for (j = 0; j < len && !found; j++) {
if (i != j && cores_are_equal(i, j, midr_array, freq_array)) {
if(j > i) {
latest_id++;
ids_array[i] = latest_id;
}
else {
ids_array[i] = ids_array[j];
}
found = true;
}
}
if(!found) {
latest_id++;
ids_array[i] = latest_id;
}
}
return latest_id+1;
}
void init_cpu_info(struct cpuInfo* cpu) {
cpu->next_cpu = NULL;
}
// We assume all cpus share the same hardware
// capabilities but I'm not sure it is always
// true...
// ARM32 https://elixir.bootlin.com/linux/latest/source/arch/arm/include/uapi/asm/hwcap.h
// ARM64 https://elixir.bootlin.com/linux/latest/source/arch/arm64/include/uapi/asm/hwcap.h
struct features* get_features_info() {
struct features* feat = malloc(sizeof(struct features));
bool *ptr = &(feat->AES);
for(uint32_t i = 0; i < sizeof(struct features)/sizeof(bool); i++, ptr++) {
*ptr = false;
}
errno = 0;
long hwcaps = getauxval(AT_HWCAP);
if(errno == ENOENT) {
printWarn("Unable to retrieve AT_HWCAP using getauxval");
}
#ifdef __aarch64__
else {
feat->AES = hwcaps & HWCAP_AES;
feat->CRC32 = hwcaps & HWCAP_CRC32;
feat->SHA1 = hwcaps & HWCAP_SHA1;
feat->SHA2 = hwcaps & HWCAP_SHA2;
feat->NEON = hwcaps & HWCAP_ASIMD;
}
#else
else {
feat->NEON = hwcaps & HWCAP_NEON;
}
hwcaps = getauxval(AT_HWCAP2);
if(errno == ENOENT) {
printWarn("Unable to retrieve AT_HWCAP2 using getauxval");
}
else {
feat->AES = hwcaps & HWCAP2_AES;
feat->CRC32 = hwcaps & HWCAP2_CRC32;
feat->SHA1 = hwcaps & HWCAP2_SHA1;
feat->SHA2 = hwcaps & HWCAP2_SHA2;
}
#endif
return feat;
}
struct cpuInfo* get_cpu_info() {
struct cpuInfo* cpu = malloc(sizeof(struct cpuInfo));
init_cpu_info(cpu);
int ncores = get_ncores_from_cpuinfo();
bool success = false;
int32_t* freq_array = malloc(sizeof(uint32_t) * ncores);
uint32_t* midr_array = malloc(sizeof(uint32_t) * ncores);
uint32_t* ids_array = malloc(sizeof(uint32_t) * ncores);
for(int i=0; i < ncores; i++) {
midr_array[i] = get_midr_from_cpuinfo(i, &success);
if(!success) {
printWarn("Unable to fetch MIDR for core %d. This is probably because the core is offline", i);
midr_array[i] = midr_array[0];
}
freq_array[i] = get_max_freq_from_file(i, false);
if(freq_array[i] == UNKNOWN_FREQ) {
printWarn("Unable to fetch max frequency for core %d. This is probably because the core is offline", i);
freq_array[i] = freq_array[0];
}
}
uint32_t sockets = fill_ids_from_midr(midr_array, freq_array, ids_array, ncores);
struct cpuInfo* ptr = cpu;
int midr_idx = 0;
int tmp_midr_idx = 0;
for(uint32_t i=0; i < sockets; i++) {
if(i > 0) {
ptr->next_cpu = malloc(sizeof(struct cpuInfo));
ptr = ptr->next_cpu;
init_cpu_info(ptr);
tmp_midr_idx = midr_idx;
while(cores_are_equal(midr_idx, tmp_midr_idx, midr_array, freq_array)) tmp_midr_idx++;
midr_idx = tmp_midr_idx;
}
ptr->midr = midr_array[midr_idx];
ptr->arch = get_uarch_from_midr(ptr->midr, ptr);
ptr->feat = get_features_info();
ptr->freq = get_frequency_info(midr_idx);
ptr->cach = get_cache_info(ptr);
ptr->topo = get_topology_info(ptr, ptr->cach, midr_array, i, ncores);
}
cpu->num_cpus = sockets;
cpu->hv = malloc(sizeof(struct hypervisor));
cpu->hv->present = false;
cpu->soc = get_soc();
return cpu;
}
char* get_str_topology(struct cpuInfo* cpu, struct topology* topo, bool dual_socket) {
uint32_t size = 3+7+1;
char* string = malloc(sizeof(char)*size);
snprintf(string, size, "%d cores", topo->total_cores);
return string;
}
char* get_str_peak_performance(struct cpuInfo* cpu) {
//7 for GFLOP/s and 6 for digits,eg 412.14
uint32_t size = 7+6+1+1;
assert(strlen(STRING_UNKNOWN)+1 <= size);
char* string = malloc(sizeof(char)*size);
struct cpuInfo* ptr = cpu;
//First check we have consistent data
for(int i=0; i < cpu->num_cpus; ptr = ptr->next_cpu, i++) {
if(get_freq(ptr->freq) == UNKNOWN_FREQ) {
snprintf(string, strlen(STRING_UNKNOWN)+1, STRING_UNKNOWN);
return string;
}
}
double flops = 0.0;
ptr = cpu;
for(int i=0; i < cpu->num_cpus; ptr = ptr->next_cpu, i++) {
flops += ptr->topo->total_cores * (get_freq(ptr->freq) * 1000000);
}
if(cpu->feat->NEON) flops = flops * 4;
if(flops >= (double)1000000000000.0)
snprintf(string,size,"%.2f TFLOP/s",flops/1000000000000);
else if(flops >= 1000000000.0)
snprintf(string,size,"%.2f GFLOP/s",flops/1000000000);
else
snprintf(string,size,"%.2f MFLOP/s",flops/1000000);
return string;
}
char* get_str_features(struct cpuInfo* cpu) {
struct features* feat = cpu->feat;
char* string = malloc(sizeof(char) * 25);
uint32_t len = 0;
if(feat->NEON) {
strcat(string, "NEON,");
len += 5;
}
if(feat->SHA1) {
strcat(string, "SHA1,");
len += 5;
}
if(feat->SHA2) {
strcat(string, "SHA2,");
len += 5;
}
if(feat->AES) {
strcat(string, "AES,");
len += 4;
}
if(feat->CRC32) {
strcat(string, "CRC32,");
len += 6;
}
if(len > 0) {
string[len-1] = '\0';
return string;
}
else
return NULL;
}
void print_debug(struct cpuInfo* cpu) {
int ncores = get_ncores_from_cpuinfo();
bool success = false;
for(int i=0; i < ncores; i++) {
printf("[Core %d] ", i);
long freq = get_max_freq_from_file(i, false);
uint32_t midr = get_midr_from_cpuinfo(i, &success);
if(!success) {
printWarn("Unable to fetch MIDR for core %d. This is probably because the core is offline", i);
printf("0x%.8X ", get_midr_from_cpuinfo(0, &success));
}
else {
printf("0x%.8X ", midr);
}
if(freq == UNKNOWN_FREQ) {
printWarn("Unable to fetch max frequency for core %d. This is probably because the core is offline", i);
printf("%ld MHz\n", get_max_freq_from_file(0, false));
}
else {
printf("%ld MHz\n", freq);
}
}
}
void free_topo_struct(struct topology* topo) {
free(topo);
}

70
src/arm/midr.h Normal file
View File

@@ -0,0 +1,70 @@
#ifndef __MIDR__
#define __MIDR__
#include "../common/cpu.h"
struct cpuInfo* get_cpu_info();
uint32_t get_nsockets(struct topology* topo);
char* get_str_topology(struct cpuInfo* cpu, struct topology* topo, bool dual_socket);
char* get_str_peak_performance(struct cpuInfo* cpu);
char* get_str_features(struct cpuInfo* cpu);
void print_debug(struct cpuInfo* cpu);
void free_topo_struct(struct topology* topo);
// Code taken from cpuinfo (https://github.com/pytorch/cpuinfo/blob/master/src/arm/midr.h)
#define CPUINFO_ARM_MIDR_IMPLEMENTER_MASK UINT32_C(0xFF000000)
#define CPUINFO_ARM_MIDR_VARIANT_MASK UINT32_C(0x00F00000)
#define CPUINFO_ARM_MIDR_ARCHITECTURE_MASK UINT32_C(0x000F0000)
#define CPUINFO_ARM_MIDR_PART_MASK UINT32_C(0x0000FFF0)
#define CPUINFO_ARM_MIDR_REVISION_MASK UINT32_C(0x0000000F)
#define CPUINFO_ARM_MIDR_IMPLEMENTER_OFFSET 24
#define CPUINFO_ARM_MIDR_VARIANT_OFFSET 20
#define CPUINFO_ARM_MIDR_ARCHITECTURE_OFFSET 16
#define CPUINFO_ARM_MIDR_PART_OFFSET 4
#define CPUINFO_ARM_MIDR_REVISION_OFFSET 0
inline static uint32_t midr_set_implementer(uint32_t midr, uint32_t implementer) {
return (midr & ~CPUINFO_ARM_MIDR_IMPLEMENTER_MASK) |
((implementer << CPUINFO_ARM_MIDR_IMPLEMENTER_OFFSET) & CPUINFO_ARM_MIDR_IMPLEMENTER_MASK);
}
inline static uint32_t midr_set_variant(uint32_t midr, uint32_t variant) {
return (midr & ~CPUINFO_ARM_MIDR_VARIANT_MASK) |
((variant << CPUINFO_ARM_MIDR_VARIANT_OFFSET) & CPUINFO_ARM_MIDR_VARIANT_MASK);
}
inline static uint32_t midr_set_architecture(uint32_t midr, uint32_t architecture) {
return (midr & ~CPUINFO_ARM_MIDR_ARCHITECTURE_MASK) |
((architecture << CPUINFO_ARM_MIDR_ARCHITECTURE_OFFSET) & CPUINFO_ARM_MIDR_ARCHITECTURE_MASK);
}
inline static uint32_t midr_set_part(uint32_t midr, uint32_t part) {
return (midr & ~CPUINFO_ARM_MIDR_PART_MASK) |
((part << CPUINFO_ARM_MIDR_PART_OFFSET) & CPUINFO_ARM_MIDR_PART_MASK);
}
inline static uint32_t midr_set_revision(uint32_t midr, uint32_t revision) {
return (midr & ~CPUINFO_ARM_MIDR_REVISION_MASK) |
((revision << CPUINFO_ARM_MIDR_REVISION_OFFSET) & CPUINFO_ARM_MIDR_REVISION_MASK);
}
inline static uint32_t midr_get_variant(uint32_t midr) {
return (midr & CPUINFO_ARM_MIDR_VARIANT_MASK) >> CPUINFO_ARM_MIDR_VARIANT_OFFSET;
}
inline static uint32_t midr_get_implementer(uint32_t midr) {
return (midr & CPUINFO_ARM_MIDR_IMPLEMENTER_MASK) >> CPUINFO_ARM_MIDR_IMPLEMENTER_OFFSET;
}
inline static uint32_t midr_get_part(uint32_t midr) {
return (midr & CPUINFO_ARM_MIDR_PART_MASK) >> CPUINFO_ARM_MIDR_PART_OFFSET;
}
inline static uint32_t midr_get_revision(uint32_t midr) {
return (midr & CPUINFO_ARM_MIDR_REVISION_MASK) >> CPUINFO_ARM_MIDR_REVISION_OFFSET;
}
#endif

568
src/arm/soc.c Normal file
View File

@@ -0,0 +1,568 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include "soc.h"
#include "socs.h"
#include "udev.h"
#include "../common/global.h"
#define min(a,b) (((a)<(b))?(a):(b))
#define STRING_UNKNOWN "Unknown"
static char* soc_trademark_string[] = {
[SOC_VENDOR_SNAPDRAGON] = "Snapdragon ",
[SOC_VENDOR_MEDIATEK] = "MediaTek ",
[SOC_VENDOR_EXYNOS] = "Exynos ",
[SOC_VENDOR_KIRIN] = "Kirin ",
[SOC_VENDOR_BROADCOM] = "Broadcom BCM",
};
void fill_soc(struct system_on_chip* soc, char* soc_name, SOC soc_model, int32_t process) {
soc->soc_model = soc_model;
soc->soc_vendor = get_soc_vendor_from_soc(soc_model);
soc->process = process;
int len = strlen(soc_name) + strlen(soc_trademark_string[soc->soc_vendor]) + 1;
soc->soc_name = malloc(sizeof(char) * len);
memset(soc->soc_name, 0, sizeof(char) * len);
sprintf(soc->soc_name, "%s%s", soc_trademark_string[soc->soc_vendor], soc_name);
}
bool match_soc(struct system_on_chip* soc, char* raw_name, char* expected_name, char* soc_name, SOC soc_model, int32_t process) {
if(strlen(raw_name) > strlen(expected_name))
return false;
int len = strlen(raw_name);
if(strncmp(raw_name, expected_name, len) != 0) {
return false;
}
else {
fill_soc(soc, soc_name, soc_model, process);
return true;
}
}
char* toupperstr(char* str) {
int len = strlen(str) + 1;
char* ret = malloc(sizeof(char) * len);
memset(ret, 0, sizeof(char) * len);
for(int i=0; i < len; i++) {
ret[i] = toupper((unsigned char) str[i]);
}
return ret;
}
#define SOC_START if (false) {}
#define SOC_EQ(raw_name, expected_name, soc_name, soc_model, soc, process) \
else if (match_soc(soc, raw_name, expected_name, soc_name, soc_model, process)) return true;
#define SOC_END else { return false; }
// https://en.wikipedia.org/wiki/Raspberry_Pi
// http://phonedb.net/index.php?m=processor&id=562&c=broadcom_bcm21663
// https://hwbot.org/hardware/processors#key=bcmxxx
bool match_broadcom(char* soc_name, struct system_on_chip* soc) {
char* tmp;
if((tmp = strstr(soc_name, "BCM")) == NULL)
return false;
SOC_START
SOC_EQ(tmp, "BCM2835", "2835", SOC_BCM_2835, soc, 65)
SOC_EQ(tmp, "BCM2836", "2836", SOC_BCM_2836, soc, 40)
SOC_EQ(tmp, "BCM2837", "2837", SOC_BCM_2837, soc, 40)
SOC_EQ(tmp, "BCM2837B0", "2837B0", SOC_BCM_2837B0, soc, 40)
SOC_EQ(tmp, "BCM2711", "2711", SOC_BCM_2711, soc, 28)
SOC_EQ(tmp, "BCM21553", "21553", SOC_BCM_21553, soc, 65)
SOC_EQ(tmp, "BCM21553-Thunderbird", "21553 Thunderbird", SOC_BCM_21553T, soc, 65)
SOC_EQ(tmp, "BCM21663", "21663", SOC_BCM_21663, soc, 40)
SOC_EQ(tmp, "BCM21664", "21664", SOC_BCM_21664, soc, 40)
SOC_EQ(tmp, "BCM28155", "28155", SOC_BCM_28155, soc, 40)
SOC_EQ(tmp, "BCM23550", "23550", SOC_BCM_23550, soc, 40)
SOC_EQ(tmp, "BCM28145", "28145", SOC_BCM_28145, soc, 40)
SOC_EQ(tmp, "BCM2157", "2157", SOC_BCM_2157, soc, 65)
SOC_EQ(tmp, "BCM21654", "21654", SOC_BCM_21654, soc, 40)
SOC_END
}
// https://www.techinsights.com/
// https://datasheetspdf.com/pdf-file/1316605/HiSilicon/Hi3660/1
bool match_hisilicon(char* soc_name, struct system_on_chip* soc) {
char* tmp;
if((tmp = strstr(soc_name, "Hi")) == NULL)
return false;
SOC_START
SOC_EQ(tmp, "Hi3620GFC", "K3V2", SOC_HISILICON_3620, soc, 40)
//SOC_EQ(tmp, "?", "K3V2E", SOC_KIRIN, soc, ?)
//SOC_EQ(tmp, "?", "620", SOC_KIRIN, soc, 28)
//SOC_EQ(tmp, "?", "650", SOC_KIRIN, soc, 16)
//SOC_EQ(tmp, "?", "655", SOC_KIRIN, soc, 16)
//SOC_EQ(tmp, "?", "658", SOC_KIRIN, soc, 16)
//SOC_EQ(tmp, "?", "659", SOC_KIRIN, soc, 16)
//SOC_EQ(tmp, "?", "710", SOC_KIRIN, soc, 12)
//SOC_EQ(tmp, "?", "710A", SOC_KIRIN, soc, 12)
//SOC_EQ(tmp, "?", "710F", SOC_KIRIN, soc, 12)
//SOC_EQ(tmp, "?", "810", SOC_KIRIN, soc, 7)
//SOC_EQ(tmp, "?", "820", SOC_KIRIN, soc, 7)
//SOC_EQ(tmp, "?", "9000", SOC_KIRIN, soc, 5)
//SOC_EQ(tmp, "?", "9000E", SOC_KIRIN, soc, 5)
//SOC_EQ(tmp, "?", "910", SOC_KIRIN, soc, 28)
//SOC_EQ(tmp, "?", "910T", SOC_KIRIN, soc, 28)
SOC_EQ(tmp, "Hi3630", "920", SOC_HISILICON_3630, soc, 28)
//SOC_EQ(tmp, "?", "925", SOC_KIRIN, soc, 28)
//SOC_EQ(tmp, "?", "930", SOC_KIRIN, soc, ?)
//SOC_EQ(tmp, "?", "935", SOC_KIRIN, soc, ?)
SOC_EQ(tmp, "Hi3650", "950", SOC_HISILICON_3650, soc, 16)
//SOC_EQ(tmp, "?", "955", SOC_KIRIN, soc, ?)
SOC_EQ(tmp, "Hi3660", "960", SOC_HISILICON_3660, soc, 16)
//SOC_EQ(tmp, "?", "960S", SOC_KIRIN, soc, 16)
SOC_EQ(tmp, "Hi3670", "970", SOC_HISILICON_3670, soc, 10)
SOC_EQ(tmp, "Hi3680", "980", SOC_HISILICON_3680, soc, 7)
//SOC_EQ(tmp, "?", "985", SOC_KIRIN, soc, 7)
SOC_EQ(tmp, "Hi3690", "990", SOC_HISILICON_3690, soc, 7)
SOC_END
}
bool match_exynos(char* soc_name, struct system_on_chip* soc) {
char* tmp;
if((tmp = strstr(soc_name, "universal")) == NULL)
return false;
SOC_START
// universalXXXX //
SOC_EQ(tmp, "universal3475", "3475", SOC_EXYNOS_3475, soc, 28)
SOC_EQ(tmp, "universal4210", "4210", SOC_EXYNOS_4210, soc, 45)
SOC_EQ(tmp, "universal4212", "4212", SOC_EXYNOS_4212, soc, 32)
SOC_EQ(tmp, "universal4412", "4412", SOC_EXYNOS_4412, soc, 32)
SOC_EQ(tmp, "universal5250", "5250", SOC_EXYNOS_5250, soc, 32)
SOC_EQ(tmp, "universal5410", "5410", SOC_EXYNOS_5410, soc, 28)
SOC_EQ(tmp, "universal5420", "5420", SOC_EXYNOS_5420, soc, 28)
SOC_EQ(tmp, "universal5422", "5422", SOC_EXYNOS_5422, soc, 28)
SOC_EQ(tmp, "universal5430", "5430", SOC_EXYNOS_5430, soc, 20)
SOC_EQ(tmp, "universal5433", "5433", SOC_EXYNOS_5433, soc, 20)
SOC_EQ(tmp, "universal5260", "5260", SOC_EXYNOS_5260, soc, 28)
SOC_EQ(tmp, "universal7270", "7270", SOC_EXYNOS_7270, soc, 14)
SOC_EQ(tmp, "universal7420", "7420", SOC_EXYNOS_7420, soc, 14)
SOC_EQ(tmp, "universal7570", "7570", SOC_EXYNOS_7570, soc, 14)
SOC_EQ(tmp, "universal7870", "7870", SOC_EXYNOS_7870, soc, 14)
SOC_EQ(tmp, "universal7872", "7872", SOC_EXYNOS_7872, soc, 14)
SOC_EQ(tmp, "universal7880", "7880", SOC_EXYNOS_7880, soc, 14)
SOC_EQ(tmp, "universal7884", "7884", SOC_EXYNOS_7884, soc, 14)
SOC_EQ(tmp, "universal7885", "7885", SOC_EXYNOS_7885, soc, 14)
SOC_EQ(tmp, "universal7904", "7904", SOC_EXYNOS_7904, soc, 14)
SOC_EQ(tmp, "universal8890", "8890", SOC_EXYNOS_8890, soc, 14)
SOC_EQ(tmp, "universal8895", "8895", SOC_EXYNOS_8895, soc, 10)
SOC_EQ(tmp, "universal9110", "9110", SOC_EXYNOS_9110, soc, 14)
SOC_EQ(tmp, "universal9609", "9609", SOC_EXYNOS_9609, soc, 10)
SOC_EQ(tmp, "universal9610", "9610", SOC_EXYNOS_9610, soc, 10)
SOC_EQ(tmp, "universal9611", "9611", SOC_EXYNOS_9611, soc, 10)
SOC_EQ(tmp, "universal9810", "9810", SOC_EXYNOS_9810, soc, 10)
SOC_EQ(tmp, "universal9820", "9820", SOC_EXYNOS_9820, soc, 8)
SOC_EQ(tmp, "universal9825", "9825", SOC_EXYNOS_9825, soc, 7)
// New exynos. Dont know if they will work //
SOC_EQ(tmp, "universal1080", "1080", SOC_EXYNOS_1080, soc, 5)
SOC_EQ(tmp, "universal990", "990", SOC_EXYNOS_990, soc, 7)
SOC_EQ(tmp, "universal980", "980", SOC_EXYNOS_980, soc, 8)
SOC_EQ(tmp, "universal880", "880", SOC_EXYNOS_880, soc, 8)
SOC_END
}
bool match_mediatek(char* soc_name, struct system_on_chip* soc) {
char* tmp;
if((tmp = strstr(soc_name, "MT")) == NULL)
return false;
SOC_START
// Dimensity //
SOC_EQ(tmp, "MT6889", "Dimensity 1000", SOC_MTK_MT6889, soc, 7)
SOC_EQ(tmp, "MT6885Z", "Dimensity 1000L", SOC_MTK_MT6885Z, soc, 7)
//SOC_EQ(tmp, "?", "Dimensity 700", SOC_MTK_, soc, 7)
SOC_EQ(tmp, "MT6853", "Dimensity 720", SOC_MTK_MT6853, soc, 7)
SOC_EQ(tmp, "MT6873", "Dimensity 800", SOC_MTK_MT6873, soc, 7)
SOC_EQ(tmp, "MT6875", "Dimensity 820", SOC_MTK_MT6875, soc, 7)
// Helio //
SOC_EQ(tmp, "MT6761D", "Helio A20", SOC_MTK_MT6761D, soc, 12)
SOC_EQ(tmp, "MT6761", "Helio A22", SOC_MTK_MT6761, soc, 12)
SOC_EQ(tmp, "MT6762D", "Helio A25", SOC_MTK_MT6762D, soc, 12)
//SOC_EQ(tmp, "?", "Helio G25", SOC_MTK_, soc, 12)
//SOC_EQ(tmp, "?", "Helio G35", SOC_MTK_, soc, 12)
//SOC_EQ(tmp, "?", "Helio G70", SOC_MTK_, soc, 12)
//SOC_EQ(tmp, "?", "Helio G80", SOC_MTK_, soc, 12)
//SOC_EQ(tmp, "?", "Helio G90", SOC_MTK_, soc, 12)
//SOC_EQ(tmp, "?", "Helio G90T", SOC_MTK_, soc, 12)
//SOC_EQ(tmp, "?", "Helio G95", SOC_MTK_, soc, 12)
SOC_EQ(tmp, "MT6755", "Helio P10", SOC_MTK_MT6755M, soc, 28)
SOC_EQ(tmp, "MT6755M", "Helio P10 M", SOC_MTK_MT6755M, soc, 28)
SOC_EQ(tmp, "MT6755T", "Helio P15", SOC_MTK_MT6755T, soc, 28)
SOC_EQ(tmp, "MT6757", "Helio P20", SOC_MTK_MT6757, soc, 16)
SOC_EQ(tmp, "MT6762", "Helio P22", SOC_MTK_MT6762, soc, 12)
SOC_EQ(tmp, "MT6763V", "Helio P23", SOC_MTK_MT6763V, soc, 16)
SOC_EQ(tmp, "MT6763T", "Helio P23", SOC_MTK_MT6763T, soc, 16)
SOC_EQ(tmp, "MT6757CD", "Helio P25", SOC_MTK_MT6757CD, soc, 16)
SOC_EQ(tmp, "MT6758", "Helio P30", SOC_MTK_MT6758, soc, 16)
SOC_EQ(tmp, "MT6765", "Helio P35", SOC_MTK_MT6765, soc, 12)
SOC_EQ(tmp, "MT6771", "Helio P60", SOC_MTK_MT6771, soc, 12)
SOC_EQ(tmp, "MT6768", "Helio P65", SOC_MTK_MT6768, soc, 12)
SOC_EQ(tmp, "MT6771T", "Helio P70", SOC_MTK_MT6771, soc, 12)
SOC_EQ(tmp, "MT6771V", "Helio P70", SOC_MTK_MT6771, soc, 12)
SOC_EQ(tmp, "MT6779", "Helio P90", SOC_MTK_MT6779, soc, 12)
//SOC_EQ(tmp, "?", "Helio P95", SOC_MTK_, soc, 12)
SOC_EQ(tmp, "MT6795", "Helio X10", SOC_MTK_MT6795, soc, 28)
SOC_EQ(tmp, "MT6795T", "Helio X10 T", SOC_MTK_MT6795, soc, 28)
SOC_EQ(tmp, "MT6797", "Helio X20", SOC_MTK_MT6797, soc, 20)
SOC_EQ(tmp, "MT6797M", "Helio X20 M", SOC_MTK_MT6797, soc, 20)
SOC_EQ(tmp, "MT6797D", "Helio X23", SOC_MTK_MT6797, soc, 20)
SOC_EQ(tmp, "MT6797T", "Helio X25", SOC_MTK_MT6797T, soc, 20)
SOC_EQ(tmp, "MT6797X", "Helio X27", SOC_MTK_MT6797X, soc, 20)
SOC_EQ(tmp, "MT6799", "Helio X30", SOC_MTK_MT6799, soc, 10)
// MT XXXX //
SOC_EQ(tmp, "MT6515", "MT6515", SOC_MTK_MT6515, soc, 40)
SOC_EQ(tmp, "MT6516", "MT6516", SOC_MTK_MT6516, soc, 65)
SOC_EQ(tmp, "MT6517", "MT6517", SOC_MTK_MT6517, soc, 40)
SOC_EQ(tmp, "MT6572", "MT6572", SOC_MTK_MT6572, soc, 28)
SOC_EQ(tmp, "MT6572M", "MT6572M", SOC_MTK_MT6572M, soc, 28)
SOC_EQ(tmp, "MT6573", "MT6573", SOC_MTK_MT6573, soc, 65)
SOC_EQ(tmp, "MT6575", "MT6575", SOC_MTK_MT6575, soc, 40)
SOC_EQ(tmp, "MT6577", "MT6577", SOC_MTK_MT6577, soc, 40)
SOC_EQ(tmp, "MT6577T", "MT6577T", SOC_MTK_MT6577T, soc, 40)
SOC_EQ(tmp, "MT6580", "MT6580", SOC_MTK_MT6580, soc, 28)
SOC_EQ(tmp, "MT6582", "MT6582", SOC_MTK_MT6582, soc, 28)
SOC_EQ(tmp, "MT6582M", "MT6582M", SOC_MTK_MT6582M, soc, 28)
SOC_EQ(tmp, "MT6589", "MT6589", SOC_MTK_MT6589, soc, 28)
SOC_EQ(tmp, "MT6589T", "MT6589T", SOC_MTK_MT6589T, soc, 28)
SOC_EQ(tmp, "MT6592", "MT6592", SOC_MTK_MT6592, soc, 28)
SOC_EQ(tmp, "MT6595", "MT6595", SOC_MTK_MT6595, soc, 28)
SOC_EQ(tmp, "MT6732", "MT6732", SOC_MTK_MT6732, soc, 28)
SOC_EQ(tmp, "MT6735", "MT6735", SOC_MTK_MT6735, soc, 28)
SOC_EQ(tmp, "MT6735M", "MT6735M", SOC_MTK_MT6735M, soc, 28)
SOC_EQ(tmp, "MT6735P", "MT6735P", SOC_MTK_MT6735P, soc, 28)
SOC_EQ(tmp, "MT6737", "MT6737", SOC_MTK_MT6737, soc, 28)
SOC_EQ(tmp, "MT6737M", "MT6737M", SOC_MTK_MT6737M, soc, 28)
SOC_EQ(tmp, "MT6737T", "MT6737T", SOC_MTK_MT6737T, soc, 28)
SOC_EQ(tmp, "MT6739", "MT6739", SOC_MTK_MT6739, soc, 28)
SOC_EQ(tmp, "MT6750", "MT6750", SOC_MTK_MT6750, soc, 28)
SOC_EQ(tmp, "MT6750S", "MT6750S", SOC_MTK_MT6750S, soc, 28)
SOC_EQ(tmp, "MT6750T", "MT6750T", SOC_MTK_MT6750T, soc, 28)
SOC_EQ(tmp, "MT6752", "MT6752", SOC_MTK_MT6752, soc, 28)
SOC_EQ(tmp, "MT6753", "MT6753", SOC_MTK_MT6753, soc, 28)
SOC_EQ(tmp, "MT6850", "MT6850", SOC_MTK_MT6850, soc, 28)
SOC_EQ(tmp, "MT8121", "MT8121", SOC_MTK_MT8121, soc, 40)
SOC_EQ(tmp, "MT8125", "MT8125", SOC_MTK_MT8125, soc, 40)
SOC_EQ(tmp, "MT8127", "MT8127", SOC_MTK_MT8127, soc, 32)
SOC_EQ(tmp, "MT8135", "MT8135", SOC_MTK_MT8135, soc, 28)
SOC_EQ(tmp, "MT8163A", "MT8163A", SOC_MTK_MT8163A, soc, 28)
SOC_EQ(tmp, "MT8163B", "MT8163B", SOC_MTK_MT8163B, soc, 28)
SOC_EQ(tmp, "MT8167B", "MT8167B", SOC_MTK_MT8167B, soc, 28)
SOC_EQ(tmp, "MT8173", "MT8173", SOC_MTK_MT8173, soc, 28)
SOC_EQ(tmp, "MT8176", "MT8176", SOC_MTK_MT8176, soc, 28)
SOC_EQ(tmp, "MT8321", "MT8321", SOC_MTK_MT8321, soc, 28)
SOC_EQ(tmp, "MT8382", "MT8382", SOC_MTK_MT8382, soc, 28)
SOC_EQ(tmp, "MT8581", "MT8581", SOC_MTK_MT8581, soc, 28)
SOC_EQ(tmp, "MT8735", "MT8735", SOC_MTK_MT8735, soc, 28)
SOC_EQ(tmp, "MT8765B", "MT8765B", SOC_MTK_MT8765B, soc, 28)
SOC_EQ(tmp, "MT8783", "MT8783", SOC_MTK_MT8783, soc, 28)
SOC_END
}
/*
* APQ: Application Processor Qualcomm
* MSM: Mobile Station Modem
* In a APQXXXX or MSMXXXX, the second digit represents:
* *------------------*
* | Value | Meaning |
* *------------------*
* | 0 | No modem |
* | 2 | HPSA+ |
* | 6 | CDMA |
* | 9 | LTE |
* *------------------*
* Ref: https://www.tomshardware.com/reviews/snapdragon-801-performance-xperia-z2,3777-2.html
* TWO-HEADED SNAPDRAGON TAKES FLIGHT By Linley Gwennap
*
* If Qualcomm official website reports the SoC name without the initial two or three SKU name,
* we assume APQ if second number is 0, or MSM if second number is different than 0
*
* All SoC names here have been retrieved from official Qualcomm resources. However, Linux kernel
* and Android may report the SoC with slightly different. Therefore, this function needs some
* rework (e.g, debug with http://specdevice.com/unmoderated.php?lang=en)
*/
bool match_qualcomm(char* soc_name, struct system_on_chip* soc) {
char* tmp;
char* soc_name_upper = toupperstr(soc_name);
if((tmp = strstr(soc_name_upper, "MSM")) != NULL);
else if((tmp = strstr(soc_name_upper, "SDM")) != NULL);
else if((tmp = strstr(soc_name_upper, "APQ")) != NULL);
else if((tmp = strstr(soc_name_upper, "SM")) != NULL);
else if((tmp = strstr(soc_name_upper, "QM")) != NULL);
else if((tmp = strstr(soc_name_upper, "QSD")) != NULL);
else return false;
SOC_START
// Snapdragon S1 //
SOC_EQ(tmp, "QSD8650", "S1", SOC_SNAPD_QSD8650, soc, 65)
SOC_EQ(tmp, "QSD8250", "S1", SOC_SNAPD_QSD8250, soc, 65)
SOC_EQ(tmp, "MSM7627", "S1", SOC_SNAPD_MSM7627, soc, 65)
SOC_EQ(tmp, "MSM7227", "S1", SOC_SNAPD_MSM7227, soc, 65)
SOC_EQ(tmp, "MSM7627A", "S1", SOC_SNAPD_MSM7627A, soc, 45)
SOC_EQ(tmp, "MSM7227A", "S1", SOC_SNAPD_MSM7227A, soc, 45)
SOC_EQ(tmp, "MSM7625", "S1", SOC_SNAPD_MSM7625, soc, 65)
SOC_EQ(tmp, "MSM7225", "S1", SOC_SNAPD_MSM7225, soc, 65)
SOC_EQ(tmp, "MSM7625A", "S1", SOC_SNAPD_MSM7625A, soc, 45)
SOC_EQ(tmp, "MSM7225A", "S1", SOC_SNAPD_MSM7225A, soc, 45)
// Snapdragon S2 //
SOC_EQ(tmp, "MSM8655", "S2", SOC_SNAPD_MSM8655, soc, 45)
SOC_EQ(tmp, "MSM8255", "S2", SOC_SNAPD_MSM8255, soc, 45)
SOC_EQ(tmp, "APQ8055", "S2", SOC_SNAPD_APQ8055, soc, 45)
SOC_EQ(tmp, "MSM7630", "S2", SOC_SNAPD_MSM7630, soc, 45)
SOC_EQ(tmp, "MSM7230", "S2", SOC_SNAPD_MSM7230, soc, 45)
// Snapdragon S3 //
SOC_EQ(tmp, "MSM8660", "S3", SOC_SNAPD_MSM8660, soc, 45)
SOC_EQ(tmp, "MSM8260", "S3", SOC_SNAPD_MSM8260, soc, 45)
SOC_EQ(tmp, "APQ8060", "S3", SOC_SNAPD_APQ8060, soc, 45)
// Snapdragon S4 //
SOC_EQ(tmp, "MSM8225", "S4 Play", SOC_SNAPD_MSM8225, soc, 45)
SOC_EQ(tmp, "MSM8625", "S4 Play", SOC_SNAPD_MSM8625, soc, 45)
SOC_EQ(tmp, "APQ8060A", "S4 Plus", SOC_SNAPD_APQ8060A, soc, 28)
SOC_EQ(tmp, "MSM8960", "S4 Plus", SOC_SNAPD_MSM8960, soc, 28)
SOC_EQ(tmp, "MSM8260A", "S4 Plus", SOC_SNAPD_MSM8260A, soc, 28)
SOC_EQ(tmp, "MSM8627", "S4 Plus", SOC_SNAPD_MSM8627, soc, 28)
SOC_EQ(tmp, "MSM8227", "S4 Plus", SOC_SNAPD_MSM8227, soc, 28)
SOC_EQ(tmp, "APQ8064", "S4 Pro", SOC_SNAPD_APQ8064, soc, 28)
SOC_EQ(tmp, "MSM8960T", "S4 Pro", SOC_SNAPD_MSM8960T, soc, 28)
// Snapdragon 2XX //
SOC_EQ(tmp, "MSM8110", "200", SOC_SNAPD_MSM8110, soc, 28)
SOC_EQ(tmp, "MSM8210", "200", SOC_SNAPD_MSM8210, soc, 28)
SOC_EQ(tmp, "MSM8610", "200", SOC_SNAPD_MSM8610, soc, 28)
SOC_EQ(tmp, "MSM8112", "200", SOC_SNAPD_MSM8112, soc, 28)
SOC_EQ(tmp, "MSM8212", "200", SOC_SNAPD_MSM8212, soc, 28)
SOC_EQ(tmp, "MSM8612", "200", SOC_SNAPD_MSM8612, soc, 28)
SOC_EQ(tmp, "MSM8225Q", "200", SOC_SNAPD_MSM8225Q, soc, 45)
SOC_EQ(tmp, "MSM8625Q", "200", SOC_SNAPD_MSM8625Q, soc, 45)
SOC_EQ(tmp, "MSM8208", "208", SOC_SNAPD_MSM8208, soc, 28)
SOC_EQ(tmp, "MSM8905", "205", SOC_SNAPD_MSM8905, soc, 28)
SOC_EQ(tmp, "MSM8909", "210 / 212", SOC_SNAPD_MSM8909, soc, 28) // In the future, we can differenciate them using frequency
SOC_EQ(tmp, "QM215", "215", SOC_SNAPD_QM215, soc, 28)
// Snapdragon 4XX //
SOC_EQ(tmp, "APQ8028", "400", SOC_SNAPD_APQ8028, soc, 28)
SOC_EQ(tmp, "MSM8228", "400", SOC_SNAPD_MSM8228, soc, 28)
SOC_EQ(tmp, "MSM8628", "400", SOC_SNAPD_MSM8628, soc, 28)
SOC_EQ(tmp, "MSM8928", "400", SOC_SNAPD_MSM8928, soc, 28)
SOC_EQ(tmp, "MSM8926", "400", SOC_SNAPD_MSM8926, soc, 28)
SOC_EQ(tmp, "APQ8030AB", "400", SOC_SNAPD_APQ8030AB, soc, 28)
SOC_EQ(tmp, "MSM8226", "400", SOC_SNAPD_MSM8226, soc, 28)
SOC_EQ(tmp, "MSM8230AB", "400", SOC_SNAPD_MSM8230AB, soc, 28)
SOC_EQ(tmp, "MSM8626", "400", SOC_SNAPD_MSM8626, soc, 28)
SOC_EQ(tmp, "MSM8630", "400", SOC_SNAPD_MSM8630, soc, 28)
SOC_EQ(tmp, "MSM8630AB", "400", SOC_SNAPD_MSM8630AB, soc, 28)
SOC_EQ(tmp, "MSM8930", "400", SOC_SNAPD_MSM8930, soc, 28)
SOC_EQ(tmp, "MSM8930AB", "400", SOC_SNAPD_MSM8930AB, soc, 28)
SOC_EQ(tmp, "MSM8916", "410 / 412", SOC_SNAPD_MSM8916, soc, 28)
SOC_EQ(tmp, "MSM8929", "415", SOC_SNAPD_MSM8929, soc, 28)
SOC_EQ(tmp, "MSM8917", "425", SOC_SNAPD_MSM8917, soc, 28)
SOC_EQ(tmp, "MSM8920", "427", SOC_SNAPD_MSM8920, soc, 28)
SOC_EQ(tmp, "SDM429", "429", SOC_SNAPD_SDM429, soc, 12)
SOC_EQ(tmp, "MSM8937", "430", SOC_SNAPD_MSM8937, soc, 28)
SOC_EQ(tmp, "MSM8940", "435", SOC_SNAPD_MSM8940, soc, 28)
SOC_EQ(tmp, "SDM439", "439", SOC_SNAPD_SDM439, soc, 12)
SOC_EQ(tmp, "SDM450", "450", SOC_SNAPD_SDM450, soc, 14)
SOC_EQ(tmp, "SM4250-AA", "460", SOC_SNAPD_SM4250_AA, soc, 11)
// Snapdragon 6XX //
SOC_EQ(tmp, "APQ8064T", "600", SOC_SNAPD_APQ8064T, soc, 28)
SOC_EQ(tmp, "APQ8064M", "600", SOC_SNAPD_APQ8064M, soc, 28)
SOC_EQ(tmp, "MSM8936", "610", SOC_SNAPD_MSM8936, soc, 28)
SOC_EQ(tmp, "MSM8939", "615 / 616", SOC_SNAPD_MSM8939, soc, 28)
SOC_EQ(tmp, "MSM8952", "617", SOC_SNAPD_MSM8952, soc, 28)
SOC_EQ(tmp, "MSM8953", "625", SOC_SNAPD_MSM8953, soc, 14)
SOC_EQ(tmp, "MSM8953 PRO", "626", SOC_SNAPD_MSM8953_PRO, soc, 14)
SOC_EQ(tmp, "SDM630", "630", SOC_SNAPD_SDM630, soc, 14)
SOC_EQ(tmp, "SDM632", "632", SOC_SNAPD_SDM632, soc, 14)
SOC_EQ(tmp, "SDM636", "636", SOC_SNAPD_SDM636, soc, 14)
SOC_EQ(tmp, "MSM8956", "650", SOC_SNAPD_MSM8956, soc, 28)
SOC_EQ(tmp, "MSM8976", "652", SOC_SNAPD_MSM8976, soc, 28)
SOC_EQ(tmp, "MSM8976 PRO", "653", SOC_SNAPD_MSM8976_PRO, soc, 28)
SOC_EQ(tmp, "SDM660", "660", SOC_SNAPD_SDM660, soc, 14)
SOC_EQ(tmp, "SM6115", "662", SOC_SNAPD_SM6115, soc, 11)
SOC_EQ(tmp, "SM6125", "665", SOC_SNAPD_SM6125, soc, 11)
SOC_EQ(tmp, "SDM670", "670", SOC_SNAPD_SDM670, soc, 10)
SOC_EQ(tmp, "SM6150", "675", SOC_SNAPD_SM6150, soc, 11)
SOC_EQ(tmp, "SM6350", "690", SOC_SNAPD_SM6350, soc, 8)
// Snapdragon 7XX //
SOC_EQ(tmp, "SDM710", "710", SOC_SNAPD_SDM710, soc, 10)
SOC_EQ(tmp, "SDM712", "712", SOC_SNAPD_SDM712, soc, 10)
SOC_EQ(tmp, "SM7125", "720G", SOC_SNAPD_SM7125, soc, 8)
SOC_EQ(tmp, "SM7150-AA", "730", SOC_SNAPD_SM7150_AA, soc, 8)
SOC_EQ(tmp, "SM7150-AB", "730G", SOC_SNAPD_SM7150_AB, soc, 8)
SOC_EQ(tmp, "SM7150-AC", "732G", SOC_SNAPD_SM7150_AC, soc, 8)
SOC_EQ(tmp, "SM7225", "750G", SOC_SNAPD_SM7225, soc, 8)
SOC_EQ(tmp, "SM7250-AA", "765", SOC_SNAPD_SM7250_AA, soc, 7)
SOC_EQ(tmp, "SM7250-AB", "765G", SOC_SNAPD_SM7250_AB, soc, 7)
SOC_EQ(tmp, "SM7250-AC", "768G", SOC_SNAPD_SM7250_AC, soc, 7)
// Snapdragon 8XX //
SOC_EQ(tmp, "MSM8974AA", "800", SOC_SNAPD_MSM8974AA, soc, 28)
SOC_EQ(tmp, "MSM8974AB", "800", SOC_SNAPD_MSM8974AB, soc, 28)
SOC_EQ(tmp, "MSM8974AC", "800", SOC_SNAPD_MSM8974AC, soc, 28)
SOC_EQ(tmp, "MSM8974PRO-AB", "801", SOC_SNAPD_MSM8974PRO_AB, soc, 28)
SOC_EQ(tmp, "MSM8974PRO-AC", "801", SOC_SNAPD_MSM8974PRO_AC, soc, 28)
SOC_EQ(tmp, "APQ8084", "805", SOC_SNAPD_APQ8084, soc, 28)
SOC_EQ(tmp, "MSM8992", "808", SOC_SNAPD_MSM8992, soc, 20)
SOC_EQ(tmp, "MSM8994", "810", SOC_SNAPD_MSM8994, soc, 20)
SOC_EQ(tmp, "MSM8996", "820", SOC_SNAPD_MSM8996, soc, 14)
SOC_EQ(tmp, "MSM8996 PRO A", "821", SOC_SNAPD_MSM8996_PRO_A, soc, 14)
SOC_EQ(tmp, "MSM8998", "835", SOC_SNAPD_MSM8998, soc, 10)
SOC_EQ(tmp, "APQ8098", "835", SOC_SNAPD_APQ8098, soc, 10)
SOC_EQ(tmp, "SDM845", "845", SOC_SNAPD_SDM845, soc, 10)
SOC_EQ(tmp, "SDM850", "850", SOC_SNAPD_SDM850, soc, 10)
SOC_EQ(tmp, "SM8150", "855", SOC_SNAPD_SM8150, soc, 7)
SOC_EQ(tmp, "SM8150-AC", "855+", SOC_SNAPD_SM8150_AC, soc, 7)
SOC_EQ(tmp, "SM8250", "865", SOC_SNAPD_SM8250, soc, 7)
SOC_EQ(tmp, "SM8250-AB", "865+", SOC_SNAPD_SM8250_AB, soc, 7)
SOC_EQ(tmp, "SM8350", "888", SOC_SNAPD_SM8350, soc, 5)
SOC_END
}
bool match_special(char* soc_name, struct system_on_chip* soc) {
char* tmp;
// Xiaomi hides Redmi Note 8/8T under "Qualcomm Technologies, Inc TRINKET"
if((tmp = strstr(soc_name, "TRINKET")) != NULL) {
fill_soc(soc, "665", SOC_SNAPD_SM6125, 11);
return true;
}
// Snapdragon 730 reported as "Qualcomm Technologies, Inc. SDMMAGPIE"
if((tmp = strstr(soc_name, "SDMMAGPIE")) != NULL) {
fill_soc(soc, "730", SOC_SNAPD_SM7150_AA, 8);
return true;
}
return false;
}
struct system_on_chip* parse_soc_from_string(struct system_on_chip* soc) {
char* raw_name = soc->raw_name;
if(match_special(raw_name, soc))
return soc;
if (match_qualcomm(raw_name, soc))
return soc;
if(match_mediatek(raw_name, soc))
return soc;
if(match_exynos(raw_name, soc))
return soc;
if(match_hisilicon(raw_name, soc))
return soc;
match_broadcom(raw_name, soc);
return soc;
}
#ifdef __ANDROID__
#include <sys/system_properties.h>
static inline int android_property_get(const char* key, char* value) {
return __system_property_get(key, value);
}
struct system_on_chip* guess_soc_from_android(struct system_on_chip* soc) {
char tmp[100];
int property_len = 0;
property_len = android_property_get("ro.mediatek.platform", (char *) &tmp);
if(property_len > 0) {
soc->raw_name = malloc(sizeof(char) * (property_len + 1));
strncpy(soc->raw_name, tmp, property_len + 1);
soc->raw_name[property_len] = '\0';
soc->soc_vendor = SOC_VENDOR_UNKNOWN;
return parse_soc_from_string(soc);
}
property_len = android_property_get("ro.product.board", (char *) &tmp);
if(property_len > 0) {
soc->raw_name = malloc(sizeof(char) * (property_len + 1));
strncpy(soc->raw_name, tmp, property_len + 1);
soc->raw_name[property_len] = '\0';
soc->soc_vendor = SOC_VENDOR_UNKNOWN;
return parse_soc_from_string(soc);
}
return soc;
}
#endif
struct system_on_chip* guess_soc_from_cpuinfo(struct system_on_chip* soc) {
char* tmp = get_hardware_from_cpuinfo(&strlen);
if(tmp != NULL) {
soc->raw_name = tmp;
return parse_soc_from_string(soc);
}
return soc;
}
struct system_on_chip* get_soc() {
struct system_on_chip* soc = malloc(sizeof(struct system_on_chip));
soc->raw_name = NULL;
soc->soc_vendor = SOC_VENDOR_UNKNOWN;
soc->process = UNKNOWN;
soc = guess_soc_from_cpuinfo(soc);
if(soc->soc_vendor == SOC_VENDOR_UNKNOWN) {
if(soc->raw_name != NULL)
printWarn("SoC detection failed using /proc/cpuinfo: Found '%s' string", soc->raw_name);
else
printWarn("SoC detection failed using /proc/cpuinfo: No string found");
#ifdef __ANDROID__
soc = guess_soc_from_android(soc);
if(soc->raw_name == NULL)
printWarn("SoC detection failed using Android: No string found");
else if(soc->soc_vendor == SOC_VENDOR_UNKNOWN)
printWarn("SoC detection failed using Android: Found '%s' string", soc->raw_name);
#endif
}
if(soc->raw_name == NULL) {
soc->raw_name = malloc(sizeof(char) * (strlen(STRING_UNKNOWN)+1));
snprintf(soc->raw_name, strlen(STRING_UNKNOWN)+1, STRING_UNKNOWN);
}
return soc;
}
char* get_soc_name(struct system_on_chip* soc) {
if(soc->soc_vendor == SOC_VENDOR_UNKNOWN)
return soc->raw_name;
return soc->soc_name;
}
VENDOR get_soc_vendor(struct system_on_chip* soc) {
return soc->soc_vendor;
}
char* get_str_process(struct system_on_chip* soc) {
char* str;
if(soc->process == UNKNOWN) {
str = malloc(sizeof(char) * (strlen(STRING_UNKNOWN)+1));
snprintf(str, strlen(STRING_UNKNOWN)+1, STRING_UNKNOWN);
}
else {
str = malloc(sizeof(char) * 5);
memset(str, 0, sizeof(char) * 5);
snprintf(str, 5, "%dnm", soc->process);
}
return str;
}

31
src/arm/soc.h Normal file
View File

@@ -0,0 +1,31 @@
#ifndef __SOC__
#define __SOC__
#include "../common/cpu.h"
#include <stdint.h>
typedef int32_t SOC;
enum {
SOC_VENDOR_UNKNOWN,
SOC_VENDOR_SNAPDRAGON,
SOC_VENDOR_MEDIATEK,
SOC_VENDOR_EXYNOS,
SOC_VENDOR_KIRIN,
SOC_VENDOR_BROADCOM
};
struct system_on_chip {
SOC soc_model;
VENDOR soc_vendor;
int32_t process;
char* soc_name;
char* raw_name;
};
struct system_on_chip* get_soc();
char* get_soc_name(struct system_on_chip* soc);
VENDOR get_soc_vendor(struct system_on_chip* soc);
char* get_str_process(struct system_on_chip* soc);
#endif

264
src/arm/socs.h Normal file
View File

@@ -0,0 +1,264 @@
#ifndef __SOCS__
#define __SOCS__
#include "soc.h"
// List of supported SOCs
enum {
// Broadcom //
SOC_BCM_2835,
SOC_BCM_2836,
SOC_BCM_2837,
SOC_BCM_2837B0,
SOC_BCM_2711,
SOC_BCM_21553,
SOC_BCM_21553T,
SOC_BCM_21663,
SOC_BCM_21664,
SOC_BCM_28155,
SOC_BCM_23550,
SOC_BCM_28145,
SOC_BCM_2157,
SOC_BCM_21654,
// Hisilicon //
SOC_HISILICON_3620,
SOC_HISILICON_3630,
SOC_HISILICON_3650,
SOC_HISILICON_3660,
SOC_HISILICON_3670,
SOC_HISILICON_3680,
SOC_HISILICON_3690,
// Exynos //
SOC_EXYNOS_3475,
SOC_EXYNOS_4210,
SOC_EXYNOS_4212,
SOC_EXYNOS_4412,
SOC_EXYNOS_5250,
SOC_EXYNOS_5410,
SOC_EXYNOS_5420,
SOC_EXYNOS_5422,
SOC_EXYNOS_5430,
SOC_EXYNOS_5433,
SOC_EXYNOS_5260,
SOC_EXYNOS_7270,
SOC_EXYNOS_7420,
SOC_EXYNOS_7570,
SOC_EXYNOS_7870,
SOC_EXYNOS_7872,
SOC_EXYNOS_7880,
SOC_EXYNOS_7884,
SOC_EXYNOS_7885,
SOC_EXYNOS_7904,
SOC_EXYNOS_8890,
SOC_EXYNOS_8895,
SOC_EXYNOS_9110,
SOC_EXYNOS_9609,
SOC_EXYNOS_9610,
SOC_EXYNOS_9611,
SOC_EXYNOS_9810,
SOC_EXYNOS_9820,
SOC_EXYNOS_9825,
SOC_EXYNOS_1080,
SOC_EXYNOS_990,
SOC_EXYNOS_980,
SOC_EXYNOS_880,
// Mediatek //
SOC_MTK_MT6889,
SOC_MTK_MT6885Z,
SOC_MTK_MT6853,
SOC_MTK_MT6873,
SOC_MTK_MT6875,
SOC_MTK_MT6761D,
SOC_MTK_MT6761,
SOC_MTK_MT6762D,
SOC_MTK_MT6755,
SOC_MTK_MT6755M,
SOC_MTK_MT6755T,
SOC_MTK_MT6757,
SOC_MTK_MT6762,
SOC_MTK_MT6763V,
SOC_MTK_MT6763T,
SOC_MTK_MT6757CD,
SOC_MTK_MT6758,
SOC_MTK_MT6765,
SOC_MTK_MT6771,
SOC_MTK_MT6768,
SOC_MTK_MT6771T,
SOC_MTK_MT6771V,
SOC_MTK_MT6779,
SOC_MTK_MT6795,
SOC_MTK_MT6795T,
SOC_MTK_MT6797,
SOC_MTK_MT6797M,
SOC_MTK_MT6797D,
SOC_MTK_MT6797T,
SOC_MTK_MT6797X,
SOC_MTK_MT6799,
SOC_MTK_MT6515,
SOC_MTK_MT6516,
SOC_MTK_MT6517,
SOC_MTK_MT6572,
SOC_MTK_MT6572M,
SOC_MTK_MT6573,
SOC_MTK_MT6575,
SOC_MTK_MT6577,
SOC_MTK_MT6577T,
SOC_MTK_MT6580,
SOC_MTK_MT6582,
SOC_MTK_MT6582M,
SOC_MTK_MT6589,
SOC_MTK_MT6589T,
SOC_MTK_MT6592,
SOC_MTK_MT6595,
SOC_MTK_MT6732,
SOC_MTK_MT6735,
SOC_MTK_MT6735M,
SOC_MTK_MT6735P,
SOC_MTK_MT6737,
SOC_MTK_MT6737M,
SOC_MTK_MT6737T,
SOC_MTK_MT6739,
SOC_MTK_MT6750,
SOC_MTK_MT6750S,
SOC_MTK_MT6750T,
SOC_MTK_MT6752,
SOC_MTK_MT6753,
SOC_MTK_MT6850,
SOC_MTK_MT8121,
SOC_MTK_MT8125,
SOC_MTK_MT8127,
SOC_MTK_MT8135,
SOC_MTK_MT8163A,
SOC_MTK_MT8163B,
SOC_MTK_MT8167B,
SOC_MTK_MT8173,
SOC_MTK_MT8176,
SOC_MTK_MT8321,
SOC_MTK_MT8382,
SOC_MTK_MT8581,
SOC_MTK_MT8735,
SOC_MTK_MT8765B,
SOC_MTK_MT8783,
// Snapdragon //
SOC_SNAPD_QSD8650,
SOC_SNAPD_QSD8250,
SOC_SNAPD_MSM7627,
SOC_SNAPD_MSM7227,
SOC_SNAPD_MSM7627A,
SOC_SNAPD_MSM7227A,
SOC_SNAPD_MSM7625,
SOC_SNAPD_MSM7225,
SOC_SNAPD_MSM7625A,
SOC_SNAPD_MSM7225A,
SOC_SNAPD_MSM8655,
SOC_SNAPD_MSM8255,
SOC_SNAPD_APQ8055,
SOC_SNAPD_MSM7630,
SOC_SNAPD_MSM7230,
SOC_SNAPD_MSM8660,
SOC_SNAPD_MSM8260,
SOC_SNAPD_APQ8060,
SOC_SNAPD_MSM8225,
SOC_SNAPD_MSM8625,
SOC_SNAPD_APQ8060A,
SOC_SNAPD_MSM8960,
SOC_SNAPD_MSM8260A,
SOC_SNAPD_MSM8627,
SOC_SNAPD_MSM8227,
SOC_SNAPD_APQ8064,
SOC_SNAPD_MSM8960T,
SOC_SNAPD_MSM8110,
SOC_SNAPD_MSM8210,
SOC_SNAPD_MSM8610,
SOC_SNAPD_MSM8112,
SOC_SNAPD_MSM8212,
SOC_SNAPD_MSM8612,
SOC_SNAPD_MSM8225Q,
SOC_SNAPD_MSM8625Q,
SOC_SNAPD_MSM8208,
SOC_SNAPD_MSM8905,
SOC_SNAPD_MSM8909,
SOC_SNAPD_QM215,
SOC_SNAPD_APQ8028,
SOC_SNAPD_MSM8228,
SOC_SNAPD_MSM8628,
SOC_SNAPD_MSM8928,
SOC_SNAPD_MSM8926,
SOC_SNAPD_APQ8030AB,
SOC_SNAPD_MSM8226,
SOC_SNAPD_MSM8230AB,
SOC_SNAPD_MSM8626,
SOC_SNAPD_MSM8630,
SOC_SNAPD_MSM8630AB,
SOC_SNAPD_MSM8930,
SOC_SNAPD_MSM8930AB,
SOC_SNAPD_MSM8916,
SOC_SNAPD_MSM8929,
SOC_SNAPD_MSM8917,
SOC_SNAPD_MSM8920,
SOC_SNAPD_SDM429,
SOC_SNAPD_MSM8937,
SOC_SNAPD_MSM8940,
SOC_SNAPD_SDM439,
SOC_SNAPD_SDM450,
SOC_SNAPD_SM4250_AA,
SOC_SNAPD_APQ8064T,
SOC_SNAPD_APQ8064M,
SOC_SNAPD_MSM8936,
SOC_SNAPD_MSM8939,
SOC_SNAPD_MSM8952,
SOC_SNAPD_MSM8953,
SOC_SNAPD_MSM8953_PRO,
SOC_SNAPD_SDM630,
SOC_SNAPD_SDM632,
SOC_SNAPD_SDM636,
SOC_SNAPD_MSM8956,
SOC_SNAPD_MSM8976,
SOC_SNAPD_MSM8976_PRO,
SOC_SNAPD_SDM660,
SOC_SNAPD_SM6115,
SOC_SNAPD_SM6125,
SOC_SNAPD_SDM670,
SOC_SNAPD_SM6150,
SOC_SNAPD_SM6350,
SOC_SNAPD_SDM710,
SOC_SNAPD_SDM712,
SOC_SNAPD_SM7125,
SOC_SNAPD_SM7150_AA,
SOC_SNAPD_SM7150_AB,
SOC_SNAPD_SM7150_AC,
SOC_SNAPD_SM7225,
SOC_SNAPD_SM7250_AA,
SOC_SNAPD_SM7250_AB,
SOC_SNAPD_SM7250_AC,
SOC_SNAPD_MSM8974AA,
SOC_SNAPD_MSM8974AB,
SOC_SNAPD_MSM8974AC,
SOC_SNAPD_MSM8974PRO_AB,
SOC_SNAPD_MSM8974PRO_AC,
SOC_SNAPD_APQ8084,
SOC_SNAPD_MSM8992,
SOC_SNAPD_MSM8994,
SOC_SNAPD_MSM8996,
SOC_SNAPD_MSM8996_PRO_A,
SOC_SNAPD_MSM8998,
SOC_SNAPD_APQ8098,
SOC_SNAPD_SDM845,
SOC_SNAPD_SDM850,
SOC_SNAPD_SM8150,
SOC_SNAPD_SM8150_AC,
SOC_SNAPD_SM8250,
SOC_SNAPD_SM8250_AB,
SOC_SNAPD_SM8350,
};
inline static VENDOR get_soc_vendor_from_soc(SOC soc) {
if(soc >= SOC_BCM_2835 && soc <= SOC_BCM_21654) return SOC_VENDOR_BROADCOM;
else if(soc >= SOC_HISILICON_3620 && soc <= SOC_HISILICON_3690) return SOC_VENDOR_KIRIN;
else if(soc >= SOC_EXYNOS_3475 && soc <= SOC_EXYNOS_880) return SOC_VENDOR_EXYNOS;
else if(soc >= SOC_MTK_MT6889 && soc <= SOC_MTK_MT8783) return SOC_VENDOR_MEDIATEK;
else if(soc >= SOC_SNAPD_QSD8650 && soc <= SOC_SNAPD_SM8350) return SOC_VENDOR_SNAPDRAGON;
return SOC_VENDOR_UNKNOWN;
}
#endif

23
src/arm/socs_generation.sh Executable file
View File

@@ -0,0 +1,23 @@
#!/bin/bash -u
SOC_LIST="$(grep SOC_EQ soc.c | grep -v '//SOC_EQ' | grep -v 'define' | cut -d',' -f2 | sed 's/"//')"
IFS=$'"'
for soc in $SOC_LIST
do
# CLEAN
soc=$(echo $soc | tr -d '\n')
soc="${soc:1}"
# REPLACE
soc=$(echo $soc | sed "s/BCM/BCM_/g")
soc=$(echo $soc | sed "s/universal/EXYNOS_/g")
soc=$(echo $soc | sed "s/Hi/HISILICON_/g")
soc=$(echo $soc | sed "s/^MSM/SNAPD_MSM/g" | sed "s/SDM/SNAPD_SDM/g" | sed "s/APQ/SNAPD_APQ/g" | sed "s/^SM/SNAPD_SM/g" | sed "s/QM/SNAPD_QM/g" | sed "s/QSD/SNAPD_QSD/g")
soc=$(echo $soc | sed "s/MT/MTK_MT/g")
soc=$(echo $soc | sed "s/-/_/g" | sed "s/ /_/g")
echo ' SOC_'"$soc"','
done
unset IFS

282
src/arm/uarch.c Normal file
View File

@@ -0,0 +1,282 @@
#include <stdbool.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <inttypes.h>
#include "uarch.h"
#include "../common/global.h"
#define STRING_UNKNOWN "Unknown"
// Data not available
#define NA -1
typedef uint32_t MICROARCH;
typedef uint32_t ISA;
struct uarch {
MICROARCH uarch;
ISA isa;
char* uarch_str;
char* isa_str;
// int32_t process; process depends on SoC
};
enum {
ISA_ARMv7_A,
ISA_ARMv8_A,
ISA_ARMv8_A_AArch32,
ISA_ARMv8_1_A,
ISA_ARMv8_2_A,
ISA_ARMv8_3_A,
};
enum {
UARCH_UNKNOWN,
// ARM
UARCH_ARM7,
UARCH_ARM9,
UARCH_ARM11, // ARM 1136, ARM 1156, ARM 1176, or ARM 11MPCore.
UARCH_CORTEX_A5,
UARCH_CORTEX_A7,
UARCH_CORTEX_A8,
UARCH_CORTEX_A9,
UARCH_CORTEX_A12,
UARCH_CORTEX_A15,
UARCH_CORTEX_A17,
UARCH_CORTEX_A32,
UARCH_CORTEX_A35,
UARCH_CORTEX_A53,
UARCH_CORTEX_A55r0, // ARM Cortex-A55 revision 0 (restricted dual-issue capabilities compared to revision 1+).
UARCH_CORTEX_A55,
UARCH_CORTEX_A57,
UARCH_CORTEX_A65,
UARCH_CORTEX_A72,
UARCH_CORTEX_A73,
UARCH_CORTEX_A75,
UARCH_CORTEX_A76,
UARCH_CORTEX_A77,
UARCH_CORTEX_A78,
UARCH_NEOVERSE_N1,
UARCH_NEOVERSE_E1,
UARCH_SCORPION,
UARCH_KRAIT,
UARCH_KYRO,
UARCH_FALKOR,
UARCH_SAPHIRA,
UARCH_DENVER,
UARCH_DENVER2,
UARCH_CARMEL,
// SAMSUNG
UARCH_EXYNOS_M1, // Samsung Exynos M1 (Exynos 8890 big cores)
UARCH_EXYNOS_M2, // Samsung Exynos M2 (Exynos 8895 big cores)
UARCH_EXYNOS_M3, // Samsung Exynos M3 (Exynos 9810 big cores)
UARCH_EXYNOS_M4, // Samsung Exynos M4 (Exynos 9820 big cores)
UARCH_EXYNOS_M5, // Samsung Exynos M5 (Exynos 9830 big cores)
// APPLE
UARCH_SWIFT, // Apple A6 and A6X processors.
UARCH_CYCLONE, // Apple A7 processor.
UARCH_TYPHOON, // Apple A8 and A8X processor
UARCH_TWISTER, // Apple A9 and A9X processor.
UARCH_HURRICANE, // Apple A10 and A10X processor.
UARCH_MONSOON, // Apple A11 processor (big cores).
UARCH_MISTRAL, // Apple A11 processor (little cores).
UARCH_VORTEX, // Apple A12 processor (big cores).
UARCH_TEMPEST, // Apple A12 processor (big cores).
UARCH_LIGHTNING, // Apple A13 processor (big cores).
UARCH_THUNDER, // Apple A13 processor (little cores).
// CAVIUM
UARCH_THUNDERX, // Cavium ThunderX
UARCH_THUNDERX2, // Cavium ThunderX2 (originally Broadcom Vulkan).
// MARVELL
UARCH_PJ4,
UARCH_BRAHMA_B15,
UARCH_BRAHMA_B53,
UARCH_XGENE, // Applied Micro X-Gene.
UARCH_TAISHAN_V110 // HiSilicon TaiShan v110 (Huawei Kunpeng 920 series processors).
};
static const ISA isas_uarch[] = {
[UARCH_CORTEX_A5] = ISA_ARMv7_A,
[UARCH_CORTEX_A7] = ISA_ARMv7_A,
[UARCH_CORTEX_A8] = ISA_ARMv7_A,
[UARCH_CORTEX_A9] = ISA_ARMv7_A,
[UARCH_CORTEX_A12] = ISA_ARMv7_A,
[UARCH_CORTEX_A15] = ISA_ARMv7_A,
[UARCH_CORTEX_A17] = ISA_ARMv7_A,
[UARCH_CORTEX_A32] = ISA_ARMv8_A_AArch32,
[UARCH_CORTEX_A35] = ISA_ARMv8_A,
[UARCH_CORTEX_A53] = ISA_ARMv8_A,
[UARCH_CORTEX_A55r0] = ISA_ARMv8_2_A,
[UARCH_CORTEX_A55] = ISA_ARMv8_2_A,
[UARCH_CORTEX_A57] = ISA_ARMv8_A,
[UARCH_CORTEX_A65] = ISA_ARMv8_2_A,
[UARCH_CORTEX_A72] = ISA_ARMv8_A,
[UARCH_CORTEX_A73] = ISA_ARMv8_A,
[UARCH_CORTEX_A75] = ISA_ARMv8_2_A,
[UARCH_CORTEX_A76] = ISA_ARMv8_2_A,
[UARCH_CORTEX_A77] = ISA_ARMv8_2_A,
[UARCH_CORTEX_A78] = ISA_ARMv8_2_A,
[UARCH_NEOVERSE_N1] = ISA_ARMv8_2_A,
[UARCH_NEOVERSE_E1] = ISA_ARMv8_2_A,
[UARCH_BRAHMA_B15] = ISA_ARMv7_A, // Same as Cortex-A15
[UARCH_BRAHMA_B53] = ISA_ARMv8_A, // Same as Cortex-A53
[UARCH_THUNDERX] = ISA_ARMv8_A,
[UARCH_THUNDERX2] = ISA_ARMv8_1_A,
[UARCH_TAISHAN_V110] = ISA_ARMv8_2_A,
[UARCH_DENVER] = ISA_ARMv8_A,
[UARCH_DENVER2] = ISA_ARMv8_A,
[UARCH_CARMEL] = ISA_ARMv8_A,
[UARCH_XGENE] = ISA_ARMv8_A, // https://en.wikichip.org/wiki/apm/x-gene
[UARCH_SCORPION] = ISA_ARMv7_A, // https://www.geektopia.es/es/product/qualcomm/snapdragon-s3-apq8060/
[UARCH_KRAIT] = ISA_ARMv7_A,
[UARCH_KYRO] = ISA_ARMv8_A,
[UARCH_FALKOR] = ISA_ARMv8_A,
[UARCH_SAPHIRA] = ISA_ARMv8_3_A,
[UARCH_EXYNOS_M1] = ISA_ARMv8_A,
[UARCH_EXYNOS_M2] = ISA_ARMv8_A,
[UARCH_EXYNOS_M3] = ISA_ARMv8_A,
[UARCH_EXYNOS_M4] = ISA_ARMv8_2_A,
[UARCH_EXYNOS_M5] = ISA_ARMv8_2_A,
[UARCH_PJ4] = ISA_ARMv7_A,
};
static char* isas_string[] = {
[ISA_ARMv7_A] = "ARMv7",
[ISA_ARMv8_A] = "ARMv8",
[ISA_ARMv8_A_AArch32] = "ARMv8 AArch32",
[ISA_ARMv8_1_A] = "ARMv8.1",
[ISA_ARMv8_2_A] = "ARMv8.2",
[ISA_ARMv8_3_A] = "ARMv8.3",
};
#define UARCH_START if (false) {}
#define CHECK_UARCH(arch, cpu, im_, p_, v_, r_, str, uarch, vendor) \
else if (im_ == im && p_ == p && (v_ == NA || v_ == v) && (r_ == NA || r_ == r)) fill_uarch(arch, cpu, str, uarch, vendor);
#define UARCH_END else { printBug("Unknown microarchitecture detected: IM=0x%.8X P=0x%.8X V=0x%.8X R=0x%.8X", im, p, v, r); fill_uarch(arch, cpu, "Unknown", UARCH_UNKNOWN, CPU_VENDOR_UNKNOWN); }
void fill_uarch(struct uarch* arch, struct cpuInfo* cpu, char* str, MICROARCH u, VENDOR vendor) {
arch->uarch = u;
arch->isa = isas_uarch[arch->uarch];
cpu->cpu_vendor = vendor;
arch->uarch_str = malloc(sizeof(char) * (strlen(str)+1));
strcpy(arch->uarch_str, str);
arch->isa_str = malloc(sizeof(char) * (strlen(isas_string[arch->isa])+1));
strcpy(arch->isa_str, isas_string[arch->isa]);
}
/*
* Codes are based on pytorch/cpuinfo, more precisely:
* - https://github.com/pytorch/cpuinfo/blob/master/src/arm/uarch.c
* Other sources:
* - https://elixir.bootlin.com/linux/latest/source/arch/arm64/include/asm/cputype.h
* - https://elixir.bootlin.com/linux/latest/source/arch/arm/include/asm/cputype.h
*/
struct uarch* get_uarch_from_midr(uint32_t midr, struct cpuInfo* cpu) {
struct uarch* arch = malloc(sizeof(struct uarch));
uint32_t im = midr_get_implementer(midr);
uint32_t p = midr_get_part(midr);
uint32_t v = midr_get_variant(midr);
uint32_t r = midr_get_revision(midr);
// ----------------------------------------------------------------------- //
// IM: Implementer //
// P: Part //
// V: Variant //
// R: Revision //
// ----------------------------------------------------------------------- //
// IM P V R //
UARCH_START
CHECK_UARCH(arch, cpu, 'A', 0xC05, NA, NA, "Cortex-A5", UARCH_CORTEX_A5, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xC07, NA, NA, "Cortex-A7", UARCH_CORTEX_A7, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xC08, NA, NA, "Cortex-A8", UARCH_CORTEX_A8, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xC09, NA, NA, "Cortex-A9", UARCH_CORTEX_A9, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xC0C, NA, NA, "Cortex-A12", UARCH_CORTEX_A12, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xC0E, NA, NA, "Cortex-A17", UARCH_CORTEX_A17, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xC0D, NA, NA, "Cortex-A12", UARCH_CORTEX_A12, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xC0F, NA, NA, "Cortex-A15", UARCH_CORTEX_A15, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD01, NA, NA, "Cortex-A32", UARCH_CORTEX_A32, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD03, NA, NA, "Cortex-A53", UARCH_CORTEX_A53, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD04, NA, NA, "Cortex-A35", UARCH_CORTEX_A35, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD05, NA, 0, "Cortex-A55", UARCH_CORTEX_A55r0, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD05, NA, NA, "Cortex-A55", UARCH_CORTEX_A55, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD06, NA, NA, "Cortex-A65", UARCH_CORTEX_A65, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD07, NA, NA, "Cortex-A57", UARCH_CORTEX_A57, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD08, NA, NA, "Cortex-A72", UARCH_CORTEX_A72, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD09, NA, NA, "Cortex-A73", UARCH_CORTEX_A73, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD0A, NA, NA, "Cortex-A75", UARCH_CORTEX_A75, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD0B, NA, NA, "Cortex-A76", UARCH_CORTEX_A76, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD0C, NA, NA, "Neoverse N1", UARCH_NEOVERSE_N1, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD0D, NA, NA, "Cortex-A77", UARCH_CORTEX_A77, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD0E, NA, NA, "Cortex-A76", UARCH_CORTEX_A76, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD41, NA, NA, "Cortex-A78", UARCH_CORTEX_A78, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'A', 0xD4A, NA, NA, "Neoverse E1", UARCH_NEOVERSE_E1, CPU_VENDOR_ARM)
CHECK_UARCH(arch, cpu, 'B', 0x00F, NA, NA, "Brahma B15", UARCH_BRAHMA_B15, CPU_VENDOR_BROADCOM)
CHECK_UARCH(arch, cpu, 'B', 0x100, NA, NA, "Brahma B53", UARCH_BRAHMA_B53, CPU_VENDOR_BROADCOM)
CHECK_UARCH(arch, cpu, 'B', 0x516, NA, NA, "ThunderX2", UARCH_THUNDERX2, CPU_VENDOR_CAVIUM)
CHECK_UARCH(arch, cpu, 'C', 0x0A0, NA, NA, "ThunderX", UARCH_THUNDERX, CPU_VENDOR_CAVIUM)
CHECK_UARCH(arch, cpu, 'C', 0x0A1, NA, NA, "ThunderX 88XX", UARCH_THUNDERX, CPU_VENDOR_CAVIUM)
CHECK_UARCH(arch, cpu, 'C', 0x0A2, NA, NA, "ThunderX 81XX", UARCH_THUNDERX, CPU_VENDOR_CAVIUM)
CHECK_UARCH(arch, cpu, 'C', 0x0A3, NA, NA, "ThunderX 81XX", UARCH_THUNDERX, CPU_VENDOR_CAVIUM)
CHECK_UARCH(arch, cpu, 'C', 0x0AF, NA, NA, "ThunderX2 99XX", UARCH_THUNDERX2, CPU_VENDOR_CAVIUM)
CHECK_UARCH(arch, cpu, 'H', 0xD01, NA, NA, "TaiShan v110", UARCH_TAISHAN_V110, CPU_VENDOR_HUAWUEI) // Kunpeng 920 series
CHECK_UARCH(arch, cpu, 'H', 0xD40, NA, NA, "Cortex-A76", UARCH_CORTEX_A76, CPU_VENDOR_ARM) // Kirin 980 Big/Medium cores -> Cortex-A76
CHECK_UARCH(arch, cpu, 'N', 0x000, NA, NA, "Denver", UARCH_DENVER, CPU_VENDOR_NVIDIA)
CHECK_UARCH(arch, cpu, 'N', 0x003, NA, NA, "Denver2", UARCH_DENVER2, CPU_VENDOR_NVIDIA)
CHECK_UARCH(arch, cpu, 'N', 0x004, NA, NA, "Carmel", UARCH_CARMEL, CPU_VENDOR_NVIDIA)
CHECK_UARCH(arch, cpu, 'P', 0x000, NA, NA, "Xgene", UARCH_XGENE, CPU_VENDOR_APM)
CHECK_UARCH(arch, cpu, 'Q', 0x00F, NA, NA, "Scorpion", UARCH_SCORPION, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x02D, NA, NA, "Scorpion", UARCH_KRAIT, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x04D, 1, 0, "Krait 200", UARCH_KRAIT, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x04D, 1, 4, "Krait 200", UARCH_KRAIT, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x04D, 2, 0, "Krait 300", UARCH_KRAIT, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x06F, 0, 1, "Krait 200", UARCH_KRAIT, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x06F, 0, 2, "Krait 200", UARCH_KRAIT, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x06F, 1, 0, "Krait 300", UARCH_KRAIT, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x06F, 2, 0, "Krait 400", UARCH_KRAIT, CPU_VENDOR_QUALCOMM) // Snapdragon 800 MSMxxxx
CHECK_UARCH(arch, cpu, 'Q', 0x06F, 2, 1, "Krait 400", UARCH_KRAIT, CPU_VENDOR_QUALCOMM) // Snapdragon 801 MSMxxxxPRO
CHECK_UARCH(arch, cpu, 'Q', 0x06F, 3, 1, "Krait 450", UARCH_KRAIT, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0x201, NA, NA, "Kryo Silver", UARCH_KYRO, CPU_VENDOR_QUALCOMM) // Qualcomm Snapdragon 821: Low-power Kryo "Silver"
CHECK_UARCH(arch, cpu, 'Q', 0x205, NA, NA, "Kryo Gold", UARCH_KYRO, CPU_VENDOR_QUALCOMM) // Qualcomm Snapdragon 820 & 821: High-performance Kryo "Gold"
CHECK_UARCH(arch, cpu, 'Q', 0x211, NA, NA, "Kryo Silver", UARCH_KYRO, CPU_VENDOR_QUALCOMM) // Qualcomm Snapdragon 820: Low-power Kryo "Silver"
CHECK_UARCH(arch, cpu, 'Q', 0x800, 10, NA, "Kryo 260 / 280 Gold", UARCH_CORTEX_A73, CPU_VENDOR_ARM) // Kryo 260 / Kryo 280 "Gold"
CHECK_UARCH(arch, cpu, 'Q', 0x801, 10, NA, "Kryo 260 / 280 Silver", UARCH_CORTEX_A53, CPU_VENDOR_ARM) // Kryo 260 / 280 "Silver"
CHECK_UARCH(arch, cpu, 'Q', 0x802, NA, NA, "Kryo 385 Gold", UARCH_CORTEX_A75, CPU_VENDOR_ARM) // High-performance Kryo 385 "Gold" -> Cortex-A75
CHECK_UARCH(arch, cpu, 'Q', 0x803, NA, NA, "Kryo 385 Silver", UARCH_CORTEX_A55r0, CPU_VENDOR_ARM) // Low-power Kryo 385 "Silver" -> Cortex-A55r0
CHECK_UARCH(arch, cpu, 'Q', 0x804, NA, NA, "Kryo 485 Gold", UARCH_CORTEX_A76, CPU_VENDOR_ARM) // High-performance Kryo 485 "Gold" / "Gold Prime" -> Cortex-A76
CHECK_UARCH(arch, cpu, 'Q', 0x805, NA, NA, "Kryo 485 Silver", UARCH_CORTEX_A55, CPU_VENDOR_ARM) // Low-performance Kryo 485 "Silver" -> Cortex-A55
CHECK_UARCH(arch, cpu, 'Q', 0xC00, NA, NA, "Falkor", UARCH_FALKOR, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'Q', 0xC01, NA, NA, "Saphira", UARCH_SAPHIRA, CPU_VENDOR_QUALCOMM)
CHECK_UARCH(arch, cpu, 'S', 0x001, 1, NA, "Exynos M1", UARCH_EXYNOS_M1, CPU_VENDOR_SAMSUNG) // Exynos 8890
CHECK_UARCH(arch, cpu, 'S', 0x001, 4, NA, "Exynos M2", UARCH_EXYNOS_M2, CPU_VENDOR_SAMSUNG) // Exynos 8895
CHECK_UARCH(arch, cpu, 'S', 0x002, 1, NA, "Exynos M3", UARCH_EXYNOS_M3, CPU_VENDOR_SAMSUNG) // Exynos 9810
CHECK_UARCH(arch, cpu, 'S', 0x003, 1, NA, "Exynos M4", UARCH_EXYNOS_M4, CPU_VENDOR_SAMSUNG) // Exynos 9820
CHECK_UARCH(arch, cpu, 'S', 0x004, 1, NA, "Exynos M5", UARCH_EXYNOS_M5, CPU_VENDOR_SAMSUNG) // Exynos 9820 (this one looks wrong at uarch.c ...)
CHECK_UARCH(arch, cpu, 'V', 0x581, NA, NA, "PJ4", UARCH_PJ4, CPU_VENDOR_MARVELL)
CHECK_UARCH(arch, cpu, 'V', 0x584, NA, NA, "PJ4B-MP", UARCH_PJ4, CPU_VENDOR_MARVELL)
UARCH_END
return arch;
}
char* get_str_uarch(struct cpuInfo* cpu) {
return cpu->arch->uarch_str;
}
void free_uarch_struct(struct uarch* arch) {
free(arch->uarch_str);
free(arch);
}

12
src/arm/uarch.h Normal file
View File

@@ -0,0 +1,12 @@
#ifndef __UARCH__
#define __UARCH__
#include <stdint.h>
#include "midr.h"
struct uarch* get_uarch_from_midr(uint32_t midr, struct cpuInfo* cpu);
char* get_str_uarch(struct cpuInfo* cpu);
void free_uarch_struct(struct uarch* arch);
#endif

162
src/arm/udev.c Normal file
View File

@@ -0,0 +1,162 @@
#include "udev.h"
#include "midr.h"
#define _PATH_CPUS_PRESENT _PATH_SYS_SYSTEM _PATH_SYS_CPU "/present"
#define _PATH_CPUINFO "/proc/cpuinfo"
//#define _PATH_CPUINFO "cpuinfo_debug"
#define CPUINFO_CPU_IMPLEMENTER_STR "CPU implementer\t: "
#define CPUINFO_CPU_ARCHITECTURE_STR "CPU architecture: "
#define CPUINFO_CPU_VARIANT_STR "CPU variant\t: "
#define CPUINFO_CPU_PART_STR "CPU part\t: "
#define CPUINFO_CPU_REVISION_STR "CPU revision\t: "
#define CPUINFO_HARDWARE_STR "Hardware\t: "
#define CPUINFO_CPU_STRING "processor"
// https://www.kernel.org/doc/html/latest/core-api/cpu_hotplug.html
int get_ncores_from_cpuinfo() {
// Examples:
// 0-271
// 0-5
// 0-7
int filelen;
char* buf;
if((buf = read_file(_PATH_CPUS_PRESENT, &filelen)) == NULL) {
perror("open");
return UNKNOWN;
}
int ncores = 0;
char* tmp1 = strstr(buf, "-") + 1;
char* tmp2 = strstr(buf, "\n");
char ncores_str[filelen];
memset(ncores_str, 0, sizeof(char) * filelen);
memcpy(ncores_str, tmp1, tmp2-tmp1);
char* end;
errno = 0;
ncores = strtol(ncores_str, &end, 10) + 1;
if(errno != 0) {
perror("strtol");
return UNKNOWN;
}
free(buf);
return ncores;
}
long parse_cpuinfo_field(char* buf, char* field_str, int field_base) {
char* tmp = strstr(buf, field_str);
if(tmp == NULL) return -1;
tmp += strlen(field_str);
char* end;
errno = 0;
long ret = strtol(tmp, &end, field_base);
if(errno != 0) {
perror("strtol");
return -1;
}
return ret;
}
// https://developer.arm.com/docs/ddi0595/h/aarch32-system-registers/midr
// https://static.docs.arm.com/ddi0595/h/SysReg_xml_v86A-2020-06.pdf
uint32_t get_midr_from_cpuinfo(uint32_t core, bool* success) {
int filelen;
char* buf;
*success = true;
if((buf = read_file(_PATH_CPUINFO, &filelen)) == NULL) {
perror("open");
*success = false;
return 0;
}
char* tmp = strstr(buf, CPUINFO_CPU_STRING);
uint32_t current_core = 0;
while(core != current_core && tmp != NULL) {
tmp++;
current_core++;
tmp = strstr(tmp, CPUINFO_CPU_STRING);
}
if(tmp == NULL) {
*success = false;
return 0;
}
uint32_t cpu_implementer;
uint32_t cpu_architecture;
uint32_t cpu_variant;
uint32_t cpu_part;
uint32_t cpu_revision;
uint32_t midr = 0;
long ret;
if ((ret = parse_cpuinfo_field(tmp, CPUINFO_CPU_IMPLEMENTER_STR, 16)) < 0) {
printf("Failed parsing cpu_implementer\n");
*success = false;
return 0;
}
cpu_implementer = (uint32_t) ret;
if ((ret = parse_cpuinfo_field(tmp, CPUINFO_CPU_ARCHITECTURE_STR, 10)) < 0) {
printf("Failed parsing cpu_architecture\n");
*success = false;
return 0;
}
cpu_architecture = (uint32_t) 0xF; // Why?
if ((ret = parse_cpuinfo_field(tmp, CPUINFO_CPU_VARIANT_STR, 16)) < 0) {
printf("Failed parsing cpu_variant\n");
*success = false;
return 0;
}
cpu_variant = (uint32_t) ret;
if ((ret = parse_cpuinfo_field(tmp, CPUINFO_CPU_PART_STR, 16)) < 0) {
printf("Failed parsing cpu_part\n");
*success = false;
return 0;
}
cpu_part = (uint32_t) ret;
if ((ret = parse_cpuinfo_field(tmp, CPUINFO_CPU_REVISION_STR, 10)) < 0) {
printf("Failed parsing cpu_revision\n");
*success = false;
return 0;
}
cpu_revision = (uint32_t) ret;
midr = midr_set_implementer(midr, cpu_implementer);
midr = midr_set_variant(midr, cpu_variant);
midr = midr_set_architecture(midr, cpu_architecture);
midr = midr_set_part(midr, cpu_part);
midr = midr_set_revision(midr, cpu_revision);
return midr;
}
char* get_hardware_from_cpuinfo() {
int filelen;
char* buf;
if((buf = read_file(_PATH_CPUINFO, &filelen)) == NULL) {
perror("open");
return NULL;
}
char* tmp1 = strstr(buf, CPUINFO_HARDWARE_STR);
if(tmp1 == NULL) return NULL;
tmp1 = tmp1 + strlen(CPUINFO_HARDWARE_STR);
char* tmp2 = strstr(tmp1, "\n");
int strlen = (1 + (tmp2-tmp1));
char* hardware = malloc(sizeof(char) * strlen);
memset(hardware, 0, sizeof(char) * strlen);
strncpy(hardware, tmp1, tmp2-tmp1);
return hardware;
}

12
src/arm/udev.h Normal file
View File

@@ -0,0 +1,12 @@
#ifndef __UDEV_ARM__
#define __UDEV_ARM__
#include "../common/udev.h"
#define UNKNOWN -1
int get_ncores_from_cpuinfo();
uint32_t get_midr_from_cpuinfo(uint32_t core, bool* success);
char* get_hardware_from_cpuinfo();
#endif

View File

@@ -1,49 +0,0 @@
#ifndef __ASCII__
#define __ASCII__
#define NUMBER_OF_LINES 19
#define LINE_SIZE 62
#define AMD_ASCII \
" \
\
\
\
\
\
@@@@ @@@ @@@ @@@@@@@@ ############ \
@@@@@@ @@@@@ @@@@ @@@ @@@@ ########## \
@@@ @@@ @@@@@@@@@@@@@ @@@ @@ # #### \
@@@ @@@ @@@ @@@ @@@ @@@ @@@ ### #### \
@@@@@@@@@@@@ @@@ @@@ @@@ @@@ #### ## ### \
@@@ @@@ @@@ @@@ @@@@@@@@@ ######## ## \
\
\
\
\
\
\
"
#define INTEL_ASCII \
" ################ \
####### ####### \
#### #### \
### #### \
### ### \
### ### \
# ### ### ### \
## ### ######### ###### ###### ### ### \
## ### ### ### ### #### #### ### ### \
## ### ### ### ### ### ### ### ### \
## ### ### ### ### ########## ### #### \
## ### ### ### ### ### ### ##### \
## ## ### ### ##### ######### ## ### \
### \
### \
#### #### \
##### ########## \
########## ################ \
############################### "
#endif

273
src/common/args.c Normal file
View File

@@ -0,0 +1,273 @@
#include <getopt.h>
#include <stdio.h>
#include <string.h>
#include <stdlib.h>
#include "args.h"
#include "global.h"
#define COLOR_STR_INTEL "intel"
#define COLOR_STR_AMD "amd"
#define COLOR_STR_ARM "arm"
static const char *SYTLES_STR_LIST[] = {
[STYLE_EMPTY] = NULL,
[STYLE_FANCY] = "fancy",
[STYLE_RETRO] = "retro",
[STYLE_LEGACY] = "legacy",
[STYLE_INVALID] = NULL
};
struct args_struct {
bool debug_flag;
bool help_flag;
bool verbose_flag;
bool version_flag;
STYLE style;
struct colors* colors;
};
const char args_chr[] = {
/* [ARG_CHAR_STYLE] = */ 's',
/* [ARG_CHAR_COLOR] = */ 'c',
/* [ARG_CHAR_HELP] = */ 'h',
/* [ARG_CHAR_DEBUG] = */ 'd',
/* [ARG_CHAR_VERBOSE] = */ 'v',
/* [ARG_CHAR_VERSION] = */ 'V',
};
const char *args_str[] = {
/* [ARG_CHAR_STYLE] = */ "style",
/* [ARG_CHAR_COLOR] = */ "color",
/* [ARG_CHAR_HELP] = */ "help",
/* [ARG_CHAR_DEBUG] = */ "debug",
/* [ARG_CHAR_VERBOSE] = */ "verbose",
/* [ARG_CHAR_VERSION] = */ "version",
};
static struct args_struct args;
STYLE get_style() {
return args.style;
}
struct colors* get_colors() {
return args.colors;
}
bool show_help() {
return args.help_flag;
}
bool show_version() {
return args.version_flag;
}
bool show_debug() {
return args.debug_flag;
}
bool verbose_enabled() {
return args.verbose_flag;
}
int max_arg_str_length() {
int max_len = -1;
int len = sizeof(args_str) / sizeof(args_str[0]);
for(int i=0; i < len; i++) {
max_len = max(max_len, (int) strlen(args_str[i]));
}
return max_len;
}
STYLE parse_style(char* style) {
uint8_t i = 0;
uint8_t styles_count = sizeof(SYTLES_STR_LIST) / sizeof(SYTLES_STR_LIST[0]);
while(i != styles_count && (SYTLES_STR_LIST[i] == NULL || strcmp(SYTLES_STR_LIST[i], style) != 0))
i++;
if(i == styles_count)
return STYLE_INVALID;
return i;
}
void free_colors_struct(struct colors* cs) {
free(cs->c1);
free(cs->c2);
free(cs->c3);
free(cs->c4);
free(cs);
}
bool parse_color(char* optarg_str, struct colors** cs) {
*cs = malloc(sizeof(struct colors));
(*cs)->c1 = malloc(sizeof(struct color));
(*cs)->c2 = malloc(sizeof(struct color));
(*cs)->c3 = malloc(sizeof(struct color));
(*cs)->c4 = malloc(sizeof(struct color));
struct color** c1 = &((*cs)->c1);
struct color** c2 = &((*cs)->c2);
struct color** c3 = &((*cs)->c3);
struct color** c4 = &((*cs)->c4);
int32_t ret;
char* str_to_parse = NULL;
bool free_ptr;
if(strcmp(optarg_str, COLOR_STR_INTEL) == 0) {
str_to_parse = malloc(sizeof(char) * 46);
strcpy(str_to_parse, COLOR_DEFAULT_INTEL);
free_ptr = true;
}
else if(strcmp(optarg_str, COLOR_STR_AMD) == 0) {
str_to_parse = malloc(sizeof(char) * 44);
strcpy(str_to_parse, COLOR_DEFAULT_AMD);
free_ptr = true;
}
else if(strcmp(optarg_str, COLOR_STR_ARM) == 0) {
str_to_parse = malloc(sizeof(char) * 46);
strcpy(str_to_parse, COLOR_DEFAULT_ARM);
free_ptr = true;
}
else {
str_to_parse = optarg_str;
free_ptr = false;
}
ret = sscanf(str_to_parse, "%d,%d,%d:%d,%d,%d:%d,%d,%d:%d,%d,%d",
&(*c1)->R, &(*c1)->G, &(*c1)->B,
&(*c2)->R, &(*c2)->G, &(*c2)->B,
&(*c3)->R, &(*c3)->G, &(*c3)->B,
&(*c4)->R, &(*c4)->G, &(*c4)->B);
if(ret != 12) {
printErr("Expected to read 12 values for color but read %d", ret);
return false;
}
//TODO: Refactor c1->R c2->R ... to c[i]->R
if((*c1)->R < 0 || (*c1)->R > 255) {
printErr("Red in color 1 is invalid. Must be in range (0, 255)");
return false;
}
if((*c1)->G < 0 || (*c1)->G > 255) {
printErr("Green in color 1 is invalid. Must be in range (0, 255)");
return false;
}
if((*c1)->B < 0 || (*c1)->B > 255) {
printErr("Blue in color 1 is invalid. Must be in range (0, 255)");
return false;
}
if((*c2)->R < 0 || (*c2)->R > 255) {
printErr("Red in color 2 is invalid. Must be in range (0, 255)");
return false;
}
if((*c2)->G < 0 || (*c2)->G > 255) {
printErr("Green in color 2 is invalid. Must be in range (0, 255)");
return false;
}
if((*c2)->B < 0 || (*c2)->B > 255) {
printErr("Blue in color 2 is invalid. Must be in range (0, 255)");
return false;
}
if(free_ptr) free (str_to_parse);
return true;
}
char* build_short_options() {
const char *c = args_chr;
int len = sizeof(args_chr) / sizeof(args_chr[0]);
char* str = (char *) malloc(sizeof(char) * (len*2 + 1));
memset(str, 0, sizeof(char) * (len*2 + 1));
sprintf(str, "%c:%c:%c%c%c%c",
c[ARG_STYLE], c[ARG_COLOR], c[ARG_HELP],
c[ARG_DEBUG], c[ARG_VERBOSE], c[ARG_VERSION]);
return str;
}
bool parse_args(int argc, char* argv[]) {
int opt;
int option_index = 0;
opterr = 0;
bool color_flag = false;
args.debug_flag = false;
args.verbose_flag = false;
args.help_flag = false;
args.style = STYLE_EMPTY;
args.colors = NULL;
const struct option long_options[] = {
{args_str[ARG_STYLE], required_argument, 0, args_chr[ARG_STYLE] },
{args_str[ARG_COLOR], required_argument, 0, args_chr[ARG_COLOR] },
{args_str[ARG_HELP], no_argument, 0, args_chr[ARG_HELP] },
{args_str[ARG_DEBUG], no_argument, 0, args_chr[ARG_DEBUG] },
{args_str[ARG_VERBOSE], no_argument, 0, args_chr[ARG_VERBOSE] },
{args_str[ARG_VERSION], no_argument, 0, args_chr[ARG_VERSION] },
{0, 0, 0, 0}
};
char* short_options = build_short_options();
opt = getopt_long(argc, argv, short_options, long_options, &option_index);
while (!args.help_flag && !args.debug_flag && !args.version_flag && opt != -1) {
if(opt == args_chr[ARG_COLOR]) {
if(color_flag) {
printErr("Color option specified more than once");
return false;
}
color_flag = true;
if(!parse_color(optarg, &args.colors)) {
printErr("Color parsing failed");
return false;
}
}
else if(opt == args_chr[ARG_STYLE]) {
if(args.style != STYLE_EMPTY) {
printErr("Style option specified more than once");
return false;
}
args.style = parse_style(optarg);
if(args.style == STYLE_INVALID) {
printErr("Invalid style '%s'",optarg);
return false;
}
break;
}
else if(opt == args_chr[ARG_HELP]) {
args.help_flag = true;
}
else if(opt == args_chr[ARG_VERBOSE]) {
args.verbose_flag = true;
}
else if(opt == args_chr[ARG_DEBUG]) {
args.debug_flag = true;
}
else if(opt == args_chr[ARG_VERSION]) {
args.version_flag = true;
}
else {
printWarn("Invalid options");
args.help_flag = true;
}
option_index = 0;
opt = getopt_long(argc, argv, short_options, long_options, &option_index);
}
if(optind < argc) {
printWarn("Invalid options");
args.help_flag = true;
}
if((args.help_flag + args.version_flag + color_flag) > 1) {
printWarn("You should specify just one option");
args.help_flag = true;
}
return true;
}

View File

@@ -17,11 +17,33 @@ struct colors {
struct color* c4;
};
enum {
STYLE_EMPTY,
STYLE_FANCY,
STYLE_WILD,
STYLE_RETRO,
STYLE_LEGACY,
STYLE_INVALID
};
enum {
ARG_STYLE,
ARG_COLOR,
ARG_HELP,
ARG_DEBUG,
ARG_VERBOSE,
ARG_VERSION
};
extern const char args_chr[];
extern const char *args_str[];
#include "printer.h"
int max_arg_str_length();
bool parse_args(int argc, char* argv[]);
bool show_help();
bool show_levels();
bool show_debug();
bool show_version();
bool verbose_enabled();
void free_colors_struct(struct colors* cs);

208
src/common/ascii.h Normal file
View File

@@ -0,0 +1,208 @@
#ifndef __ASCII__
#define __ASCII__
#define NUMBER_OF_LINES 19
#define LINE_SIZE 62
#define AMD_ASCII \
" \
\
\
\
\
\
@@@@ @@@ @@@ @@@@@@@@ ############ \
@@@@@@ @@@@@ @@@@@ @@@ @@@ ########## \
@@@ @@@ @@@@@@@@@@@@@ @@@ @@ # ##### \
@@@ @@@ @@@ @@@ @@@ @@@ @@ ### ##### \
@@@@@@@@@@@@ @@@ @@@ @@@ @@@ ######### ### \
@@@ @@@ @@@ @@@ @@@@@@@@@ ######## ## \
\
\
\
\
\
\
"
#define INTEL_ASCII \
" ################ \
####### ####### \
#### #### \
### #### \
### ### \
### ### \
# ### ### ### \
## ### ######### ###### ###### ### ### \
## ### ### ### ### #### #### ### ### \
## ### ### ### ### ### ### ### ### \
## ### ### ### ### ########## ### #### \
## ### ### ### ### ### ### ##### \
## ## ### ### ##### ######### ## ### \
### \
### \
#### #### \
##### ########## \
########## ################ \
############################### "
#define SNAPDRAGON_ASCII \
" \
@@######## \
@@@@@########### \
@@ @@@@@################# \
@@@@@@@@@@#################### \
@@@@@@@@@@@@##################### \
@@@@@@@@@@@@@@@#################### \
@@@@@@@@@@@@@@@@@################### \
@@@@@@@@@@@@@@@@@@@@################ \
@@@@@@@@@@@@@@@@@@@@############# \
@@@@@@@@@@@@@@@@@@############ \
@ @@@@@@@@@@@@@@@########### \
@@@@@ @@@@@@@@@@@@@########## \
@@@@@@@@@ @@@@@@@@@@@@######## \
@@@@@@@@@ @@@@@@@@@@####### \
@@@@@@@@@@@@@@@@####### \
@@@@########### \
\
"
#define MEDIATEK_ASCII \
" \
\
\
\
\
\
## ## ###### ###### # ### @@@@@@ @@@@@@ @@ @@ \
### ### # # # # #### @@ @ @@ @@ \
######## # ### # # # ## ## @@ @ @@@ @@@@ \
## ### ## # # # # ## ## @@ @ @@ @@ \
## ## ## ###### ##### # ## ## @@ @@@@@@ @@ @@ \
\
\
\
\
\
\
\
"
#define EXYNOS_ASCII \
" \
\
\
\
\
\
## ## ## \
## ## \
## \
## ## \
## ## ## \
\
SAMSUNG \
Exynos \
\
\
\
\
"
#define KIRIN_ASCII \
" \
\
\
\
\
####### \
##### #################### \
###################################### \
####################################### \
####################################### \
############################## \
########################## \
######################### \
######################## \
######################## \
######################### \
######################### \
\
"
#define BROADCOM_ASCII \
" \
################ \
######################### \
############################### \
################@@@@################ \
################@@@@@@################ \
#################@@@@@@################ \
#################@@@@@@@@################# \
#################@@@@@@@@################# \
#################@@@@##@@@@################ \
################@@@@##@@@@################ \
###############@@@@####@@@@############### \
@@@@@@@@@@####@@@@####@@@@####@@@@@@@@@@ \
######@@@@@@@@@@######@@@@@@@@@@###### \
################################## \
############################## \
######################## \
############### \
"
#define ARM_ASCII \
" \
\
\
\
\
\
############ ########## #### ###### ######## \
############### ######### ####################### \
#### #### #### ##### ####### ##### \
#### #### #### #### ##### #### \
#### #### #### #### #### #### \
#### ##### #### #### #### #### \
############### #### #### #### #### \
######## #### #### #### #### #### \
\
\
\
\
"
#define UNKNOWN_ASCII \
" \
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
\
"
static const char* ASCII_ARRAY [] = {
AMD_ASCII,
INTEL_ASCII,
ARM_ASCII,
SNAPDRAGON_ASCII,
MEDIATEK_ASCII,
EXYNOS_ASCII,
KIRIN_ASCII,
BROADCOM_ASCII,
UNKNOWN_ASCII
};
#endif

186
src/common/cpu.c Executable file
View File

@@ -0,0 +1,186 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <stdbool.h>
#include "../common/global.h"
#include "cpu.h"
#ifdef ARCH_X86
#include "../x86/uarch.h"
#elif ARCH_ARM
#include "../arm/uarch.h"
#endif
#define STRING_UNKNOWN "Unknown"
#define STRING_YES "Yes"
#define STRING_NO "No"
#define STRING_NONE "None"
#define STRING_MEGAHERZ "MHz"
#define STRING_GIGAHERZ "GHz"
#define STRING_KILOBYTES "KB"
#define STRING_MEGABYTES "MB"
VENDOR get_cpu_vendor(struct cpuInfo* cpu) {
return cpu->cpu_vendor;
}
int64_t get_freq(struct frequency* freq) {
return freq->max;
}
#ifdef ARCH_X86
char* get_str_cpu_name(struct cpuInfo* cpu) {
return cpu->cpu_name;
}
char* get_str_sockets(struct topology* topo) {
char* string = malloc(sizeof(char) * 2);
int32_t sanity_ret = snprintf(string, 2, "%d", topo->sockets);
if(sanity_ret < 0) {
printBug("get_str_sockets: snprintf returned a negative value for input: '%d'", topo->sockets);
return NULL;
}
return string;
}
uint32_t get_nsockets(struct topology* topo) {
return topo->sockets;
}
#endif
int32_t get_value_as_smallest_unit(char ** str, uint32_t value) {
int32_t sanity_ret;
*str = malloc(sizeof(char)* 11); //8 for digits, 2 for units
if(value/1024 >= 1024)
sanity_ret = snprintf(*str, 10,"%.4g"STRING_MEGABYTES, (double)value/(1<<20));
else
sanity_ret = snprintf(*str, 10,"%.4g"STRING_KILOBYTES, (double)value/(1<<10));
return sanity_ret;
}
// String functions
char* get_str_cache_two(int32_t cache_size, uint32_t physical_cores) {
// 4 for digits, 2 for units, 2 for ' (', 3 digits, 2 for units and 7 for ' Total)'
uint32_t max_size = 4+2 + 2 + 4+2 + 7 + 1;
int32_t sanity_ret;
char* string = malloc(sizeof(char) * max_size);
char* tmp1;
char* tmp2;
int32_t tmp1_len = get_value_as_smallest_unit(&tmp1, cache_size);
int32_t tmp2_len = get_value_as_smallest_unit(&tmp2, cache_size * physical_cores);
if(tmp1_len < 0) {
printBug("get_value_as_smallest_unit: snprintf returned a negative value for input: %d\n", cache_size);
return NULL;
}
if(tmp2_len < 0) {
printBug("get_value_as_smallest_unit: snprintf returned a negative value for input: %d\n", cache_size * physical_cores);
return NULL;
}
uint32_t size = tmp1_len + 2 + tmp2_len + 7 + 1;
sanity_ret = snprintf(string, size, "%s (%s Total)", tmp1, tmp2);
if(sanity_ret < 0) {
printBug("get_str_cache_two: snprintf returned a negative value for input: '%s' and '%s'\n", tmp1, tmp2);
return NULL;
}
free(tmp1);
free(tmp2);
return string;
}
char* get_str_cache_one(int32_t cache_size) {
// 4 for digits, 2 for units, 2 for ' (', 3 digits, 2 for units and 7 for ' Total)'
uint32_t max_size = 4+2 + 1;
int32_t sanity_ret;
char* string = malloc(sizeof(char) * max_size);
char* tmp;
int32_t tmp_len = get_value_as_smallest_unit(&tmp, cache_size);
if(tmp_len < 0) {
printBug("get_value_as_smallest_unit: snprintf returned a negative value for input: %d", cache_size);
return NULL;
}
uint32_t size = tmp_len + 1;
sanity_ret = snprintf(string, size, "%s", tmp);
if(sanity_ret < 0) {
printBug("get_str_cache_one: snprintf returned a negative value for input: '%s'", tmp);
return NULL;
}
free(tmp);
return string;
}
char* get_str_cache(int32_t cache_size, int32_t num_caches) {
if(num_caches > 1)
return get_str_cache_two(cache_size, num_caches);
else
return get_str_cache_one(cache_size);
}
char* get_str_l1i(struct cache* cach) {
return get_str_cache(cach->L1i->size, cach->L1i->num_caches);
}
char* get_str_l1d(struct cache* cach) {
return get_str_cache(cach->L1d->size, cach->L1d->num_caches);
}
char* get_str_l2(struct cache* cach) {
assert(cach->L2->exists);
return get_str_cache(cach->L2->size, cach->L2->num_caches);
}
char* get_str_l3(struct cache* cach) {
if(!cach->L3->exists)
return NULL;
return get_str_cache(cach->L3->size, cach->L3->num_caches);
}
char* get_str_freq(struct frequency* freq) {
//Max 3 digits and 3 for '(M/G)Hz' plus 1 for '\0'
uint32_t size = (5+1+3+1);
assert(strlen(STRING_UNKNOWN)+1 <= size);
char* string = malloc(sizeof(char)*size);
memset(string, 0, sizeof(char)*size);
if(freq->max == UNKNOWN_FREQ || freq->max < 0)
snprintf(string,strlen(STRING_UNKNOWN)+1,STRING_UNKNOWN);
else if(freq->max >= 1000)
snprintf(string,size,"%.3f "STRING_GIGAHERZ,(float)(freq->max)/1000);
else
snprintf(string,size,"%d "STRING_MEGAHERZ,freq->max);
return string;
}
void free_cache_struct(struct cache* cach) {
for(int i=0; i < 4; i++) free(cach->cach_arr[i]);
free(cach->cach_arr);
free(cach);
}
void free_freq_struct(struct frequency* freq) {
free(freq);
}
void free_hv_struct(struct hypervisor* hv) {
free(hv);
}
void free_cpuinfo_struct(struct cpuInfo* cpu) {
free_uarch_struct(cpu->arch);
free_hv_struct(cpu->hv);
#ifdef ARCH_X86
free(cpu->cpu_name);
#endif
free(cpu);
}

158
src/common/cpu.h Normal file
View File

@@ -0,0 +1,158 @@
#ifndef __CPU__
#define __CPU__
#include <stdint.h>
#include <stdbool.h>
enum {
// ARCH_X86
CPU_VENDOR_INTEL,
CPU_VENDOR_AMD,
// ARCH_ARM
CPU_VENDOR_ARM,
CPU_VENDOR_BROADCOM,
CPU_VENDOR_CAVIUM,
CPU_VENDOR_NVIDIA,
CPU_VENDOR_APM,
CPU_VENDOR_QUALCOMM,
CPU_VENDOR_HUAWUEI,
CPU_VENDOR_SAMSUNG,
CPU_VENDOR_MARVELL,
// OTHERS
CPU_VENDOR_UNKNOWN,
CPU_VENDOR_INVALID
};
enum {
HV_VENDOR_KVM,
HV_VENDOR_QEMU,
HV_VENDOR_HYPERV,
HV_VENDOR_VMWARE,
HV_VENDOR_XEN,
HV_VENDOR_PARALLELS,
HV_VENDOR_INVALID
};
#define UNKNOWN_FREQ -1
#define CPU_NAME_MAX_LENGTH 64
typedef int32_t VENDOR;
struct frequency {
int32_t base;
int32_t max;
};
struct hypervisor {
bool present;
char* hv_name;
VENDOR hv_vendor;
};
struct cach {
int32_t size;
uint8_t num_caches;
bool exists;
// plenty of more properties to include in the future...
};
struct cache {
struct cach* L1i;
struct cach* L1d;
struct cach* L2;
struct cach* L3;
struct cach** cach_arr;
uint8_t max_cache_level;
};
struct topology {
int32_t total_cores;
struct cache* cach;
#ifdef ARCH_X86
uint32_t physical_cores;
uint32_t logical_cores;
uint32_t smt_available; // Number of SMT that is currently enabled
uint32_t smt_supported; // Number of SMT that CPU supports (equal to smt_available if SMT is enabled)
uint32_t sockets;
struct apic* apic;
#endif
};
struct features {
bool AES; // Must be the first field of features struct!
#ifdef ARCH_X86
bool AVX;
bool AVX2;
bool AVX512;
bool SSE;
bool SSE2;
bool SSE3;
bool SSSE3;
bool SSE4a;
bool SSE4_1;
bool SSE4_2;
bool FMA3;
bool FMA4;
bool SHA;
#elif ARCH_ARM
bool NEON;
bool SHA1;
bool SHA2;
bool CRC32;
#endif
};
struct cpuInfo {
VENDOR cpu_vendor;
struct uarch* arch;
struct hypervisor* hv;
struct frequency* freq;
struct cache* cach;
struct topology* topo;
struct features* feat;
#ifdef ARCH_X86
// CPU name from model
char* cpu_name;
// Max cpuids levels
uint32_t maxLevels;
// Max cpuids extended levels
uint32_t maxExtendedLevels;
#elif ARCH_ARM
// Main ID register
uint32_t midr;
#endif
#ifdef ARCH_ARM
struct system_on_chip* soc;
// If SoC contains more than one CPU and they
// are different, the others will be stored in
// the next_cpu field
struct cpuInfo* next_cpu;
uint8_t num_cpus;
#endif
};
#ifdef ARCH_X86
char* get_str_cpu_name(struct cpuInfo* cpu);
char* get_str_sockets(struct topology* topo);
uint32_t get_nsockets(struct topology* topo);
#endif
VENDOR get_cpu_vendor(struct cpuInfo* cpu);
int64_t get_freq(struct frequency* freq);
char* get_str_aes(struct cpuInfo* cpu);
char* get_str_sha(struct cpuInfo* cpu);
char* get_str_l1i(struct cache* cach);
char* get_str_l1d(struct cache* cach);
char* get_str_l2(struct cache* cach);
char* get_str_l3(struct cache* cach);
char* get_str_freq(struct frequency* freq);
void free_cache_struct(struct cache* cach);
void free_freq_struct(struct frequency* freq);
void free_cpuinfo_struct(struct cpuInfo* cpu);
#endif

View File

@@ -16,8 +16,10 @@
#endif
#define LOG_LEVEL_NORMAL 0
#define LOG_LEVEL_VERBOSE 1
enum {
LOG_LEVEL_NORMAL,
LOG_LEVEL_VERBOSE
};
int LOG_LEVEL;
@@ -51,10 +53,18 @@ void printBug(const char *fmt, ...) {
vsnprintf(buffer,buffer_size, fmt, args);
va_end(args);
fprintf(stderr,RED "[ERROR]: "RESET "%s\n",buffer);
fprintf(stderr,"Please, create a new issue with this error message and your CPU in https://github.com/Dr-Noob/cpufetch/issues\n");
#ifdef ARCH_X86
fprintf(stderr,"Please, create a new issue with this error message and the output of 'cpufetch --debug' in https://github.com/Dr-Noob/cpufetch/issues\n");
#elif ARCH_ARM
fprintf(stderr,"Please, create a new issue with this error message, your smartphone/computer model and the output of 'cpufetch --debug' in https://github.com/Dr-Noob/cpufetch/issues\n");
#endif
}
void set_log_level(bool verbose) {
if(verbose) LOG_LEVEL = LOG_LEVEL_VERBOSE;
else LOG_LEVEL = LOG_LEVEL_NORMAL;
}
int max(int a, int b) {
return a > b ? a : b;
}

View File

@@ -7,5 +7,6 @@ void set_log_level(bool verbose);
void printWarn(const char *fmt, ...);
void printErr(const char *fmt, ...);
void printBug(const char *fmt, ...);
int max(int a, int b);
#endif

102
src/common/main.c Normal file
View File

@@ -0,0 +1,102 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "args.h"
#include "printer.h"
#include "global.h"
#ifdef ARCH_X86
static const char* ARCH_STR = "x86_64 build";
#include "../x86/cpuid.h"
#elif ARCH_ARM
static const char* ARCH_STR = "ARM build";
#include "../arm/midr.h"
#endif
static const char* VERSION = "0.96";
void print_help(char *argv[]) {
const char **t = args_str;
const char *c = args_chr;
int max_len = max_arg_str_length();
printf("Usage: %s [OPTION]...\n", argv[0]);
printf("Simple yet fancy CPU architecture fetching tool\n\n");
printf("Options: \n");
printf(" -%c, --%s %*s Set the color scheme (by default, cpufetch uses the system color scheme)\n", c[ARG_COLOR], t[ARG_COLOR], (int) (max_len-strlen(t[ARG_COLOR])), "");
printf(" -%c, --%s %*s Set the style of CPU art\n", c[ARG_STYLE], t[ARG_STYLE], (int) (max_len-strlen(t[ARG_STYLE])), "");
#ifdef ARCH_X86
printf(" -%c, --%s %*s Prints CPU model and cpuid levels (debug purposes)\n", c[ARG_DEBUG], t[ARG_DEBUG], (int) (max_len-strlen(t[ARG_DEBUG])), "");
#elif ARCH_ARM
printf(" -%c, --%s %*s Prints main ID register values for all cores (debug purposes)\n", c[ARG_DEBUG], t[ARG_DEBUG], (int) (max_len-strlen(t[ARG_DEBUG])), "");
#endif
printf(" -%c, --%s %*s Prints extra information (if available) about how cpufetch tried fetching information\n", c[ARG_VERBOSE], t[ARG_VERBOSE], (int) (max_len-strlen(t[ARG_VERBOSE])), "");
printf(" -%c, --%s %*s Prints this help and exit\n", c[ARG_HELP], t[ARG_HELP], (int) (max_len-strlen(t[ARG_HELP])), "");
printf(" -%c, --%s %*s Prints cpufetch version and exit\n", c[ARG_VERSION], t[ARG_VERSION], (int) (max_len-strlen(t[ARG_VERSION])), "");
printf("\nCOLORS: \n");
printf(" * \"intel\": Use Intel default color scheme \n");
printf(" * \"amd\": Use AMD default color scheme \n");
printf(" * \"arm\": Use ARM default color scheme \n");
printf(" * custom: If color argument do not match \"intel\", \"amd\" or \"arm\", a custom scheme can be specified.\n");
printf(" 4 colors must be given in RGB with the format: R,G,B:R,G,B:...\n");
printf(" The first 2 colors are the CPU art color and the next 2 colors are the text colors\n");
printf("\nSTYLES: \n");
printf(" * \"fancy\": Default style\n");
printf(" * \"retro\": Old cpufetch style\n");
printf(" * \"legacy\": Fallback style for terminals that do not support colors\n");
printf("\nEXAMPLES: \n");
printf(" Run cpufetch with Intel color scheme:\n");
printf(" ./cpufetch --color intel\n");
printf(" Run cpufetch with a custom color scheme:\n");
printf(" ./cpufetch --color 239,90,45:210,200,200:100,200,45:0,200,200\n");
printf("\nBUGS: \n");
printf(" Report bugs to https://github.com/Dr-Noob/cpufetch/issues\n");
printf("\nNOTE: \n");
printf(" Peak performance information is NOT accurate. cpufetch computes peak performance using the max\n");
printf(" frequency. However, to properly compute peak performance, you need to know the frequency of the\n");
printf(" CPU running AVX code, which is not be fetched by cpufetch since it depends on each specific CPU.\n");
printf(" For peak performance measurement see: https://github.com/Dr-Noob/peakperf\n");
}
void print_version() {
printf("cpufetch v%s (%s)\n",VERSION, ARCH_STR);
}
int main(int argc, char* argv[]) {
if(!parse_args(argc,argv))
return EXIT_FAILURE;
if(show_help()) {
print_help(argv);
return EXIT_SUCCESS;
}
if(show_version()) {
print_version();
return EXIT_SUCCESS;
}
set_log_level(verbose_enabled());
struct cpuInfo* cpu = get_cpu_info();
if(cpu == NULL)
return EXIT_FAILURE;
if(show_debug()) {
print_version();
print_debug(cpu);
return EXIT_SUCCESS;
}
if(print_cpufetch(cpu, get_style(), get_colors()))
return EXIT_SUCCESS;
else
return EXIT_FAILURE;
}

716
src/common/printer.c Normal file
View File

@@ -0,0 +1,716 @@
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <stdbool.h>
#include "printer.h"
#include "ascii.h"
#include "../common/global.h"
#include "../common/cpu.h"
#ifdef ARCH_X86
#include "../x86/uarch.h"
#include "../x86/cpuid.h"
#else
#include "../arm/uarch.h"
#include "../arm/midr.h"
#include "../arm/soc.h"
#endif
#ifdef _WIN32
#define NOMINMAX
#include <Windows.h>
#endif
#define max(a,b) (((a)>(b))?(a):(b))
#define MAX_ATTRIBUTES 100
#define COLOR_NONE ""
#define COLOR_FG_BLACK "\x1b[30;1m"
#define COLOR_FG_RED "\x1b[31;1m"
#define COLOR_FG_GREEN "\x1b[32;1m"
#define COLOR_FG_YELLOW "\x1b[33;1m"
#define COLOR_FG_BLUE "\x1b[34;1m"
#define COLOR_FG_MAGENTA "\x1b[35;1m"
#define COLOR_FG_CYAN "\x1b[36;1m"
#define COLOR_FG_WHITE "\x1b[37;1m"
#define COLOR_BG_BLACK "\x1b[40;1m"
#define COLOR_BG_RED "\x1b[41;1m"
#define COLOR_BG_GREEN "\x1b[42;1m"
#define COLOR_BG_YELLOW "\x1b[43;1m"
#define COLOR_BG_BLUE "\x1b[44;1m"
#define COLOR_BG_MAGENTA "\x1b[45;1m"
#define COLOR_BG_CYAN "\x1b[46;1m"
#define COLOR_BG_WHITE "\x1b[47;1m"
#define COLOR_RESET "\x1b[m"
enum {
#ifdef ARCH_X86
ATTRIBUTE_NAME,
#elif ARCH_ARM
ATTRIBUTE_SOC,
ATTRIBUTE_CPU_NUM,
#endif
ATTRIBUTE_HYPERVISOR,
ATTRIBUTE_UARCH,
ATTRIBUTE_TECHNOLOGY,
ATTRIBUTE_FREQUENCY,
ATTRIBUTE_SOCKETS,
ATTRIBUTE_NCORES,
ATTRIBUTE_NCORES_DUAL,
#ifdef ARCH_X86
ATTRIBUTE_AVX,
ATTRIBUTE_FMA,
#elif ARCH_ARM
ATTRIBUTE_FEATURES,
#endif
ATTRIBUTE_L1i,
ATTRIBUTE_L1d,
ATTRIBUTE_L2,
ATTRIBUTE_L3,
ATTRIBUTE_PEAK
};
static const char* ATTRIBUTE_FIELDS [] = {
#ifdef ARCH_X86
"Name:",
#elif ARCH_ARM
"SoC:",
"",
#endif
"Hypervisor:",
"Microarchitecture:",
"Technology:",
"Max Frequency:",
"Sockets:",
"Cores:",
"Cores (Total):",
#ifdef ARCH_X86
"AVX:",
"FMA:",
#elif ARCH_ARM
"Features: ",
#endif
"L1i Size:",
"L1d Size:",
"L2 Size:",
"L3 Size:",
"Peak Performance:",
};
struct attribute {
int type;
char* value;
};
struct ascii {
char art[NUMBER_OF_LINES][LINE_SIZE+1];
char color1_ascii[100];
char color2_ascii[100];
char color1_text[100];
char color2_text[100];
char ascii_chars[2];
char reset[100];
struct attribute** attributes;
uint32_t n_attributes_set;
uint32_t additional_spaces;
VENDOR vendor;
STYLE style;
};
void setAttribute(struct ascii* art, int type, char* value) {
art->attributes[art->n_attributes_set]->value = value;
art->attributes[art->n_attributes_set]->type = type;
art->n_attributes_set++;
if(art->n_attributes_set > MAX_ATTRIBUTES) {
printBug("Set %d attributes, while max value is %d!", art->n_attributes_set, MAX_ATTRIBUTES);
}
}
char* rgb_to_ansi(struct color* c, bool background, bool bold) {
char* str = malloc(sizeof(char) * 100);
if(background) {
snprintf(str, 44, "\x1b[48;2;%.3d;%.3d;%.3dm", c->R, c->G, c->B);
}
else {
if(bold)
snprintf(str, 48, "\x1b[1m\x1b[38;2;%.3d;%.3d;%.3dm", c->R, c->G, c->B);
else
snprintf(str, 44, "\x1b[38;2;%.3d;%.3d;%.3dm", c->R, c->G, c->B);
}
return str;
}
struct ascii* set_ascii(VENDOR vendor, STYLE style, struct colors* cs) {
char *COL_FANCY_1, *COL_FANCY_2, *COL_FANCY_3, *COL_FANCY_4, *COL_RETRO_1, *COL_RETRO_2, *COL_RETRO_3, *COL_RETRO_4;
struct ascii* art = malloc(sizeof(struct ascii));
art->n_attributes_set = 0;
art->additional_spaces = 0;
art->vendor = vendor;
art->attributes = malloc(sizeof(struct attribute *) * MAX_ATTRIBUTES);
for(uint32_t i=0; i < MAX_ATTRIBUTES; i++) {
art->attributes[i] = malloc(sizeof(struct attribute));
art->attributes[i]->type = 0;
art->attributes[i]->value = NULL;
}
strcpy(art->reset, COLOR_RESET);
#ifdef ARCH_X86
if(art->vendor == CPU_VENDOR_INTEL) {
COL_FANCY_1 = COLOR_BG_CYAN;
COL_FANCY_2 = COLOR_BG_WHITE;
COL_FANCY_3 = COLOR_FG_CYAN;
COL_FANCY_4 = COLOR_FG_WHITE;
art->ascii_chars[0] = '#';
}
else if(art->vendor == CPU_VENDOR_AMD) {
COL_FANCY_1 = COLOR_BG_WHITE;
COL_FANCY_2 = COLOR_BG_GREEN;
COL_FANCY_3 = COLOR_FG_WHITE;
COL_FANCY_4 = COLOR_FG_GREEN;
art->ascii_chars[0] = '@';
}
else {
printBug("Invalid CPU vendor in set_ascii (%d)", art->vendor);
return NULL;
}
#elif ARCH_ARM
if(art->vendor == SOC_VENDOR_SNAPDRAGON) {
COL_FANCY_1 = COLOR_BG_RED;
COL_FANCY_2 = COLOR_BG_WHITE;
COL_FANCY_3 = COLOR_FG_RED;
COL_FANCY_4 = COLOR_FG_WHITE;
art->ascii_chars[0] = '@';
}
else if(art->vendor == SOC_VENDOR_MEDIATEK) {
COL_FANCY_1 = COLOR_BG_BLUE;
COL_FANCY_2 = COLOR_BG_YELLOW;
COL_FANCY_3 = COLOR_FG_WHITE;
COL_FANCY_4 = COLOR_FG_BLUE;
art->ascii_chars[0] = '@';
}
else if(art->vendor == SOC_VENDOR_EXYNOS) {
COL_FANCY_1 = COLOR_BG_BLUE;
COL_FANCY_2 = COLOR_BG_WHITE;
COL_FANCY_3 = COLOR_FG_BLUE;
COL_FANCY_4 = COLOR_FG_WHITE;
art->ascii_chars[0] = '@';
}
else if(art->vendor == SOC_VENDOR_KIRIN) {
COL_FANCY_1 = COLOR_BG_WHITE;
COL_FANCY_2 = COLOR_BG_RED;
COL_FANCY_3 = COLOR_FG_WHITE;
COL_FANCY_4 = COLOR_FG_RED;
art->ascii_chars[0] = '@';
}
else if(art->vendor == SOC_VENDOR_BROADCOM) {
COL_FANCY_1 = COLOR_BG_WHITE;
COL_FANCY_2 = COLOR_BG_RED;
COL_FANCY_3 = COLOR_FG_WHITE;
COL_FANCY_4 = COLOR_FG_RED;
art->ascii_chars[0] = '@';
}
else {
COL_FANCY_1 = COLOR_BG_CYAN;
COL_FANCY_2 = COLOR_BG_CYAN;
COL_FANCY_3 = COLOR_FG_WHITE;
COL_FANCY_4 = COLOR_FG_CYAN;
art->ascii_chars[0] = '#';
}
#endif
COL_RETRO_1 = COL_FANCY_3;
COL_RETRO_2 = COL_FANCY_4;
COL_RETRO_3 = COL_RETRO_1;
COL_RETRO_4 = COL_RETRO_2;
art->ascii_chars[1] = '#';
#ifdef _WIN32
HANDLE std_handle = GetStdHandle(STD_OUTPUT_HANDLE);
DWORD console_mode;
// Attempt to enable the VT100-processing flag
GetConsoleMode(std_handle, &console_mode);
SetConsoleMode(std_handle, console_mode | ENABLE_VIRTUAL_TERMINAL_PROCESSING);
// Get the console mode flag again, to see if it successfully enabled it
GetConsoleMode(std_handle, &console_mode);
#endif
if(style == STYLE_EMPTY) {
#ifdef _WIN32
// Use fancy style if VT100-processing is enabled,
// or legacy style in other case
art->style = (console_mode & ENABLE_VIRTUAL_TERMINAL_PROCESSING) ? STYLE_FANCY : STYLE_LEGACY;
#else
art->style = STYLE_FANCY;
#endif
}
else {
art->style = style;
}
switch(art->style) {
case STYLE_LEGACY:
strcpy(art->color1_ascii, COLOR_NONE);
strcpy(art->color2_ascii, COLOR_NONE);
strcpy(art->color1_text, COLOR_NONE);
strcpy(art->color2_text, COLOR_NONE);
art->reset[0] = '\0';
break;
case STYLE_FANCY:
if(cs != NULL) {
COL_FANCY_1 = rgb_to_ansi(cs->c1, true, true);
COL_FANCY_2 = rgb_to_ansi(cs->c2, true, true);
COL_FANCY_3 = rgb_to_ansi(cs->c3, false, true);
COL_FANCY_4 = rgb_to_ansi(cs->c4, false, true);
}
art->ascii_chars[0] = ' ';
art->ascii_chars[1] = ' ';
strcpy(art->color1_ascii,COL_FANCY_1);
strcpy(art->color2_ascii,COL_FANCY_2);
strcpy(art->color1_text,COL_FANCY_3);
strcpy(art->color2_text,COL_FANCY_4);
if(cs != NULL) {
free(COL_FANCY_1);
free(COL_FANCY_2);
free(COL_FANCY_3);
free(COL_FANCY_4);
}
break;
case STYLE_RETRO:
if(cs != NULL) {
COL_RETRO_1 = rgb_to_ansi(cs->c1, false, true);
COL_RETRO_2 = rgb_to_ansi(cs->c2, false, true);
COL_RETRO_3 = rgb_to_ansi(cs->c3, false, true);
COL_RETRO_4 = rgb_to_ansi(cs->c4, false, true);
}
strcpy(art->color1_ascii,COL_RETRO_1);
strcpy(art->color2_ascii,COL_RETRO_2);
strcpy(art->color1_text,COL_RETRO_3);
strcpy(art->color2_text,COL_RETRO_4);
if(cs != NULL) {
free(COL_RETRO_1);
free(COL_RETRO_2);
free(COL_RETRO_3);
free(COL_RETRO_4);
}
break;
case STYLE_INVALID:
default:
printBug("Found invalid style (%d)", art->style);
return NULL;
}
char tmp[NUMBER_OF_LINES * LINE_SIZE + 1];
#ifdef ARCH_X86
if(art->vendor == CPU_VENDOR_INTEL)
strcpy(tmp, INTEL_ASCII);
else if(art->vendor == CPU_VENDOR_AMD)
strcpy(tmp, AMD_ASCII);
else
strcpy(tmp, UNKNOWN_ASCII);
#elif ARCH_ARM
if(art->vendor == SOC_VENDOR_SNAPDRAGON)
strcpy(tmp, SNAPDRAGON_ASCII);
else if(art->vendor == SOC_VENDOR_MEDIATEK)
strcpy(tmp, MEDIATEK_ASCII);
else if(art->vendor == SOC_VENDOR_EXYNOS)
strcpy(tmp, EXYNOS_ASCII);
else if(art->vendor == SOC_VENDOR_KIRIN)
strcpy(tmp, KIRIN_ASCII);
else if(art->vendor == SOC_VENDOR_BROADCOM)
strcpy(tmp, BROADCOM_ASCII);
else
strcpy(tmp, ARM_ASCII);
#endif
for(int i=0; i < NUMBER_OF_LINES; i++)
memcpy(art->art[i], tmp + i*LINE_SIZE, LINE_SIZE);
return art;
}
uint32_t longest_attribute_length(struct ascii* art) {
uint32_t max = 0;
uint64_t len = 0;
for(uint32_t i=0; i < art->n_attributes_set; i++) {
if(art->attributes[i]->value != NULL) {
len = strlen(ATTRIBUTE_FIELDS[art->attributes[i]->type]);
if(len > max) max = len;
}
}
return max;
}
#ifdef ARCH_X86
void print_algorithm_intel(struct ascii* art, int n, bool* flag) {
for(int i=0; i < LINE_SIZE; i++) {
if(*flag) {
if(art->art[n][i] == ' ') {
*flag = false;
printf("%s%c%s", art->color2_ascii, art->ascii_chars[1], art->reset);
}
else {
printf("%s%c%s", art->color1_ascii, art->ascii_chars[0], art->reset);
}
}
else {
if(art->art[n][i] != ' ' && art->art[n][i] != '\0') {
*flag = true;
printf("%c",' ');
}
else {
printf("%c",' ');
}
}
}
}
void print_algorithm_amd(struct ascii* art, int n, bool* flag) {
*flag = false; // dummy, just silence compiler error
for(int i=0; i < LINE_SIZE; i++) {
if(art->art[n][i] == '@')
printf("%s%c%s", art->color1_ascii, art->ascii_chars[0], art->reset);
else if(art->art[n][i] == '#')
printf("%s%c%s", art->color2_ascii, art->ascii_chars[1], art->reset);
else
printf("%c",art->art[n][i]);
}
}
void print_ascii_x86(struct ascii* art, uint32_t la, void (*callback_print_algorithm)(struct ascii* art, int i, bool* flag)) {
int attr_to_print = 0;
int attr_type;
char* attr_value;
uint32_t space_right;
uint32_t space_up = (NUMBER_OF_LINES - art->n_attributes_set)/2;
uint32_t space_down = NUMBER_OF_LINES - art->n_attributes_set - space_up;
bool flag = false;
printf("\n");
for(uint32_t n=0;n<NUMBER_OF_LINES;n++) {
callback_print_algorithm(art, n, &flag);
if(n > space_up-1 && n < NUMBER_OF_LINES-space_down) {
attr_type = art->attributes[attr_to_print]->type;
attr_value = art->attributes[attr_to_print]->value;
attr_to_print++;
space_right = 1 + (la - strlen(ATTRIBUTE_FIELDS[attr_type]));
printf("%s%s%s%*s%s%s%s\n", art->color1_text, ATTRIBUTE_FIELDS[attr_type], art->reset, space_right, "", art->color2_text, attr_value, art->reset);
}
else printf("\n");
}
printf("\n");
}
void print_ascii(struct ascii* art) {
uint32_t longest_attribute = longest_attribute_length(art);
if(art->vendor == CPU_VENDOR_INTEL)
print_ascii_x86(art, longest_attribute, &print_algorithm_intel);
else if(art->vendor == CPU_VENDOR_AMD)
print_ascii_x86(art, longest_attribute, &print_algorithm_amd);
else {
printBug("Invalid CPU vendor: %d\n", art->vendor);
}
}
bool print_cpufetch_x86(struct cpuInfo* cpu, STYLE s, struct colors* cs) {
struct ascii* art = set_ascii(get_cpu_vendor(cpu), s, cs);
if(art == NULL)
return false;
char* uarch = get_str_uarch(cpu);
char* manufacturing_process = get_str_process(cpu);
char* sockets = get_str_sockets(cpu->topo);
char* max_frequency = get_str_freq(cpu->freq);
char* n_cores = get_str_topology(cpu, cpu->topo, false);
char* n_cores_dual = get_str_topology(cpu, cpu->topo, true);
char* cpu_name = get_str_cpu_name(cpu);
char* avx = get_str_avx(cpu);
char* fma = get_str_fma(cpu);
char* l1i = get_str_l1i(cpu->cach);
char* l1d = get_str_l1d(cpu->cach);
char* l2 = get_str_l2(cpu->cach);
char* l3 = get_str_l3(cpu->cach);
char* pp = get_str_peak_performance(cpu,cpu->topo,get_freq(cpu->freq));
setAttribute(art,ATTRIBUTE_NAME,cpu_name);
if(cpu->hv->present) {
setAttribute(art, ATTRIBUTE_HYPERVISOR, cpu->hv->hv_name);
}
setAttribute(art,ATTRIBUTE_UARCH,uarch);
setAttribute(art,ATTRIBUTE_TECHNOLOGY,manufacturing_process);
setAttribute(art,ATTRIBUTE_FREQUENCY,max_frequency);
uint32_t socket_num = get_nsockets(cpu->topo);
if (socket_num > 1) {
setAttribute(art, ATTRIBUTE_SOCKETS, sockets);
setAttribute(art, ATTRIBUTE_NCORES,n_cores);
setAttribute(art, ATTRIBUTE_NCORES_DUAL, n_cores_dual);
}
else {
setAttribute(art,ATTRIBUTE_NCORES,n_cores);
}
setAttribute(art,ATTRIBUTE_AVX,avx);
setAttribute(art,ATTRIBUTE_FMA,fma);
setAttribute(art,ATTRIBUTE_L1i,l1i);
setAttribute(art,ATTRIBUTE_L1d,l1d);
setAttribute(art,ATTRIBUTE_L2,l2);
if(l3 != NULL) {
setAttribute(art,ATTRIBUTE_L3,l3);
}
setAttribute(art,ATTRIBUTE_PEAK,pp);
if(art->n_attributes_set > NUMBER_OF_LINES) {
printBug("The number of attributes set is bigger than the max that can be displayed");
return false;
}
print_ascii(art);
free(manufacturing_process);
free(max_frequency);
free(sockets);
free(n_cores);
free(n_cores_dual);
free(avx);
free(fma);
free(l1i);
free(l1d);
free(l2);
free(l3);
free(pp);
free(art->attributes);
free(art);
if(cs != NULL) free_colors_struct(cs);
free_cache_struct(cpu->cach);
free_topo_struct(cpu->topo);
free_freq_struct(cpu->freq);
free_cpuinfo_struct(cpu);
return true;
}
#endif
#ifdef ARCH_ARM
void print_algorithm_snapd_mtk(struct ascii* art, int n) {
for(int i=0; i < LINE_SIZE; i++) {
if(art->art[n][i] == '@')
printf("%s%c%s", art->color1_ascii, art->ascii_chars[0], art->reset);
else if(art->art[n][i] == '#')
printf("%s%c%s", art->color2_ascii, art->ascii_chars[1], art->reset);
else
printf("%c",art->art[n][i]);
}
}
void print_algorithm_samsung(struct ascii* art, int n) {
int y_margin = 2;
int x_margin = 2 * y_margin;
for(int i=0; i < LINE_SIZE; i++) {
if(art->art[n][i] == '#') {
printf("%s%c%s", art->color1_ascii, art->ascii_chars[0], art->reset);
}
else if((n >= y_margin && n < NUMBER_OF_LINES-y_margin) && (i >= x_margin && i < LINE_SIZE-x_margin)) {
if(art->art[n][i] == '#')
printf("%s%c%s", art->color1_ascii, art->ascii_chars[0], art->reset);
else
printf("%s%c%s","\x1b[48;2;10;10;10m" COLOR_FG_WHITE, art->art[n][i], art->reset);
}
else
printf("%c", art->art[n][i]);
}
}
void print_algorithm_arm(struct ascii* art, int n) {
for(int i=0; i < LINE_SIZE; i++) {
if(art->art[n][i] == '#')
printf("%s%c%s", art->color1_ascii, art->ascii_chars[0], art->reset);
else
printf("%c",art->art[n][i]);
}
}
void print_ascii_arm(struct ascii* art, uint32_t la, void (*callback_print_algorithm)(struct ascii* art, int n)) {
int attr_to_print = 0;
int attr_type;
char* attr_value;
uint32_t limit_up;
uint32_t limit_down;
uint32_t space_right;
uint32_t space_up = (NUMBER_OF_LINES - art->n_attributes_set)/2;
uint32_t space_down = NUMBER_OF_LINES - art->n_attributes_set - space_up;
if(art->n_attributes_set > NUMBER_OF_LINES) {
limit_up = 0;
limit_down = art->n_attributes_set;
}
else {
limit_up = space_up;
limit_down = NUMBER_OF_LINES-space_down;
}
bool add_space = false;
uint32_t len = max(art->n_attributes_set, NUMBER_OF_LINES);
for(uint32_t n=0; n < len; n++) {
if(n >= art->additional_spaces && n < NUMBER_OF_LINES + art->additional_spaces)
callback_print_algorithm(art, n - art->additional_spaces);
else
printf("%*s", LINE_SIZE, "");
if(n >= limit_up && n < limit_down) {
attr_type = art->attributes[attr_to_print]->type;
attr_value = art->attributes[attr_to_print]->value;
attr_to_print++;
if(attr_type == ATTRIBUTE_PEAK) {
add_space = false;
}
if(attr_type == ATTRIBUTE_CPU_NUM) {
printf("%s%s%s\n", art->color1_text, attr_value, art->reset);
add_space = true;
}
else {
if(add_space) {
space_right = 1 + (la - strlen(ATTRIBUTE_FIELDS[attr_type]));
printf(" %s%s%s%*s%s%s%s\n", art->color1_text, ATTRIBUTE_FIELDS[attr_type], art->reset, space_right, "", art->color2_text, attr_value, art->reset);
}
else {
space_right = 2 + 1 + (la - strlen(ATTRIBUTE_FIELDS[attr_type]));
printf("%s%s%s%*s%s%s%s\n", art->color1_text, ATTRIBUTE_FIELDS[attr_type], art->reset, space_right, "", art->color2_text, attr_value, art->reset);
}
}
}
else printf("\n");
}
}
void print_ascii(struct ascii* art) {
uint32_t longest_attribute = longest_attribute_length(art);
if(art->vendor == SOC_VENDOR_SNAPDRAGON || art->vendor == SOC_VENDOR_MEDIATEK || art->vendor == SOC_VENDOR_KIRIN || art->vendor == SOC_VENDOR_BROADCOM)
print_ascii_arm(art, longest_attribute, &print_algorithm_snapd_mtk);
else if(art->vendor == SOC_VENDOR_EXYNOS)
print_ascii_arm(art, longest_attribute, &print_algorithm_samsung);
else {
if(art->vendor != SOC_VENDOR_UNKNOWN)
printWarn("Invalid SOC vendor: %d\n", art->vendor);
print_ascii_arm(art, longest_attribute, &print_algorithm_arm);
}
}
bool print_cpufetch_arm(struct cpuInfo* cpu, STYLE s, struct colors* cs) {
struct ascii* art = set_ascii(get_soc_vendor(cpu->soc), s, cs);
if(art == NULL)
return false;
char* manufacturing_process = get_str_process(cpu->soc);
char* soc_name = get_soc_name(cpu->soc);
char* features = get_str_features(cpu);
setAttribute(art,ATTRIBUTE_SOC,soc_name);
setAttribute(art,ATTRIBUTE_TECHNOLOGY,manufacturing_process);
if(cpu->num_cpus == 1) {
char* uarch = get_str_uarch(cpu);
char* max_frequency = get_str_freq(cpu->freq);
char* n_cores = get_str_topology(cpu, cpu->topo, false);
/*
* char* l1i = get_str_l1i(cpu->cach);
* char* l1d = get_str_l1d(cpu->cach);
* char* l2 = get_str_l2(cpu->cach);
* char* l3 = get_str_l3(cpu->cach);
* Do not setAttribute for caches.
* Cache functionality may be implemented
* in the future
*/
setAttribute(art,ATTRIBUTE_UARCH,uarch);
setAttribute(art,ATTRIBUTE_FREQUENCY,max_frequency);
setAttribute(art,ATTRIBUTE_NCORES,n_cores);
if(features != NULL) {
setAttribute(art, ATTRIBUTE_FEATURES, features);
}
}
else {
struct cpuInfo* ptr = cpu;
for(int i = 0; i < cpu->num_cpus; ptr = ptr->next_cpu, i++) {
char* uarch = get_str_uarch(ptr);
char* max_frequency = get_str_freq(ptr->freq);
char* n_cores = get_str_topology(ptr, ptr->topo, false);
/*
* char* l1i = get_str_l1i(cpu->cach);
* char* l1d = get_str_l1d(cpu->cach);
* char* l2 = get_str_l2(cpu->cach);
* char* l3 = get_str_l3(cpu->cach);
* Do not setAttribute for caches.
* Cache functionality may be implemented
* in the future
*/
char* cpu_num = malloc(sizeof(char) * 9);
sprintf(cpu_num, "CPU %d:", i+1);
setAttribute(art, ATTRIBUTE_CPU_NUM, cpu_num);
setAttribute(art, ATTRIBUTE_UARCH, uarch);
setAttribute(art, ATTRIBUTE_FREQUENCY, max_frequency);
setAttribute(art, ATTRIBUTE_NCORES, n_cores);
if(features != NULL) {
setAttribute(art, ATTRIBUTE_FEATURES, features);
}
}
}
char* pp = get_str_peak_performance(cpu);
setAttribute(art,ATTRIBUTE_PEAK,pp);
if(art->n_attributes_set > NUMBER_OF_LINES) {
art->additional_spaces = (art->n_attributes_set - NUMBER_OF_LINES) / 2;
}
if(cpu->hv->present)
setAttribute(art, ATTRIBUTE_HYPERVISOR, cpu->hv->hv_name);
print_ascii(art);
free(manufacturing_process);
free(pp);
free(art->attributes);
free(art);
if(cs != NULL) free_colors_struct(cs);
free_cache_struct(cpu->cach);
free_topo_struct(cpu->topo);
free_cpuinfo_struct(cpu);
return true;
}
#endif
bool print_cpufetch(struct cpuInfo* cpu, STYLE s, struct colors* cs) {
// Sanity check of ASCII arts
int len = sizeof(ASCII_ARRAY) / sizeof(ASCII_ARRAY[0]);
for(int i=0; i < len; i++) {
const char* ascii = ASCII_ARRAY[i];
if(strlen(ascii) != (NUMBER_OF_LINES * LINE_SIZE)) {
printBug("ASCII art %d is wrong! ASCII length: %d, expected length: %d", i, strlen(ascii), (NUMBER_OF_LINES * LINE_SIZE));
return false;
}
}
#ifdef ARCH_X86
return print_cpufetch_x86(cpu, s, cs);
#elif ARCH_ARM
return print_cpufetch_arm(cpu, s, cs);
#endif
}

24
src/common/printer.h Normal file
View File

@@ -0,0 +1,24 @@
#ifndef __PRINTER__
#define __PRINTER__
typedef int STYLE;
#include "args.h"
#ifdef ARCH_X86
#include "../x86/cpuid.h"
#else
#include "../arm/midr.h"
#endif
#define COLOR_DEFAULT_INTEL "15,125,194:230,230,230:40,150,220:230,230,230"
#define COLOR_DEFAULT_AMD "250,250,250:0,154,102:250,250,250:0,154,102"
#define COLOR_DEFAULT_ARM "0,145,189:0,145,189:240,240,240:0,145,189"
#ifdef ARCH_X86
void print_levels(struct cpuInfo* cpu);
#endif
bool print_cpufetch(struct cpuInfo* cpu, STYLE s, struct colors* cs);
#endif

76
src/common/udev.c Normal file
View File

@@ -0,0 +1,76 @@
#include "udev.h"
#include "global.h"
#include "cpu.h"
char* read_file(char* path, int* len) {
int fd = open(path, O_RDONLY);
if(fd == -1) {
return NULL;
}
//File exists, read it
int bytes_read = 0;
int offset = 0;
int block = 128;
char* buf = malloc(sizeof(char)*DEFAULT_FILE_SIZE);
memset(buf, 0, sizeof(char)*DEFAULT_FILE_SIZE);
while ( (bytes_read = read(fd, buf+offset, block)) > 0 ) {
offset += bytes_read;
}
if (close(fd) == -1) {
return NULL;
}
*len = offset;
return buf;
}
long get_freq_from_file(char* path, bool hv_present) {
int filelen;
char* buf;
if((buf = read_file(path, &filelen)) == NULL) {
if(hv_present)
printWarn("Could not open '%s' (HV is present)", path);
else
printWarn("Could not open '%s'", path);
return UNKNOWN_FREQ;
}
char* end;
errno = 0;
long ret = strtol(buf, &end, 10);
if(errno != 0) {
perror("strtol");
printBug("Failed parsing '%s' file. Read data was: '%s'", path, buf);
free(buf);
return UNKNOWN_FREQ;
}
// We will be getting the frequency in KHz
// We consider it is an error if frequency is
// greater than 10 GHz or less than 100 MHz
if(ret > 10000 * 1000 || ret < 100 * 1000) {
printBug("Invalid data was read from file '%s': %ld\n", path, ret);
return UNKNOWN_FREQ;
}
free(buf);
return ret/1000;
}
long get_max_freq_from_file(uint32_t core, bool hv_present) {
char path[_PATH_FREQUENCY_MAX_LEN];
sprintf(path, "%s%s/cpu%d%s%s", _PATH_SYS_SYSTEM, _PATH_SYS_CPU, core, _PATH_FREQUENCY, _PATH_FREQUENCY_MAX);
return get_freq_from_file(path, hv_present);
}
long get_min_freq_from_file(uint32_t core, bool hv_present) {
char path[_PATH_FREQUENCY_MAX_LEN];
sprintf(path, "%s%s/cpu%d%s%s", _PATH_SYS_SYSTEM, _PATH_SYS_CPU, core, _PATH_FREQUENCY, _PATH_FREQUENCY_MIN);
return get_freq_from_file(path, hv_present);
}

28
src/common/udev.h Normal file
View File

@@ -0,0 +1,28 @@
#ifndef __UDEV__
#define __UDEV__
#include <stdio.h>
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdbool.h>
#include <unistd.h>
#include <fcntl.h>
#include <errno.h>
#include "cpu.h"
#define _PATH_SYS_SYSTEM "/sys/devices/system"
#define _PATH_SYS_CPU "/cpu"
#define _PATH_FREQUENCY "/cpufreq"
#define _PATH_FREQUENCY_MAX "/cpuinfo_max_freq"
#define _PATH_FREQUENCY_MIN "/cpuinfo_min_freq"
#define _PATH_FREQUENCY_MAX_LEN 100
#define DEFAULT_FILE_SIZE 4096
char* read_file(char* path, int* len);
long get_max_freq_from_file(uint32_t core, bool hv_present);
long get_min_freq_from_file(uint32_t core, bool hv_present);
#endif

View File

@@ -1,875 +0,0 @@
#ifdef _WIN32
#include <windows.h>
#else
#include "udev.h"
#include <unistd.h>
#endif
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <stdbool.h>
#include "cpuid.h"
#include "cpuid_asm.h"
#include "global.h"
#include "apic.h"
#define VENDOR_INTEL_STRING "GenuineIntel"
#define VENDOR_AMD_STRING "AuthenticAMD"
#define STRING_YES "Yes"
#define STRING_NO "No"
#define STRING_UNKNOWN "Unknown"
#define STRING_NONE "None"
#define STRING_MEGAHERZ "MHz"
#define STRING_GIGAHERZ "GHz"
#define STRING_KILOBYTES "KB"
#define STRING_MEGABYTES "MB"
#define CPU_NAME_MAX_LENGTH 64
#define MASK 0xFF
/*
* cpuid reference: http://www.sandpile.org/x86/cpuid.htm
* cpuid amd: https://www.amd.com/system/files/TechDocs/25481.pdf
*/
struct cpuInfo {
bool AVX;
bool AVX2;
bool AVX512;
bool SSE;
bool SSE2;
bool SSE3;
bool SSSE3;
bool SSE4a;
bool SSE4_1;
bool SSE4_2;
bool FMA3;
bool FMA4;
bool AES;
bool SHA;
VENDOR cpu_vendor;
char* cpu_name;
// Max cpuids levels
uint32_t maxLevels;
// Max cpuids extended levels
uint32_t maxExtendedLevels;
};
struct cache {
int32_t L1i;
int32_t L1d;
int32_t L2;
int32_t L3;
};
struct frequency {
int64_t base;
int64_t max;
};
void init_cpu_info(struct cpuInfo* cpu) {
cpu->AVX = false;
cpu->AVX2 = false;
cpu->AVX512 = false;
cpu->SSE = false;
cpu->SSE2 = false;
cpu->SSE3 = false;
cpu->SSSE3 = false;
cpu->SSE4a = false;
cpu->SSE4_1 = false;
cpu->SSE4_2 = false;
cpu->FMA3 = false;
cpu->FMA4 = false;
cpu->AES = false;
cpu->SHA = false;
}
void init_topology_struct(struct topology** topo) {
(*topo)->total_cores = 0;
(*topo)->physical_cores = 0;
(*topo)->logical_cores = 0;
(*topo)->smt_available = 0;
(*topo)->smt_supported = 0;
(*topo)->sockets = 0;
}
void get_cpu_vendor_internal(char* name, uint32_t ebx,uint32_t ecx,uint32_t edx) {
name[__COUNTER__] = ebx & MASK;
name[__COUNTER__] = (ebx>>8) & MASK;
name[__COUNTER__] = (ebx>>16) & MASK;
name[__COUNTER__] = (ebx>>24) & MASK;
name[__COUNTER__] = edx & MASK;
name[__COUNTER__] = (edx>>8) & MASK;
name[__COUNTER__] = (edx>>16) & MASK;
name[__COUNTER__] = (edx>>24) & MASK;
name[__COUNTER__] = ecx & MASK;
name[__COUNTER__] = (ecx>>8) & MASK;
name[__COUNTER__] = (ecx>>16) & MASK;
name[__COUNTER__] = (ecx>>24) & MASK;
}
char* get_str_cpu_name_internal() {
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
uint32_t c = 0;
char * name = malloc(sizeof(char) * CPU_NAME_MAX_LENGTH);
memset(name, 0, CPU_NAME_MAX_LENGTH);
for(int i=0; i < 3; i++) {
eax = 0x80000002 + i;
cpuid(&eax, &ebx, &ecx, &edx);
name[c++] = eax & MASK;
name[c++] = (eax>>8) & MASK;
name[c++] = (eax>>16) & MASK;
name[c++] = (eax>>24) & MASK;
name[c++] = ebx & MASK;
name[c++] = (ebx>>8) & MASK;
name[c++] = (ebx>>16) & MASK;
name[c++] = (ebx>>24) & MASK;
name[c++] = ecx & MASK;
name[c++] = (ecx>>8) & MASK;
name[c++] = (ecx>>16) & MASK;
name[c++] = (ecx>>24) & MASK;
name[c++] = edx & MASK;
name[c++] = (edx>>8) & MASK;
name[c++] = (edx>>16) & MASK;
name[c++] = (edx>>24) & MASK;
}
name[c] = '\0';
//Remove unused characters
char *str = name;
char *dest = name;
// Remove spaces before name
while (*str != '\0' && *str == ' ')str++;
// Remove spaces between the name and after it
while (*str != '\0') {
while (*str == ' ' && *(str + 1) == ' ') str++;
*dest++ = *str++;
}
*dest = '\0';
return name;
}
struct cpuInfo* get_cpu_info() {
struct cpuInfo* cpu = malloc(sizeof(struct cpuInfo));
init_cpu_info(cpu);
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
//Get max cpuid level
cpuid(&eax, &ebx, &ecx, &edx);
cpu->maxLevels = eax;
//Fill vendor
char name[13];
memset(name,0,13);
get_cpu_vendor_internal(name, ebx, ecx, edx);
if(strcmp(VENDOR_INTEL_STRING,name) == 0)
cpu->cpu_vendor = VENDOR_INTEL;
else if (strcmp(VENDOR_AMD_STRING,name) == 0)
cpu->cpu_vendor = VENDOR_AMD;
else {
cpu->cpu_vendor = VENDOR_INVALID;
printErr("Unknown CPU vendor: %s", name);
return NULL;
}
//Get max extended level
eax = 0x80000000;
ebx = 0;
ecx = 0;
edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
cpu->maxExtendedLevels = eax;
//Fill instructions support
if (cpu->maxLevels >= 0x00000001){
eax = 0x00000001;
cpuid(&eax, &ebx, &ecx, &edx);
cpu->SSE = (edx & ((int)1 << 25)) != 0;
cpu->SSE2 = (edx & ((int)1 << 26)) != 0;
cpu->SSE3 = (ecx & ((int)1 << 0)) != 0;
cpu->SSSE3 = (ecx & ((int)1 << 9)) != 0;
cpu->SSE4_1 = (ecx & ((int)1 << 19)) != 0;
cpu->SSE4_2 = (ecx & ((int)1 << 20)) != 0;
cpu->AES = (ecx & ((int)1 << 25)) != 0;
cpu->AVX = (ecx & ((int)1 << 28)) != 0;
cpu->FMA3 = (ecx & ((int)1 << 12)) != 0;
}
else {
printWarn("Can't read features information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x00000001, cpu->maxLevels);
}
if (cpu->maxLevels >= 0x00000007){
eax = 0x00000007;
ecx = 0x00000000;
cpuid(&eax, &ebx, &ecx, &edx);
cpu->AVX2 = (ebx & ((int)1 << 5)) != 0;
cpu->SHA = (ebx & ((int)1 << 29)) != 0;
cpu->AVX512 = (((ebx & ((int)1 << 16)) != 0) ||
((ebx & ((int)1 << 28)) != 0) ||
((ebx & ((int)1 << 26)) != 0) ||
((ebx & ((int)1 << 27)) != 0) ||
((ebx & ((int)1 << 31)) != 0) ||
((ebx & ((int)1 << 30)) != 0) ||
((ebx & ((int)1 << 17)) != 0) ||
((ebx & ((int)1 << 21)) != 0));
}
else {
printWarn("Can't read features information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x00000007, cpu->maxLevels);
}
if (cpu->maxExtendedLevels >= 0x80000001){
eax = 0x80000001;
cpuid(&eax, &ebx, &ecx, &edx);
cpu->SSE4a = (ecx & ((int)1 << 6)) != 0;
cpu->FMA4 = (ecx & ((int)1 << 16)) != 0;
}
else {
printWarn("Can't read features information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x80000001, cpu->maxExtendedLevels);
}
if (cpu->maxExtendedLevels >= 0x80000004){
cpu->cpu_name = get_str_cpu_name_internal();
}
else {
cpu->cpu_name = malloc(sizeof(char)*8);
sprintf(cpu->cpu_name,"Unknown");
printWarn("Can't read cpu name from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x80000004, cpu->maxExtendedLevels);
}
return cpu;
}
// Main reference: https://software.intel.com/content/www/us/en/develop/articles/intel-64-architecture-processor-topology-enumeration.html
// Very interesting resource: https://wiki.osdev.org/Detecting_CPU_Topology_(80x86)
struct topology* get_topology_info(struct cpuInfo* cpu) {
struct topology* topo = malloc(sizeof(struct topology));
topo->apic = malloc(sizeof(struct apic));
init_topology_struct(&topo);
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
// Ask the OS the total number of cores it sees
// If we have one socket, it will be same as the cpuid,
// but in dual socket it will not!
// TODO: Replace by apic?
#ifdef _WIN32
SYSTEM_INFO info;
GetSystemInfo(&info);
topo->total_cores = info.dwNumberOfProcessors;
#else
if((topo->total_cores = sysconf(_SC_NPROCESSORS_ONLN)) == -1) {
perror("sysconf");
topo->total_cores = topo->logical_cores; // fallback
}
#endif
switch(cpu->cpu_vendor) {
case VENDOR_INTEL:
if (cpu->maxLevels >= 0x00000004) {
get_topology_from_apic(cpu->maxLevels, &topo);
}
else {
printErr("Can't read topology information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x00000001, cpu->maxLevels);
topo->physical_cores = 1;
topo->logical_cores = 1;
topo->smt_available = 1;
topo->smt_supported = 1;
}
break;
case VENDOR_AMD:
if (cpu->maxExtendedLevels >= 0x80000008) {
eax = 0x80000008;
cpuid(&eax, &ebx, &ecx, &edx);
topo->logical_cores = (ecx & 0xFF) + 1;
if (cpu->maxExtendedLevels >= 0x8000001E) {
eax = 0x8000001E;
cpuid(&eax, &ebx, &ecx, &edx);
topo->smt_supported = ((ebx >> 8) & 0x03) + 1;
}
else {
printWarn("Can't read topology information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x8000001E, cpu->maxExtendedLevels);
topo->smt_supported = 1;
}
}
else {
printErr("Can't read topology information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x80000008, cpu->maxExtendedLevels);
topo->physical_cores = 1;
topo->logical_cores = 1;
topo->smt_supported = 1;
}
if (cpu->maxLevels >= 0x0000000B) {
topo->smt_available = is_smt_enabled(topo);
}
else {
printWarn("Can't read topology information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x0000000B, cpu->maxLevels);
topo->smt_available = 1;
}
topo->physical_cores = topo->logical_cores / topo->smt_available;
if(topo->smt_supported > 1)
topo->sockets = topo->total_cores / topo->smt_supported / topo->physical_cores; // Idea borrowed from lscpu
else
topo->sockets = topo->total_cores / topo->physical_cores;
break;
default:
printBug("Cant get topology because VENDOR is empty");
return NULL;
}
return topo;
}
struct cache* get_cache_info(struct cpuInfo* cpu) {
struct cache* cach = malloc(sizeof(struct cache));
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
uint32_t level;
// We use standart 0x00000004 for Intel
// We use extended 0x8000001D for AMD
if(cpu->cpu_vendor == VENDOR_INTEL) {
level = 0x00000004;
if(cpu->maxLevels < level) {
printErr("Can't read cache information from cpuid (needed level is %d, max is %d)", level, cpu->maxLevels);
return NULL;
}
}
else {
level = 0x8000001D;
if(cpu->maxExtendedLevels < level) {
printErr("Can't read cache information from cpuid (needed extended level is %d, max is %d)", level, cpu->maxExtendedLevels);
return NULL;
}
}
// We suppose there are 4 caches (at most)
for(int i=0; i < 4; i++) {
eax = level; // get cache info
ebx = 0;
ecx = i; // cache id
edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
int32_t cache_type = eax & 0x1F;
// If its 0, we tried fetching a non existing cache
if (cache_type > 0) {
int32_t cache_level = (eax >>= 5) & 0x7;
uint32_t cache_sets = ecx + 1;
uint32_t cache_coherency_line_size = (ebx & 0xFFF) + 1;
uint32_t cache_physical_line_partitions = ((ebx >>= 12) & 0x3FF) + 1;
uint32_t cache_ways_of_associativity = ((ebx >>= 10) & 0x3FF) + 1;
int32_t cache_total_size = cache_ways_of_associativity * cache_physical_line_partitions * cache_coherency_line_size * cache_sets;
switch (cache_type) {
case 1: // Data Cache (We assume this is L1d)
if(cache_level != 1) {
printBug("Found data cache at level %d (expected 1)", cache_level);
return NULL;
}
cach->L1d = cache_total_size;
break;
case 2: // Instruction Cache (We assume this is L1i)
if(cache_level != 1) {
printBug("Found instruction cache at level %d (expected 1)", cache_level);
return NULL;
}
cach->L1i = cache_total_size;
break;
case 3: // Unified Cache (This may be L2 or L3)
if(cache_level == 2) cach->L2 = cache_total_size;
else if(cache_level == 3) cach->L3 = cache_total_size;
else {
printBug("Found unified cache at level %d (expected == 2 or 3)", cache_level);
return NULL;
}
break;
default: // Unknown Type Cache
printBug("Unknown Type Cache found at ID %d", i);
return NULL;
}
}
else if(i == 2) cach->L2 = UNKNOWN;
else if(i == 3) cach->L3 = UNKNOWN;
else {
printBug("Could not find cache ID %d", i);
return NULL;
}
}
// Sanity checks. If we read values greater than this, they can't be valid ones
// The values were chosen by me
if(cach->L1i > 64 * 1024) {
printBug("Invalid L1i size: %dKB", cach->L1i/1024);
return NULL;
}
if(cach->L1d > 64 * 1024) {
printBug("Invalid L1d size: %dKB", cach->L1d/1024);
return NULL;
}
if(cach->L2 != UNKNOWN) {
if(cach->L3 != UNKNOWN && cach->L2 > 2 * 1048576) {
printBug("Invalid L2 size: %dMB", cach->L2/(1048576));
return NULL;
}
else if(cach->L2 > 100 * 1048576) {
printBug("Invalid L2 size: %dMB", cach->L2/(1048576));
return NULL;
}
}
if(cach->L3 != UNKNOWN && cach->L3 > 100 * 1048576) {
printBug("Invalid L3 size: %dMB", cach->L3/(1048576));
return NULL;
}
if(cach->L2 == UNKNOWN) {
printBug("Could not find L2 cache");
return NULL;
}
return cach;
}
struct frequency* get_frequency_info(struct cpuInfo* cpu) {
struct frequency* freq = malloc(sizeof(struct frequency));
if(cpu->maxLevels < 0x16) {
#ifdef _WIN32
printErr("Can't read frequency information from cpuid (needed level is %d, max is %d)", 0x16, cpu->maxLevels);
freq->base = UNKNOWN;
freq->max = UNKNOWN;
#else
printWarn("Can't read frequency information from cpuid (needed level is %d, max is %d). Using udev", 0x16, cpu->maxLevels);
freq->base = UNKNOWN;
freq->max = get_max_freq_from_file();
#endif
}
else {
uint32_t eax = 0x16;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
freq->base = eax;
freq->max = ebx;
}
return freq;
}
uint32_t get_nsockets(struct topology* topo) {
return topo->sockets;
}
int64_t get_freq(struct frequency* freq) {
return freq->max;
}
VENDOR get_cpu_vendor(struct cpuInfo* cpu) {
return cpu->cpu_vendor;
}
void debug_cpu_info(struct cpuInfo* cpu) {
printf("AVX=%s\n", cpu->AVX ? "true" : "false");
printf("AVX2=%s\n", cpu->AVX2 ? "true" : "false");
printf("AVX512=%s\n\n", cpu->AVX512 ? "true" : "false");
printf("SSE=%s\n", cpu->SSE ? "true" : "false");
printf("SSE2=%s\n", cpu->SSE2 ? "true" : "false");
printf("SSE3=%s\n", cpu->SSE3 ? "true" : "false");
printf("SSSE3=%s\n", cpu->SSSE3 ? "true" : "false");
printf("SSE4a=%s\n", cpu->SSE4a ? "true" : "false");
printf("SSE4_1=%s\n", cpu->SSE4_1 ? "true" : "false");
printf("SSE4_2=%s\n\n", cpu->SSE4_2 ? "true" : "false");
printf("FMA3=%s\n", cpu->FMA3 ? "true" : "false");
printf("FMA4=%s\n\n", cpu->FMA4 ? "true" : "false");
printf("AES=%s\n", cpu->AES ? "true" : "false");
printf("SHA=%s\n", cpu->SHA ? "true" : "false");
}
void debug_cache(struct cache* cach) {
printf("L1i=%dB\n",cach->L1i);
printf("L1d=%dB\n",cach->L1d);
printf("L2=%dB\n",cach->L2);
printf("L3=%dB\n",cach->L3);
}
void debug_frequency(struct frequency* freq) {
#ifdef _WIN32
printf("maxf=%I64d Mhz\n",freq->max);
printf("basef=%I64d Mhz\n",freq->base);
#else
printf("maxf=%ld Mhz\n",freq->max);
printf("basef=%ld Mhz\n",freq->base);
#endif
}
/*** STRING FUNCTIONS ***/
char* get_str_peak_performance(struct cpuInfo* cpu, struct topology* topo, int64_t freq) {
/***
PP = PeakPerformance
SP = SinglePrecision
PP(SP) =
N_CORES *
FREQUENCY *
2(Two vector units) *
2(If cpu has fma) *
16(If AVX512), 8(If AVX), 4(If SSE) *
***/
//7 for GFLOP/s and 6 for digits,eg 412.14
uint32_t size = 7+6+1+1;
assert(strlen(STRING_UNKNOWN)+1 <= size);
char* string = malloc(sizeof(char)*size);
//First check we have consistent data
if(freq == UNKNOWN) {
snprintf(string,strlen(STRING_UNKNOWN)+1,STRING_UNKNOWN);
return string;
}
double flops = topo->physical_cores*(freq*1000000);
// Intel USUALLY has two VPUs. I have never seen an AMD
// with two VPUs.
if(cpu->cpu_vendor == VENDOR_INTEL) flops = flops * 2;
if(cpu->FMA3 || cpu->FMA4)
flops = flops*2;
if(cpu->AVX512)
flops = flops*16;
else if(cpu->AVX || cpu->AVX2)
flops = flops*8;
else if(cpu->SSE)
flops = flops*4;
if(flops >= (double)1000000000000.0)
snprintf(string,size,"%.2f TFLOP/s",flops/1000000000000);
else if(flops >= 1000000000.0)
snprintf(string,size,"%.2f GFLOP/s",flops/1000000000);
else
snprintf(string,size,"%.2f MFLOP/s",flops/1000000);
return string;
}
// TODO: Refactoring
char* get_str_topology(struct cpuInfo* cpu, struct topology* topo, bool dual_socket) {
char* string;
if(topo->smt_supported > 1) {
//3 for digits, 21 for ' cores (SMT disabled)' which is the longest possible output
uint32_t size = 3+21+1;
string = malloc(sizeof(char)*size);
if(dual_socket) {
if(topo->smt_available > 1)
snprintf(string, size, "%d cores (%d threads)",topo->physical_cores * topo->sockets, topo->logical_cores * topo->sockets);
else {
if(cpu->cpu_vendor == VENDOR_AMD)
snprintf(string, size, "%d cores (SMT disabled)",topo->physical_cores * topo->sockets);
else
snprintf(string, size, "%d cores (HT disabled)",topo->physical_cores * topo->sockets);
}
}
else {
if(topo->smt_available > 1)
snprintf(string, size, "%d cores (%d threads)",topo->physical_cores,topo->logical_cores);
else {
if(cpu->cpu_vendor == VENDOR_AMD)
snprintf(string, size, "%d cores (SMT disabled)",topo->physical_cores);
else
snprintf(string, size, "%d cores (HT disabled)",topo->physical_cores);
}
}
}
else {
uint32_t size = 3+7+1;
string = malloc(sizeof(char)*size);
if(dual_socket)
snprintf(string, size, "%d cores",topo->physical_cores * topo->sockets);
else
snprintf(string, size, "%d cores",topo->physical_cores);
}
return string;
}
char* get_str_sockets(struct topology* topo) {
char* string = malloc(sizeof(char) * 2);
int32_t sanity_ret = snprintf(string, 2, "%d", topo->sockets);
if(sanity_ret < 0) {
printBug("get_str_sockets: snprintf returned a negative value for input: '%d'", topo->sockets);
return NULL;
}
return string;
}
char* get_str_cpu_name(struct cpuInfo* cpu) {
return cpu->cpu_name;
}
char* get_str_avx(struct cpuInfo* cpu) {
//If all AVX are available, it will use up to 15
char* string = malloc(sizeof(char)*15+1);
if(!cpu->AVX)
snprintf(string,2+1,"No");
else if(!cpu->AVX2)
snprintf(string,3+1,"AVX");
else if(!cpu->AVX512)
snprintf(string,8+1,"AVX,AVX2");
else
snprintf(string,15+1,"AVX,AVX2,AVX512");
return string;
}
char* get_str_sse(struct cpuInfo* cpu) {
uint32_t last = 0;
uint32_t SSE_sl = 4;
uint32_t SSE2_sl = 5;
uint32_t SSE3_sl = 5;
uint32_t SSSE3_sl = 6;
uint32_t SSE4a_sl = 6;
uint32_t SSE4_1_sl = 7;
uint32_t SSE4_2_sl = 7;
char* string = malloc(sizeof(char)*SSE_sl+SSE2_sl+SSE3_sl+SSSE3_sl+SSE4a_sl+SSE4_1_sl+SSE4_2_sl+1);
if(cpu->SSE) {
snprintf(string+last,SSE_sl+1,"SSE,");
last+=SSE_sl;
}
if(cpu->SSE2) {
snprintf(string+last,SSE2_sl+1,"SSE2,");
last+=SSE2_sl;
}
if(cpu->SSE3) {
snprintf(string+last,SSE3_sl+1,"SSE3,");
last+=SSE3_sl;
}
if(cpu->SSSE3) {
snprintf(string+last,SSSE3_sl+1,"SSSE3,");
last+=SSSE3_sl;
}
if(cpu->SSE4a) {
snprintf(string+last,SSE4a_sl+1,"SSE4a,");
last+=SSE4a_sl;
}
if(cpu->SSE4_1) {
snprintf(string+last,SSE4_1_sl+1,"SSE4_1,");
last+=SSE4_1_sl;
}
if(cpu->SSE4_2) {
snprintf(string+last,SSE4_2_sl+1,"SSE4_2,");
last+=SSE4_2_sl;
}
//Purge last comma
string[last-1] = '\0';
return string;
}
char* get_str_fma(struct cpuInfo* cpu) {
char* string = malloc(sizeof(char)*9+1);
if(!cpu->FMA3)
snprintf(string,2+1,"No");
else if(!cpu->FMA4)
snprintf(string,4+1,"FMA3");
else
snprintf(string,9+1,"FMA3,FMA4");
return string;
}
char* get_str_aes(struct cpuInfo* cpu) {
char* string = malloc(sizeof(char)*3+1);
if(cpu->AES)
snprintf(string,3+1,STRING_YES);
else
snprintf(string,2+1,STRING_NO);
return string;
}
char* get_str_sha(struct cpuInfo* cpu) {
char* string = malloc(sizeof(char)*3+1);
if(cpu->SHA)
snprintf(string,3+1,STRING_YES);
else
snprintf(string,2+1,STRING_NO);
return string;
}
int32_t get_value_as_smallest_unit(char ** str, uint32_t value) {
int32_t sanity_ret;
*str = malloc(sizeof(char)* 11); //8 for digits, 2 for units
if(value/1024 >= 1024)
sanity_ret = snprintf(*str, 10,"%.4g"STRING_MEGABYTES, (double)value/(1<<20));
else
sanity_ret = snprintf(*str, 10,"%.4g"STRING_KILOBYTES, (double)value/(1<<10));
return sanity_ret;
}
// String functions
char* get_str_cache_two(int32_t cache_size, uint32_t physical_cores) {
// 4 for digits, 2 for units, 2 for ' (', 3 digits, 2 for units and 7 for ' Total)'
uint32_t max_size = 4+2 + 2 + 4+2 + 7 + 1;
int32_t sanity_ret;
char* string = malloc(sizeof(char) * max_size);
char* tmp1;
char* tmp2;
int32_t tmp1_len = get_value_as_smallest_unit(&tmp1, cache_size);
int32_t tmp2_len = get_value_as_smallest_unit(&tmp2, cache_size * physical_cores);
if(tmp1_len < 0) {
printBug("get_value_as_smallest_unit: snprintf returned a negative value for input: %d\n", cache_size);
return NULL;
}
if(tmp2_len < 0) {
printBug("get_value_as_smallest_unit: snprintf returned a negative value for input: %d\n", cache_size * physical_cores);
return NULL;
}
uint32_t size = tmp1_len + 2 + tmp2_len + 7 + 1;
sanity_ret = snprintf(string, size, "%s (%s Total)", tmp1, tmp2);
if(sanity_ret < 0) {
printBug("get_str_cache_two: snprintf returned a negative value for input: '%s' and '%s'\n", tmp1, tmp2);
return NULL;
}
free(tmp1);
free(tmp2);
return string;
}
char* get_str_cache_one(int32_t cache_size) {
// 4 for digits, 2 for units, 2 for ' (', 3 digits, 2 for units and 7 for ' Total)'
uint32_t max_size = 4+2 + 1;
int32_t sanity_ret;
char* string = malloc(sizeof(char) * max_size);
char* tmp;
int32_t tmp_len = get_value_as_smallest_unit(&tmp, cache_size);
if(tmp_len < 0) {
printBug("get_value_as_smallest_unit: snprintf returned a negative value for input: %d", cache_size);
return NULL;
}
uint32_t size = tmp_len + 1;
sanity_ret = snprintf(string, size, "%s", tmp);
if(sanity_ret < 0) {
printBug("get_str_cache_one: snprintf returned a negative value for input: '%s'", tmp);
return NULL;
}
free(tmp);
return string;
}
char* get_str_cache(int32_t cache_size, struct topology* topo, bool llc) {
if(topo->sockets == 1) {
if(llc)
return get_str_cache_one(cache_size);
else
return get_str_cache_two(cache_size, topo->physical_cores);
}
else {
if(llc)
return get_str_cache_two(cache_size, topo->sockets);
else
return get_str_cache_two(cache_size, topo->physical_cores * topo->sockets);
}
}
char* get_str_l1i(struct cache* cach, struct topology* topo) {
return get_str_cache(cach->L1i, topo, false);
}
char* get_str_l1d(struct cache* cach, struct topology* topo) {
return get_str_cache(cach->L1d, topo, false);
}
char* get_str_l2(struct cache* cach, struct topology* topo) {
assert(cach->L2 != UNKNOWN);
if(cach->L3 == UNKNOWN)
return get_str_cache(cach->L2, topo, true);
else
return get_str_cache(cach->L2, topo, false);
}
char* get_str_l3(struct cache* cach, struct topology* topo) {
if(cach->L3 == UNKNOWN)
return NULL;
return get_str_cache(cach->L3, topo, true);
}
char* get_str_freq(struct frequency* freq) {
//Max 3 digits and 3 for '(M/G)Hz' plus 1 for '\0'
uint32_t size = (4+3+1);
assert(strlen(STRING_UNKNOWN)+1 <= size);
char* string = malloc(sizeof(char)*size);
if(freq->max == UNKNOWN)
snprintf(string,strlen(STRING_UNKNOWN)+1,STRING_UNKNOWN);
else if(freq->max >= 1000)
snprintf(string,size,"%.2f"STRING_GIGAHERZ,(float)(freq->max)/1000);
else
snprintf(string,size,"%.2f"STRING_MEGAHERZ,(float)(freq->max));
return string;
}
void print_levels(struct cpuInfo* cpu, char* cpu_name) {
printf("%s\n", cpu_name);
printf("- Max standart level: 0x%.8X\n", cpu->maxLevels);
printf("- Max extended level: 0x%.8X\n", cpu->maxExtendedLevels);
}
void free_topo_struct(struct topology* topo) {
free(topo);
}
void free_cache_struct(struct cache* cach) {
free(cach);
}
void free_freq_struct(struct frequency* freq) {
free(freq);
}

View File

@@ -1,68 +0,0 @@
#ifndef __CPUID__
#define __CPUID__
#include <stdint.h>
#define VENDOR_EMPTY 0
#define VENDOR_INTEL 1
#define VENDOR_AMD 2
#define VENDOR_INVALID 3
#define UNKNOWN -1
struct cpuInfo;
struct frequency;
struct cache;
struct topology {
int64_t total_cores;
uint32_t physical_cores;
uint32_t logical_cores;
uint32_t smt_available; // Number of SMT that is currently enabled
uint32_t smt_supported; // Number of SMT that CPU supports (equal to smt_available if SMT is enabled)
uint32_t sockets;
struct apic* apic;
};
typedef int32_t VENDOR;
struct cpuInfo* get_cpu_info();
VENDOR get_cpu_vendor(struct cpuInfo* cpu);
uint32_t get_nsockets(struct topology* topo);
int64_t get_freq(struct frequency* freq);
struct cache* get_cache_info(struct cpuInfo* cpu);
struct frequency* get_frequency_info(struct cpuInfo* cpu);
struct topology* get_topology_info(struct cpuInfo* cpu);
char* get_str_cpu_name(struct cpuInfo* cpu);
char* get_str_ncores(struct cpuInfo* cpu);
char* get_str_avx(struct cpuInfo* cpu);
char* get_str_sse(struct cpuInfo* cpu);
char* get_str_fma(struct cpuInfo* cpu);
char* get_str_aes(struct cpuInfo* cpu);
char* get_str_sha(struct cpuInfo* cpu);
char* get_str_l1i(struct cache* cach, struct topology* topo);
char* get_str_l1d(struct cache* cach, struct topology* topo);
char* get_str_l2(struct cache* cach, struct topology* topo);
char* get_str_l3(struct cache* cach, struct topology* topo);
char* get_str_freq(struct frequency* freq);
char* get_str_sockets(struct topology* topo);
char* get_str_topology(struct cpuInfo* cpu, struct topology* topo, bool dual_socket);
char* get_str_peak_performance(struct cpuInfo* cpu, struct topology* topo, int64_t freq);
void print_levels(struct cpuInfo* cpu, char* cpu_name);
void free_cpuinfo_struct(struct cpuInfo* cpu);
void free_cache_struct(struct cache* cach);
void free_topo_struct(struct topology* topo);
void free_freq_struct(struct frequency* freq);
void debug_cpu_info(struct cpuInfo* cpu);
void debug_cache(struct cache* cach);
void debug_frequency(struct frequency* freq);
#endif

View File

@@ -1,74 +0,0 @@
#include <stdio.h>
#include <stdlib.h>
#include "args.h"
#include "printer.h"
#include "cpuid.h"
#include "global.h"
static const char* VERSION = "0.6";
void print_help(char *argv[]) {
printf("Usage: %s [--version] [--help] [--levels] [--style fancy|retro|legacy] [--color 'R,G,B:R,G,B:R,G,B:R,G,B']\n\
Options: \n\
--color Set a custom color scheme. 4 colors must be specified in RGB with the format: R,G,B:R,G,B:...\n\
These colors correspond to the ASCII art color (2 colors) and for the text colors (next 2)\n\
Suggested color (Intel): --color 15,125,194:230,230,230:40,150,220:230,230,230\n\
Suggested color (AMD): --color 250,250,250:0,154,102:250,250,250:0,154,102\n\
--style Set the style of the ASCII art:\n\
* fancy \n\
* retro \n\
* legacy \n\
--help Prints this help and exit\n\
--levels Prints CPU model and cpuid levels (debug purposes)\n\
--version Prints cpufetch version and exit\n",
argv[0]);
}
void print_version() {
printf("cpufetch v%s\n",VERSION);
}
int main(int argc, char* argv[]) {
if(!parse_args(argc,argv))
return EXIT_FAILURE;
if(show_help()) {
print_help(argv);
return EXIT_SUCCESS;
}
if(show_version()) {
print_version();
return EXIT_SUCCESS;
}
set_log_level(verbose_enabled());
struct cpuInfo* cpu = get_cpu_info();
if(cpu == NULL)
return EXIT_FAILURE;
if(show_levels()) {
print_version();
print_levels(cpu, get_str_cpu_name(cpu));
return EXIT_SUCCESS;
}
struct cache* cach = get_cache_info(cpu);
if(cach == NULL)
return EXIT_FAILURE;
struct frequency* freq = get_frequency_info(cpu);
if(freq == NULL)
return EXIT_FAILURE;
struct topology* topo = get_topology_info(cpu);
if(topo == NULL)
return EXIT_FAILURE;
if(print_cpufetch(cpu, cach, freq, topo, get_style(), get_colors()))
return EXIT_SUCCESS;
else
return EXIT_FAILURE;
}

View File

@@ -1,384 +0,0 @@
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <stdbool.h>
#include "printer.h"
#include "ascii.h"
#include "global.h"
#define COL_NONE ""
#define COL_INTEL_FANCY_1 "\x1b[46;1m"
#define COL_INTEL_FANCY_2 "\x1b[47;1m"
#define COL_INTEL_FANCY_3 "\x1b[36;1m"
#define COL_INTEL_FANCY_4 "\x1b[37;1m"
#define COL_INTEL_RETRO_1 "\x1b[36;1m"
#define COL_INTEL_RETRO_2 "\x1b[37;1m"
#define COL_AMD_FANCY_1 "\x1b[47;1m"
#define COL_AMD_FANCY_2 "\x1b[42;1m"
#define COL_AMD_FANCY_3 "\x1b[37;1m"
#define COL_AMD_FANCY_4 "\x1b[32;1m"
#define COL_AMD_RETRO_1 "\x1b[37;1m"
#define COL_AMD_RETRO_2 "\x1b[32;1m"
#define RESET "\x1b[m"
#define TITLE_NAME "Name:"
#define TITLE_FREQUENCY "Frequency:"
#define TITLE_SOCKETS "Sockets:"
#define TITLE_NCORES "Cores:"
#define TITLE_NCORES_DUAL "Cores (Total):"
#define TITLE_AVX "AVX:"
#define TITLE_SSE "SSE:"
#define TITLE_FMA "FMA:"
#define TITLE_AES "AES:"
#define TITLE_SHA "SHA:"
#define TITLE_L1i "L1i Size:"
#define TITLE_L1d "L1d Size:"
#define TITLE_L2 "L2 Size:"
#define TITLE_L3 "L3 Size:"
#define TITLE_PEAK "Peak Perf.:"
#define MAX_ATTRIBUTE_COUNT 15
#define ATTRIBUTE_NAME 0
#define ATTRIBUTE_FREQUENCY 1
#define ATTRIBUTE_SOCKETS 2
#define ATTRIBUTE_NCORES 3
#define ATTRIBUTE_NCORES_DUAL 4
#define ATTRIBUTE_AVX 5
#define ATTRIBUTE_SSE 6
#define ATTRIBUTE_FMA 7
#define ATTRIBUTE_AES 8
#define ATTRIBUTE_SHA 9
#define ATTRIBUTE_L1i 10
#define ATTRIBUTE_L1d 11
#define ATTRIBUTE_L2 12
#define ATTRIBUTE_L3 13
#define ATTRIBUTE_PEAK 14
static const char* ATTRIBUTE_FIELDS [MAX_ATTRIBUTE_COUNT] = { TITLE_NAME, TITLE_FREQUENCY, TITLE_SOCKETS,
TITLE_NCORES, TITLE_NCORES_DUAL,
TITLE_AVX, TITLE_SSE,
TITLE_FMA, TITLE_AES, TITLE_SHA,
TITLE_L1i, TITLE_L1d, TITLE_L2, TITLE_L3,
TITLE_PEAK
};
static const int ATTRIBUTE_LIST[MAX_ATTRIBUTE_COUNT] = { ATTRIBUTE_NAME, ATTRIBUTE_FREQUENCY, ATTRIBUTE_SOCKETS,
ATTRIBUTE_NCORES, ATTRIBUTE_NCORES_DUAL, ATTRIBUTE_AVX,
ATTRIBUTE_SSE, ATTRIBUTE_FMA, ATTRIBUTE_AES, ATTRIBUTE_SHA,
ATTRIBUTE_L1i, ATTRIBUTE_L1d, ATTRIBUTE_L2, ATTRIBUTE_L3,
ATTRIBUTE_PEAK };
struct ascii {
char art[NUMBER_OF_LINES][LINE_SIZE];
char color1_ascii[100];
char color2_ascii[100];
char color1_text[100];
char color2_text[100];
char ascii_chars[2];
char reset[100];
char* attributes[MAX_ATTRIBUTE_COUNT];
uint32_t n_attributes_set;
VENDOR vendor;
};
void setAttribute(struct ascii* art, int type, char* value) {
art->attributes[type] = value;
art->n_attributes_set++;
}
char* rgb_to_ansi(struct color* c, bool background, bool bold) {
char* str = malloc(sizeof(char) * 100);
if(background) {
snprintf(str, 44, "\x1b[48;2;%.3d;%.3d;%.3dm", c->R, c->G, c->B);
}
else {
if(bold)
snprintf(str, 48, "\x1b[1m\x1b[38;2;%.3d;%.3d;%.3dm", c->R, c->G, c->B);
else
snprintf(str, 44, "\x1b[38;2;%.3d;%.3d;%.3dm", c->R, c->G, c->B);
}
return str;
}
struct ascii* set_ascii(VENDOR cpuVendor, STYLE style, struct colors* cs) {
// Sanity checks //
for(int i=0; i < MAX_ATTRIBUTE_COUNT; i++) {
if(ATTRIBUTE_FIELDS[i] == NULL) {
printBug("Attribute field at position %d is empty", i);
return NULL;
}
if(i > 0 && ATTRIBUTE_LIST[i] == 0) {
printBug("Attribute list at position %d is empty", i);
return NULL;
}
}
char *COL_FANCY_1, *COL_FANCY_2, *COL_FANCY_3, *COL_FANCY_4, *COL_RETRO_1, *COL_RETRO_2, *COL_RETRO_3, *COL_RETRO_4;
struct ascii* art = malloc(sizeof(struct ascii));
art->n_attributes_set = 0;
art->vendor = cpuVendor;
for(int i=0; i < MAX_ATTRIBUTE_COUNT; i++)
art->attributes[i] = NULL;
strcpy(art->reset,RESET);
if(cpuVendor == VENDOR_INTEL) {
COL_FANCY_1 = COL_INTEL_FANCY_1;
COL_FANCY_2 = COL_INTEL_FANCY_2;
COL_FANCY_3 = COL_INTEL_FANCY_3;
COL_FANCY_4 = COL_INTEL_FANCY_4;
COL_RETRO_1 = COL_INTEL_RETRO_1;
COL_RETRO_2 = COL_INTEL_RETRO_2;
COL_RETRO_3 = COL_INTEL_RETRO_1;
COL_RETRO_4 = COL_INTEL_RETRO_2;
art->ascii_chars[0] = '#';
}
else {
COL_FANCY_1 = COL_AMD_FANCY_1;
COL_FANCY_2 = COL_AMD_FANCY_2;
COL_FANCY_3 = COL_AMD_FANCY_3;
COL_FANCY_4 = COL_AMD_FANCY_4;
COL_RETRO_1 = COL_AMD_RETRO_1;
COL_RETRO_2 = COL_AMD_RETRO_2;
COL_RETRO_3 = COL_AMD_RETRO_1;
COL_RETRO_4 = COL_AMD_RETRO_2;
art->ascii_chars[0] = '@';
}
art->ascii_chars[1] = '#';
// If style is emtpy, set the default style
if(style == STYLE_EMPTY) {
#ifdef _WIN32
style = STYLE_LEGACY;
#else
style = STYLE_FANCY;
#endif
}
switch(style) {
case STYLE_LEGACY:
strcpy(art->color1_ascii,COL_NONE);
strcpy(art->color2_ascii,COL_NONE);
strcpy(art->color1_text,COL_NONE);
strcpy(art->color2_text,COL_NONE);
art->reset[0] = '\0';
break;
case STYLE_FANCY:
if(cs != NULL) {
COL_FANCY_1 = rgb_to_ansi(cs->c1, true, true);
COL_FANCY_2 = rgb_to_ansi(cs->c2, true, true);
COL_FANCY_3 = rgb_to_ansi(cs->c3, false, true);
COL_FANCY_4 = rgb_to_ansi(cs->c4, false, true);
}
art->ascii_chars[0] = ' ';
art->ascii_chars[1] = ' ';
strcpy(art->color1_ascii,COL_FANCY_1);
strcpy(art->color2_ascii,COL_FANCY_2);
strcpy(art->color1_text,COL_FANCY_3);
strcpy(art->color2_text,COL_FANCY_4);
if(cs != NULL) {
free(COL_FANCY_1);
free(COL_FANCY_2);
free(COL_FANCY_3);
free(COL_FANCY_4);
}
break;
case STYLE_RETRO:
if(cs != NULL) {
COL_RETRO_1 = rgb_to_ansi(cs->c1, false, true);
COL_RETRO_2 = rgb_to_ansi(cs->c2, false, true);
COL_RETRO_3 = rgb_to_ansi(cs->c3, false, true);
COL_RETRO_4 = rgb_to_ansi(cs->c4, false, true);
}
strcpy(art->color1_ascii,COL_RETRO_1);
strcpy(art->color2_ascii,COL_RETRO_2);
strcpy(art->color1_text,COL_RETRO_3);
strcpy(art->color2_text,COL_RETRO_4);
if(cs != NULL) {
free(COL_RETRO_1);
free(COL_RETRO_2);
free(COL_RETRO_3);
free(COL_RETRO_4);
}
break;
case STYLE_INVALID:
default:
printBug("Found invalid style (%d)",style);
return NULL;
}
char tmp[NUMBER_OF_LINES*LINE_SIZE];
if(cpuVendor == VENDOR_INTEL) strcpy(tmp, INTEL_ASCII);
else strcpy(tmp, AMD_ASCII);
for(int i=0; i < NUMBER_OF_LINES; i++)
strncpy(art->art[i], tmp + i*LINE_SIZE, LINE_SIZE);
return art;
}
uint32_t get_next_attribute(struct ascii* art, uint32_t last_attr) {
last_attr++;
while(art->attributes[last_attr] == NULL) last_attr++;
return last_attr;
}
void print_ascii_intel(struct ascii* art, uint32_t la) {
bool flag = false;
int attr_to_print = -1;
uint32_t space_right;
uint32_t space_up = (NUMBER_OF_LINES - art->n_attributes_set)/2;
uint32_t space_down = NUMBER_OF_LINES - art->n_attributes_set - space_up;
for(uint32_t n=0;n<NUMBER_OF_LINES;n++) {
for(int i=0;i<LINE_SIZE;i++) {
if(flag) {
if(art->art[n][i] == ' ') {
flag = false;
printf("%s%c%s", art->color2_ascii, art->ascii_chars[1], art->reset);
}
else
printf("%s%c%s", art->color1_ascii, art->ascii_chars[0], art->reset);
}
else {
if(art->art[n][i] != ' ' && art->art[n][i] != '\0') {
flag = true;
printf("%c",' ');
}
else
printf("%c",' ');
}
}
if(n > space_up-1 && n < NUMBER_OF_LINES-space_down) {
attr_to_print = get_next_attribute(art, attr_to_print);
space_right = 1 + (la - strlen(ATTRIBUTE_FIELDS[attr_to_print]));
printf("%s%s%s%*s%s%s%s\n",art->color1_text, ATTRIBUTE_FIELDS[attr_to_print], art->reset, space_right, "", art->color2_text, art->attributes[attr_to_print], art->reset);
}
else printf("\n");
}
}
void print_ascii_amd(struct ascii* art, uint32_t la) {
int attr_to_print = -1;
uint32_t space_right;
uint32_t space_up = (NUMBER_OF_LINES - art->n_attributes_set)/2;
uint32_t space_down = NUMBER_OF_LINES - art->n_attributes_set - space_up;
for(uint32_t n=0;n<NUMBER_OF_LINES;n++) {
for(int i=0;i<LINE_SIZE;i++) {
if(art->art[n][i] == '@')
printf("%s%c%s", art->color1_ascii, art->ascii_chars[0], art->reset);
else if(art->art[n][i] == '#')
printf("%s%c%s", art->color2_ascii, art->ascii_chars[1], art->reset);
else
printf("%c",art->art[n][i]);
}
if(n > space_up-1 && n < NUMBER_OF_LINES-space_down) {
attr_to_print = get_next_attribute(art, attr_to_print);
space_right = 1 + (la - strlen(ATTRIBUTE_FIELDS[attr_to_print]));
printf("%s%s%s%*s%s%s%s\n",art->color1_text, ATTRIBUTE_FIELDS[attr_to_print], art->reset, space_right, "", art->color2_text, art->attributes[attr_to_print], art->reset);
}
else printf("\n");
}
}
uint32_t longest_attribute_length(struct ascii* art) {
uint32_t max = 0;
uint64_t len = 0;
for(int i=0; i < MAX_ATTRIBUTE_COUNT; i++) {
if(art->attributes[i] != NULL) {
len = strlen(ATTRIBUTE_FIELDS[i]);
if(len > max) max = len;
}
}
return max;
}
void print_ascii(struct ascii* art) {
uint32_t longest_attribute = longest_attribute_length(art);
if(art->vendor == VENDOR_INTEL)
print_ascii_intel(art, longest_attribute);
else
print_ascii_amd(art, longest_attribute);
}
bool print_cpufetch(struct cpuInfo* cpu, struct cache* cach, struct frequency* freq, struct topology* topo, STYLE s, struct colors* cs) {
struct ascii* art = set_ascii(get_cpu_vendor(cpu), s, cs);
if(art == NULL)
return false;
char* cpu_name = get_str_cpu_name(cpu);
char* sockets = get_str_sockets(topo);
char* max_frequency = get_str_freq(freq);
char* n_cores = get_str_topology(cpu, topo, false);
char* n_cores_dual = get_str_topology(cpu, topo, true);
char* avx = get_str_avx(cpu);
char* sse = get_str_sse(cpu);
char* fma = get_str_fma(cpu);
char* aes = get_str_aes(cpu);
char* sha = get_str_sha(cpu);
char* l1i = get_str_l1i(cach, topo);
char* l1d = get_str_l1d(cach, topo);
char* l2 = get_str_l2(cach, topo);
char* l3 = get_str_l3(cach, topo);
char* pp = get_str_peak_performance(cpu,topo,get_freq(freq));
setAttribute(art,ATTRIBUTE_NAME,cpu_name);
setAttribute(art,ATTRIBUTE_FREQUENCY,max_frequency);
setAttribute(art,ATTRIBUTE_NCORES,n_cores);
setAttribute(art,ATTRIBUTE_AVX,avx);
setAttribute(art,ATTRIBUTE_SSE,sse);
setAttribute(art,ATTRIBUTE_FMA,fma);
setAttribute(art,ATTRIBUTE_AES,aes);
setAttribute(art,ATTRIBUTE_SHA,sha);
setAttribute(art,ATTRIBUTE_L1i,l1i);
setAttribute(art,ATTRIBUTE_L1d,l1d);
setAttribute(art,ATTRIBUTE_L2,l2);
setAttribute(art,ATTRIBUTE_PEAK,pp);
uint32_t socket_num = get_nsockets(topo);
if (socket_num > 1) {
setAttribute(art, ATTRIBUTE_SOCKETS, sockets);
setAttribute(art, ATTRIBUTE_NCORES_DUAL, n_cores_dual);
}
if(l3 != NULL) {
setAttribute(art,ATTRIBUTE_L3,l3);
}
if(art->n_attributes_set > NUMBER_OF_LINES) {
printBug("The number of attributes set is bigger than the max that can be displayed");
return false;
}
print_ascii(art);
free(cpu_name);
free(max_frequency);
free(sockets);
free(n_cores);
free(n_cores_dual);
free(avx);
free(sse);
free(fma);
free(aes);
free(sha);
free(l1i);
free(l1d);
free(l2);
free(l3);
free(pp);
free(cpu);
free(art);
if(cs != NULL) free_colors_struct(cs);
free_cache_struct(cach);
free_topo_struct(topo);
free_freq_struct(freq);
return true;
}

View File

@@ -1,19 +0,0 @@
#ifndef __PRINTER__
#define __PRINTER__
typedef int STYLE;
#include "args.h"
#include "cpuid.h"
#define STYLES_COUNT 3
#define STYLE_INVALID -2
#define STYLE_EMPTY -1
#define STYLE_FANCY 0
#define STYLE_RETRO 1
#define STYLE_LEGACY 2
bool print_cpufetch(struct cpuInfo* cpu, struct cache* cach, struct frequency* freq, struct topology* topo, STYLE s, struct colors* cs);
#endif

View File

@@ -1,74 +0,0 @@
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <fcntl.h>
#include <errno.h>
#include "global.h"
#include "cpuid.h"
#define _PATH_SYS_SYSTEM "/sys/devices/system"
#define _PATH_SYS_CPU _PATH_SYS_SYSTEM"/cpu"
#define _PATH_ONE_CPU _PATH_SYS_CPU"/cpu0"
#define _PATH_FREQUENCY _PATH_ONE_CPU"/cpufreq"
#define _PATH_FREQUENCY_MAX _PATH_FREQUENCY"/cpuinfo_max_freq"
#define _PATH_FREQUENCY_MIN _PATH_FREQUENCY"/cpuinfo_min_freq"
#define DEFAULT_FILE_SIZE 4096
long get_freq_from_file(char* path) {
int fd = open(path, O_RDONLY);
if(fd == -1) {
perror("open");
printBug("Could not open '%s'", path);
return UNKNOWN;
}
//File exists, read it
int bytes_read = 0;
int offset = 0;
int block = 1;
char* buf = malloc(sizeof(char)*DEFAULT_FILE_SIZE);
memset(buf, 0, sizeof(char)*DEFAULT_FILE_SIZE);
while ( (bytes_read = read(fd, buf+offset, block)) > 0 ) {
offset += bytes_read;
}
char* end;
errno = 0;
long ret = strtol(buf, &end, 10);
if(errno != 0) {
perror("strtol");
printBug("Failed parsing '%s' file. Read data was: '%s'", path, buf);
free(buf);
return UNKNOWN;
}
// We will be getting the frequency in KHz
// We consider it is an error if frequency is
// greater than 10 GHz or less than 100 MHz
if(ret > 10000 * 1000 || ret < 100 * 1000) {
printBug("Invalid data was read from file '%s': %ld\n", path, ret);
return UNKNOWN;
}
free(buf);
if (close(fd) == -1) {
perror("close");
printErr("Closing '%s' failed\n", path);
}
return ret/1000;
}
long get_max_freq_from_file() {
return get_freq_from_file(_PATH_FREQUENCY_MAX);
}
long get_min_freq_from_file() {
return get_freq_from_file(_PATH_FREQUENCY_MIN);
}

View File

@@ -1,7 +0,0 @@
#ifndef __UDEV__
#define __UDEV__
long get_max_freq_from_file();
long get_min_freq_from_file();
#endif

429
src/x86/apic.c Normal file
View File

@@ -0,0 +1,429 @@
#ifdef _WIN32
#define NOMINMAX
#include <windows.h>
#elif defined __linux__
#define _GNU_SOURCE
#include <sched.h>
#elif defined __APPLE__
#define UNUSED(x) (void)(x)
#endif
#include <stdlib.h>
#include <stdint.h>
#include <string.h>
#include <stdio.h>
#include <unistd.h>
#include "apic.h"
#include "cpuid_asm.h"
#include "../common/global.h"
/*
* bit_scan_reverse and create_mask code taken from:
* https://software.intel.com/content/www/us/en/develop/articles/intel-64-architecture-processor-topology-enumeration.html
*/
unsigned char bit_scan_reverse(uint32_t* index, uint64_t mask) {
for(uint64_t i = (8 * sizeof(uint64_t)); i > 0; i--) {
if((mask & (1LL << (i-1))) != 0) {
*index = (uint64_t) (i-1);
break;
}
}
return (unsigned char) (mask != 0);
}
uint32_t create_mask(uint32_t num_entries, uint32_t *mask_width) {
uint32_t i = 0;
uint64_t k = 0;
// NearestPo2(numEntries) is the nearest power of 2 integer that is not less than numEntries
// The most significant bit of (numEntries * 2 -1) matches the above definition
k = (uint64_t)(num_entries) * 2 -1;
if (bit_scan_reverse(&i, k) == 0) {
if (mask_width) *mask_width = 0;
return 0;
}
if (mask_width) *mask_width = i;
if (i == 31) return (uint32_t ) -1;
return (1 << i) -1;
}
uint32_t get_apic_id(bool x2apic_id) {
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
if(x2apic_id) {
eax = 0x0000000B;
cpuid(&eax, &ebx, &ecx, &edx);
return edx;
}
else {
eax = 0x00000001;
cpuid(&eax, &ebx, &ecx, &edx);
return (ebx >> 24);
}
}
#ifndef __APPLE__
bool bind_to_cpu(int cpu_id) {
#ifdef _WIN32
HANDLE process = GetCurrentProcess();
DWORD_PTR processAffinityMask = 1 << cpu_id;
return SetProcessAffinityMask(process, processAffinityMask);
#else
cpu_set_t currentCPU;
CPU_ZERO(&currentCPU);
CPU_SET(cpu_id, &currentCPU);
if (sched_setaffinity (0, sizeof(currentCPU), &currentCPU) == -1) {
perror("sched_setaffinity");
return false;
}
return true;
#endif
}
#endif
bool fill_topo_masks_apic(struct topology* topo) {
uint32_t eax = 0x00000001;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
uint32_t core_plus_smt_id_max_cnt;
uint32_t core_id_max_cnt;
uint32_t smt_id_per_core_max_cnt;
cpuid(&eax, &ebx, &ecx, &edx);
core_plus_smt_id_max_cnt = (ebx >> 16) & 0xFF;
eax = 0x00000004;
ecx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
core_id_max_cnt = (eax >> 26) + 1;
smt_id_per_core_max_cnt = core_plus_smt_id_max_cnt / core_id_max_cnt;
topo->apic->smt_mask = create_mask(smt_id_per_core_max_cnt, &(topo->apic->smt_mask_width));
topo->apic->core_mask = create_mask(core_id_max_cnt,&(topo->apic->pkg_mask_shift));
topo->apic->pkg_mask_shift += topo->apic->smt_mask_width;
topo->apic->core_mask <<= topo->apic->smt_mask_width;
topo->apic->pkg_mask = (-1) ^ (topo->apic->core_mask | topo->apic->smt_mask);
return true;
}
bool fill_topo_masks_x2apic(struct topology* topo) {
int32_t level_type;
int32_t level_shift;
int32_t coreplus_smt_mask = 0;
bool level2 = false;
bool level1 = false;
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
uint32_t i = 0;
while(true) {
eax = 0x0000000B;
ecx = i;
cpuid(&eax, &ebx, &ecx, &edx);
if(ebx == 0) break;
level_type = (ecx >> 8) & 0xFF;
level_shift = eax & 0xFFF;
switch(level_type) {
case 1: // SMT
topo->apic->smt_mask = ~(0xFFFFFFFF << level_shift);
topo->apic->smt_mask_width = level_shift;
topo->smt_supported = ebx & 0xFFFF;
level1 = true;
break;
case 2: // Core
coreplus_smt_mask = ~(0xFFFFFFFF << level_shift);
topo->apic->pkg_mask_shift = level_shift;
topo->apic->pkg_mask = (-1) ^ coreplus_smt_mask;
level2 = true;
break;
default:
printErr("Found invalid level when querying topology: %d", level_type);
break;
}
i++; // sublevel to query
}
if (level1 && level2) {
topo->apic->core_mask = coreplus_smt_mask ^ topo->apic->smt_mask;
}
else if (!level2 && level1) {
topo->apic->core_mask = 0;
topo->apic->pkg_mask_shift = topo->apic->smt_mask_width;
topo->apic->pkg_mask = (-1) ^ topo->apic->smt_mask;
}
else {
printErr("SMT level was not found when querying topology");
return false;
}
return true;
}
// Not a very elegant solution. The width should always be as long
// as the number of cores, but in the case of Xeon Phi KNL it is not
uint32_t max_apic_id_size(uint32_t** cache_id_apic, struct topology* topo) {
uint32_t max = 0;
for(int i=0; i < topo->cach->max_cache_level; i++) {
for(int j=0; j < topo->total_cores; j++) {
if(cache_id_apic[j][i] > max) max = cache_id_apic[j][i];
}
}
max++;
if(max > (uint32_t) topo->total_cores) return max;
return topo->total_cores;
}
bool build_topo_from_apic(uint32_t* apic_pkg, uint32_t* apic_smt, uint32_t** cache_id_apic, struct topology* topo) {
uint32_t size = max_apic_id_size(cache_id_apic, topo);
uint32_t* sockets = malloc(sizeof(uint32_t) * size);
uint32_t* smt = malloc(sizeof(uint32_t) * size);
uint32_t* apic_id = malloc(sizeof(uint32_t) * size);
uint32_t num_caches = 0;
memset(sockets, 0, sizeof(uint32_t) * size);
memset(smt, 0, sizeof(uint32_t) * size);
memset(apic_id, 0, sizeof(uint32_t) * size);
// System topology
for(int i=0; i < topo->total_cores; i++) {
sockets[apic_pkg[i]] = 1;
smt[apic_smt[i]] = 1;
}
for(int i=0; i < topo->total_cores; i++) {
if(sockets[i] != 0)
topo->sockets++;
if(smt[i] != 0)
topo->smt_available++;
}
topo->logical_cores = topo->total_cores / topo->sockets;
topo->physical_cores = topo->logical_cores / topo->smt_available;
// Cache topology
for(int i=0; i < topo->cach->max_cache_level; i++) {
num_caches = 0;
memset(apic_id, 0, sizeof(uint32_t) * size);
for(int c=0; c < topo->total_cores; c++) {
apic_id[cache_id_apic[c][i]]++;
}
for(uint32_t c=0; c < size; c++) {
if(apic_id[c] > 0) num_caches++;
}
topo->cach->cach_arr[i]->num_caches = num_caches;
}
free(sockets);
free(smt);
free(apic_id);
return true;
}
void get_cache_topology_from_apic(struct topology* topo) {
uint32_t eax = 0x00000004;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
for(int i=0; i < topo->cach->max_cache_level; i++) {
eax = 0x00000004;
ecx = i;
cpuid(&eax, &ebx, &ecx, &edx);
uint32_t SMTMaxCntPerEachCache = ((eax >> 14) & 0x7FF) + 1;
uint32_t dummy;
topo->apic->cache_select_mask[i] = create_mask(SMTMaxCntPerEachCache,&dummy);
}
}
bool apic_array_full(uint32_t* apic_ids, int n) {
for(int i=0; i < n; i++) {
if(apic_ids[i] == (uint32_t) -1) return false;
}
return true;
}
void add_apic_to_array(uint32_t apic, uint32_t* apic_ids, int n) {
int i=0;
int last=0;
bool found = false;
while(!found && i < n) {
if(apic_ids[i] == apic) found = true;
if(apic_ids[i] != (uint32_t) -1) last = i+1;
i++;
}
if(!found) {
apic_ids[last] = apic;
//printf("Added %d\n", apic);
}
}
bool fill_apic_ids(uint32_t* apic_ids, int n, bool x2apic_id) {
#ifdef __APPLE__
// macOS extremely dirty approach...
printf("cpufetch is computing APIC IDs, please wait...\n");
bool end = false;
uint32_t apic;
for(int i=0; i < n; i++) apic_ids[i] = (uint32_t) -1;
while(!end) {
apic = get_apic_id(x2apic_id);
add_apic_to_array(apic, apic_ids, n);
end = apic_array_full(apic_ids, n);
usleep(1000);
}
#else
for(int i=0; i < n; i++) {
if(!bind_to_cpu(i)) {
printErr("Failed binding to CPU %d", i);
return false;
}
apic_ids[i] = get_apic_id(x2apic_id);
}
#endif
return true;
}
bool get_topology_from_apic(struct cpuInfo* cpu, struct topology* topo) {
uint32_t apic_id;
uint32_t* apic_ids = malloc(sizeof(uint32_t) * topo->total_cores);
uint32_t* apic_pkg = malloc(sizeof(uint32_t) * topo->total_cores);
uint32_t* apic_core = malloc(sizeof(uint32_t) * topo->total_cores);
uint32_t* apic_smt = malloc(sizeof(uint32_t) * topo->total_cores);
uint32_t** cache_smt_id_apic = malloc(sizeof(uint32_t*) * topo->total_cores);
uint32_t** cache_id_apic = malloc(sizeof(uint32_t*) * topo->total_cores);
bool x2apic_id;
if(cpu->maxLevels >= 0x0000000B) {
uint32_t eax = 0x0000000B;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
if(ebx == 0) x2apic_id = false;
else x2apic_id = true;
}
else {
x2apic_id = false;
}
for(int i=0; i < topo->total_cores; i++) {
cache_smt_id_apic[i] = malloc(sizeof(uint32_t) * (topo->cach->max_cache_level));
cache_id_apic[i] = malloc(sizeof(uint32_t) * (topo->cach->max_cache_level));
}
topo->apic->cache_select_mask = malloc(sizeof(uint32_t) * (topo->cach->max_cache_level));
topo->apic->cache_id_apic = malloc(sizeof(uint32_t) * (topo->cach->max_cache_level));
if(x2apic_id) {
if(!fill_topo_masks_x2apic(topo))
return false;
}
else {
if(!fill_topo_masks_apic(topo))
return false;
}
get_cache_topology_from_apic(topo);
if(!fill_apic_ids(apic_ids, topo->total_cores, x2apic_id))
return false;
for(int i=0; i < topo->total_cores; i++) {
apic_id = apic_ids[i];
apic_pkg[i] = (apic_id & topo->apic->pkg_mask) >> topo->apic->pkg_mask_shift;
apic_core[i] = (apic_id & topo->apic->core_mask) >> topo->apic->smt_mask_width;
apic_smt[i] = apic_id & topo->apic->smt_mask;
for(int c=0; c < topo->cach->max_cache_level; c++) {
cache_smt_id_apic[i][c] = apic_id & topo->apic->cache_select_mask[c];
cache_id_apic[i][c] = apic_id & (-1 ^ topo->apic->cache_select_mask[c]);
}
}
/* DEBUG
for(int i=0; i < topo->cach->max_cache_level; i++) {
printf("[CACH %1d]", i);
for(int j=0; j < topo->total_cores; j++)
printf("[%03d]", cache_id_apic[j][i]);
printf("\n");
}
for(int i=0; i < topo->total_cores; i++)
printf("[%2d] 0x%.8X\n", i, apic_pkg[i]);
printf("\n");
for(int i=0; i < topo->total_cores; i++)
printf("[%2d] 0x%.8X\n", i, apic_core[i]);
printf("\n");
for(int i=0; i < topo->total_cores; i++)
printf("[%2d] 0x%.8X\n", i, apic_smt[i]);*/
bool ret = build_topo_from_apic(apic_pkg, apic_smt, cache_id_apic, topo);
// Assumption: If we cant get smt_available, we assume it is equal to smt_supported...
if (!x2apic_id) {
printWarn("Can't read SMT from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x0000000B, cpu->maxLevels);
topo->smt_supported = topo->smt_available;
}
free(apic_pkg);
free(apic_core);
free(apic_smt);
for(int i=0; i < topo->total_cores; i++) {
free(cache_smt_id_apic[i]);
free(cache_id_apic[i]);
}
free(cache_smt_id_apic);
free(cache_id_apic);
return ret;
}
uint32_t is_smt_enabled_amd(struct topology* topo) {
#ifdef __APPLE__
UNUSED(topo);
return 1;
#else
uint32_t id;
for(int i = 0; i < topo->total_cores; i++) {
if(!bind_to_cpu(i)) {
printErr("Failed binding to CPU %d", i);
return false;
}
id = get_apic_id(false) & 1; // get the last bit
if(id == 1) return 2; // We assume there isn't any AMD CPU with more than 2th per core.
}
return 1;
#endif
}

View File

@@ -10,9 +10,11 @@ struct apic {
uint32_t core_mask;
uint32_t smt_mask_width;
uint32_t smt_mask;
uint32_t* cache_select_mask;
uint32_t* cache_id_apic;
};
bool get_topology_from_apic(uint32_t cpuid_max_levels, struct topology** topo);
uint32_t is_smt_enabled(struct topology* topo);
bool get_topology_from_apic(struct cpuInfo* cpu, struct topology* topo);
uint32_t is_smt_enabled_amd(struct topology* topo);
#endif

906
src/x86/cpuid.c Executable file
View File

@@ -0,0 +1,906 @@
#ifdef _WIN32
#define NOMINMAX
#include <windows.h>
#else
#include "../common/udev.h"
#include <unistd.h>
#endif
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>
#include <stdbool.h>
#include "cpuid.h"
#include "cpuid_asm.h"
#include "../common/global.h"
#include "apic.h"
#include "uarch.h"
#define CPU_VENDOR_INTEL_STRING "GenuineIntel"
#define CPU_VENDOR_AMD_STRING "AuthenticAMD"
static const char *hv_vendors_string[] = {
[HV_VENDOR_KVM] = "KVMKVMKVM",
[HV_VENDOR_QEMU] = "TCGTCGTCGTCG",
[HV_VENDOR_HYPERV] = "Microsoft Hv",
[HV_VENDOR_VMWARE] = "VMwareVMware",
[HV_VENDOR_XEN] = "XenVMMXenVMM",
[HV_VENDOR_PARALLELS] = "lrpepyh vr",
};
static char *hv_vendors_name[] = {
[HV_VENDOR_KVM] = "KVM",
[HV_VENDOR_QEMU] = "QEMU",
[HV_VENDOR_HYPERV] = "Microsoft Hyper-V",
[HV_VENDOR_VMWARE] = "VMware",
[HV_VENDOR_XEN] = "Xen",
[HV_VENDOR_PARALLELS] = "Parallels",
[HV_VENDOR_INVALID] = "Unknown"
};
#define STRING_UNKNOWN "Unknown"
#define HYPERVISOR_NAME_MAX_LENGTH 17
#define MASK 0xFF
/*
* cpuid reference: http://www.sandpile.org/x86/cpuid.htm
* cpuid amd: https://www.amd.com/system/files/TechDocs/25481.pdf
*/
void init_topology_struct(struct topology* topo, struct cache* cach) {
topo->total_cores = 0;
topo->physical_cores = 0;
topo->logical_cores = 0;
topo->smt_available = 0;
topo->smt_supported = 0;
topo->sockets = 0;
topo->apic = malloc(sizeof(struct apic));
topo->cach = cach;
}
void init_cache_struct(struct cache* cach) {
cach->L1i = malloc(sizeof(struct cach));
cach->L1d = malloc(sizeof(struct cach));
cach->L2 = malloc(sizeof(struct cach));
cach->L3 = malloc(sizeof(struct cach));
cach->cach_arr = malloc(sizeof(struct cach*) * 4);
cach->cach_arr[0] = cach->L1i;
cach->cach_arr[1] = cach->L1d;
cach->cach_arr[2] = cach->L2;
cach->cach_arr[3] = cach->L3;
cach->max_cache_level = 0;
cach->L1i->exists = false;
cach->L1d->exists = false;
cach->L2->exists = false;
cach->L3->exists = false;
}
void get_name_cpuid(char* name, uint32_t reg1, uint32_t reg2, uint32_t reg3) {
uint32_t c = 0;
name[c++] = reg1 & MASK;
name[c++] = (reg1>>8) & MASK;
name[c++] = (reg1>>16) & MASK;
name[c++] = (reg1>>24) & MASK;
name[c++] = reg2 & MASK;
name[c++] = (reg2>>8) & MASK;
name[c++] = (reg2>>16) & MASK;
name[c++] = (reg2>>24) & MASK;
name[c++] = reg3 & MASK;
name[c++] = (reg3>>8) & MASK;
name[c++] = (reg3>>16) & MASK;
name[c++] = (reg3>>24) & MASK;
}
char* get_str_cpu_name_internal() {
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
uint32_t c = 0;
char * name = malloc(sizeof(char) * CPU_NAME_MAX_LENGTH);
memset(name, 0, CPU_NAME_MAX_LENGTH);
for(int i=0; i < 3; i++) {
eax = 0x80000002 + i;
cpuid(&eax, &ebx, &ecx, &edx);
name[c++] = eax & MASK;
name[c++] = (eax>>8) & MASK;
name[c++] = (eax>>16) & MASK;
name[c++] = (eax>>24) & MASK;
name[c++] = ebx & MASK;
name[c++] = (ebx>>8) & MASK;
name[c++] = (ebx>>16) & MASK;
name[c++] = (ebx>>24) & MASK;
name[c++] = ecx & MASK;
name[c++] = (ecx>>8) & MASK;
name[c++] = (ecx>>16) & MASK;
name[c++] = (ecx>>24) & MASK;
name[c++] = edx & MASK;
name[c++] = (edx>>8) & MASK;
name[c++] = (edx>>16) & MASK;
name[c++] = (edx>>24) & MASK;
}
name[c] = '\0';
//Remove unused characters
char *str = name;
char *dest = name;
// Remove spaces before name
while (*str != '\0' && *str == ' ')str++;
// Remove spaces between the name and after it
while (*str != '\0') {
while (*str == ' ' && *(str + 1) == ' ') str++;
*dest++ = *str++;
}
*dest = '\0';
return name;
}
struct uarch* get_cpu_uarch(struct cpuInfo* cpu) {
uint32_t eax = 0x00000001;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
uint32_t stepping = eax & 0xF;
uint32_t model = (eax >> 4) & 0xF;
uint32_t emodel = (eax >> 16) & 0xF;
uint32_t family = (eax >> 8) & 0xF;
uint32_t efamily = (eax >> 20) & 0xFF;
return get_uarch_from_cpuid(cpu, efamily, family, emodel, model, (int)stepping);
}
struct hypervisor* get_hp_info(bool hv_present) {
struct hypervisor* hv = malloc(sizeof(struct hypervisor));
if(!hv_present) {
hv->present = false;
return hv;
}
hv->present = true;
uint32_t eax = 0x40000000;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
char name[13];
memset(name, 0, 13);
get_name_cpuid(name, ebx, ecx, edx);
bool found = false;
uint8_t len = sizeof(hv_vendors_string) / sizeof(hv_vendors_string[0]);
for(uint8_t v=0; v < len && !found; v++) {
if(strcmp(hv_vendors_string[v], name) == 0) {
hv->hv_vendor = v;
found = true;
}
}
if(!found) {
hv->hv_vendor = HV_VENDOR_INVALID;
printWarn("Unknown hypervisor vendor: %s", name);
}
hv->hv_name = hv_vendors_name[hv->hv_vendor];
return hv;
}
struct cpuInfo* get_cpu_info() {
struct cpuInfo* cpu = malloc(sizeof(struct cpuInfo));
struct features* feat = malloc(sizeof(struct features));
cpu->feat = feat;
bool *ptr = &(feat->AES);
for(uint32_t i = 0; i < sizeof(struct features)/sizeof(bool); i++, ptr++) {
*ptr = false;
}
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
//Get max cpuid level
cpuid(&eax, &ebx, &ecx, &edx);
cpu->maxLevels = eax;
//Fill vendor
char name[13];
memset(name,0,13);
get_name_cpuid(name, ebx, edx, ecx);
if(strcmp(CPU_VENDOR_INTEL_STRING,name) == 0)
cpu->cpu_vendor = CPU_VENDOR_INTEL;
else if (strcmp(CPU_VENDOR_AMD_STRING,name) == 0)
cpu->cpu_vendor = CPU_VENDOR_AMD;
else {
cpu->cpu_vendor = CPU_VENDOR_INVALID;
printErr("Unknown CPU vendor: %s", name);
return NULL;
}
//Get max extended level
eax = 0x80000000;
ebx = 0;
ecx = 0;
edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
cpu->maxExtendedLevels = eax;
//Fill instructions support
if (cpu->maxLevels >= 0x00000001){
eax = 0x00000001;
cpuid(&eax, &ebx, &ecx, &edx);
feat->SSE = (edx & ((int)1 << 25)) != 0;
feat->SSE2 = (edx & ((int)1 << 26)) != 0;
feat->SSE3 = (ecx & ((int)1 << 0)) != 0;
feat->SSSE3 = (ecx & ((int)1 << 9)) != 0;
feat->SSE4_1 = (ecx & ((int)1 << 19)) != 0;
feat->SSE4_2 = (ecx & ((int)1 << 20)) != 0;
feat->AES = (ecx & ((int)1 << 25)) != 0;
feat->AVX = (ecx & ((int)1 << 28)) != 0;
feat->FMA3 = (ecx & ((int)1 << 12)) != 0;
bool hv_present = (ecx & ((int)1 << 31)) != 0;
if((cpu->hv = get_hp_info(hv_present)) == NULL)
return NULL;
}
else {
printWarn("Can't read features information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x00000001, cpu->maxLevels);
}
if (cpu->maxLevels >= 0x00000007){
eax = 0x00000007;
ecx = 0x00000000;
cpuid(&eax, &ebx, &ecx, &edx);
feat->AVX2 = (ebx & ((int)1 << 5)) != 0;
feat->SHA = (ebx & ((int)1 << 29)) != 0;
feat->AVX512 = (((ebx & ((int)1 << 16)) != 0) ||
((ebx & ((int)1 << 28)) != 0) ||
((ebx & ((int)1 << 26)) != 0) ||
((ebx & ((int)1 << 27)) != 0) ||
((ebx & ((int)1 << 31)) != 0) ||
((ebx & ((int)1 << 30)) != 0) ||
((ebx & ((int)1 << 17)) != 0) ||
((ebx & ((int)1 << 21)) != 0));
}
else {
printWarn("Can't read features information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x00000007, cpu->maxLevels);
}
if (cpu->maxExtendedLevels >= 0x80000001){
eax = 0x80000001;
cpuid(&eax, &ebx, &ecx, &edx);
feat->SSE4a = (ecx & ((int)1 << 6)) != 0;
feat->FMA4 = (ecx & ((int)1 << 16)) != 0;
}
else {
printWarn("Can't read features information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x80000001, cpu->maxExtendedLevels);
}
if (cpu->maxExtendedLevels >= 0x80000004){
cpu->cpu_name = get_str_cpu_name_internal();
}
else {
cpu->cpu_name = malloc(sizeof(char)*8);
sprintf(cpu->cpu_name,"Unknown");
printWarn("Can't read cpu name from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x80000004, cpu->maxExtendedLevels);
}
cpu->arch = get_cpu_uarch(cpu);
cpu->freq = get_frequency_info(cpu);
cpu->cach = get_cache_info(cpu);
cpu->topo = get_topology_info(cpu, cpu->cach);
if(cpu->cach == NULL || cpu->topo == NULL) {
return NULL;
}
return cpu;
}
bool get_cache_topology_amd(struct cpuInfo* cpu, struct topology* topo) {
if(cpu->maxExtendedLevels >= 0x8000001D) {
uint32_t i, eax, ebx, ecx, edx, num_sharing_cache, cache_type, cache_level;
i = 0;
do {
eax = 0x8000001D;
ebx = 0;
ecx = i; // cache id
edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
cache_type = eax & 0x1F;
if(cache_type > 0) {
num_sharing_cache = ((eax >> 14) & 0xFFF) + 1;
cache_level = (eax >>= 5) & 0x7;
switch (cache_type) {
case 1: // Data Cache (We assume this is L1d)
if(cache_level != 1) {
printBug("Found data cache at level %d (expected 1)", cache_level);
return false;
}
topo->cach->L1d->num_caches = topo->logical_cores / num_sharing_cache;
break;
case 2: // Instruction Cache (We assume this is L1i)
if(cache_level != 1) {
printBug("Found instruction cache at level %d (expected 1)", cache_level);
return false;
}
topo->cach->L1i->num_caches = topo->logical_cores / num_sharing_cache;
break;
case 3: // Unified Cache (This may be L2 or L3)
if(cache_level == 2) {
topo->cach->L2->num_caches = topo->logical_cores / num_sharing_cache;
}
else if(cache_level == 3) {
topo->cach->L3->num_caches = topo->logical_cores / num_sharing_cache;
}
else {
printWarn("Found unknown unified cache at level %d", cache_level);
}
break;
default: // Unknown cache type
printBug("Unknown cache type %d with level %d found at i=%d", cache_type, cache_level, i);
return false;
}
}
i++;
} while (cache_type > 0);
}
else {
printWarn("Can't read topology information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X). Guessing cache sizes", 0x8000001D, cpu->maxExtendedLevels);
topo->cach->L1i->num_caches = topo->physical_cores;
topo->cach->L1d->num_caches = topo->physical_cores;
if(topo->cach->L3->exists) {
topo->cach->L2->num_caches = topo->physical_cores;
topo->cach->L3->num_caches = 1;
}
else {
topo->cach->L2->num_caches = 1;
}
}
return true;
}
// Main reference: https://software.intel.com/content/www/us/en/develop/articles/intel-64-architecture-processor-topology-enumeration.html
// Very interesting resource: https://wiki.osdev.org/Detecting_CPU_Topology_(80x86)
struct topology* get_topology_info(struct cpuInfo* cpu, struct cache* cach) {
struct topology* topo = malloc(sizeof(struct topology));
init_topology_struct(topo, cach);
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
// Ask the OS the total number of cores it sees
// If we have one socket, it will be same as the cpuid,
// but in dual socket it will not!
// TODO: Replace by apic?
#ifdef _WIN32
SYSTEM_INFO info;
GetSystemInfo(&info);
topo->total_cores = info.dwNumberOfProcessors;
#else
if((topo->total_cores = sysconf(_SC_NPROCESSORS_ONLN)) == -1) {
perror("sysconf");
topo->total_cores = topo->logical_cores; // fallback
}
#endif
switch(cpu->cpu_vendor) {
case CPU_VENDOR_INTEL:
if (cpu->maxLevels >= 0x00000004) {
get_topology_from_apic(cpu, topo);
}
else {
printWarn("Can't read topology information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x00000001, cpu->maxLevels);
topo->physical_cores = 1;
topo->logical_cores = 1;
topo->smt_available = 1;
topo->smt_supported = 1;
}
break;
case CPU_VENDOR_AMD:
if (cpu->maxExtendedLevels >= 0x80000008) {
eax = 0x80000008;
cpuid(&eax, &ebx, &ecx, &edx);
topo->logical_cores = (ecx & 0xFF) + 1;
if (cpu->maxExtendedLevels >= 0x8000001E) {
eax = 0x8000001E;
cpuid(&eax, &ebx, &ecx, &edx);
topo->smt_supported = ((ebx >> 8) & 0x03) + 1;
}
else {
printWarn("Can't read topology information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x8000001E, cpu->maxExtendedLevels);
topo->smt_supported = 1;
}
}
else {
printWarn("Can't read topology information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x80000008, cpu->maxExtendedLevels);
topo->physical_cores = 1;
topo->logical_cores = 1;
topo->smt_supported = 1;
}
if (cpu->maxLevels >= 0x00000001) {
if(topo->smt_supported > 1)
topo->smt_available = is_smt_enabled_amd(topo);
else
topo->smt_available = 1;
}
else {
printWarn("Can't read topology information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x0000000B, cpu->maxLevels);
topo->smt_available = 1;
}
topo->physical_cores = topo->logical_cores / topo->smt_available;
if(topo->smt_supported > 1)
topo->sockets = topo->total_cores / topo->smt_supported / topo->physical_cores; // Idea borrowed from lscpu
else
topo->sockets = topo->total_cores / topo->physical_cores;
get_cache_topology_amd(cpu, topo);
break;
default:
printBug("Cant get topology because VENDOR is empty");
return NULL;
}
return topo;
}
struct cache* get_cache_info_amd_fallback(struct cache* cach) {
uint32_t eax = 0x80000005;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
cach->L1d->size = (ecx >> 24) * 1024;
cach->L1i->size = (edx >> 24) * 1024;
eax = 0x80000006;
cpuid(&eax, &ebx, &ecx, &edx);
cach->L2->size = (ecx >> 16) * 1024;
cach->L3->size = (edx >> 18) * 512 * 1024;
cach->L1i->exists = cach->L1i->size > 0;
cach->L1d->exists = cach->L1d->size > 0;
cach->L2->exists = cach->L2->size > 0;
cach->L3->exists = cach->L3->size > 0;
if(cach->L3->exists)
cach->max_cache_level = 4;
else
cach->max_cache_level = 3;
return cach;
}
struct cache* get_cache_info_general(struct cache* cach, uint32_t level) {
uint32_t eax = 0;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
int i=0;
int32_t cache_type;
do {
eax = level; // get cache info
ebx = 0;
ecx = i; // cache id
edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
cache_type = eax & 0x1F;
// If its 0, we tried fetching a non existing cache
if (cache_type > 0) {
int32_t cache_level = (eax >>= 5) & 0x7;
uint32_t cache_sets = ecx + 1;
uint32_t cache_coherency_line_size = (ebx & 0xFFF) + 1;
uint32_t cache_physical_line_partitions = ((ebx >>= 12) & 0x3FF) + 1;
uint32_t cache_ways_of_associativity = ((ebx >>= 10) & 0x3FF) + 1;
int32_t cache_total_size = cache_ways_of_associativity * cache_physical_line_partitions * cache_coherency_line_size * cache_sets;
cach->max_cache_level++;
switch (cache_type) {
case 1: // Data Cache (We assume this is L1d)
if(cache_level != 1) {
printBug("Found data cache at level %d (expected 1)", cache_level);
return NULL;
}
cach->L1d->size = cache_total_size;
cach->L1d->exists = true;
break;
case 2: // Instruction Cache (We assume this is L1i)
if(cache_level != 1) {
printBug("Found instruction cache at level %d (expected 1)", cache_level);
return NULL;
}
cach->L1i->size = cache_total_size;
cach->L1i->exists = true;
break;
case 3: // Unified Cache (This may be L2 or L3)
if(cache_level == 2) {
cach->L2->size = cache_total_size;
cach->L2->exists = true;
}
else if(cache_level == 3) {
cach->L3->size = cache_total_size;
cach->L3->exists = true;
}
else {
printWarn("Found unknown unified cache at level %d (size is %d bytes)", cache_level, cache_total_size);
}
break;
default: // Unknown cache type
printBug("Unknown cache type %d with level %d found at i=%d", cache_type, cache_level, i);
return NULL;
}
}
i++;
} while (cache_type > 0);
return cach;
}
struct cache* get_cache_info(struct cpuInfo* cpu) {
struct cache* cach = malloc(sizeof(struct cache));
init_cache_struct(cach);
uint32_t level;
// We use standard 0x00000004 for Intel
// We use extended 0x8000001D for AMD
// or 0x80000005/6 for old AMD
if(cpu->cpu_vendor == CPU_VENDOR_INTEL) {
level = 0x00000004;
if(cpu->maxLevels < level) {
printWarn("Can't read cache information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", level, cpu->maxLevels);
return NULL;
}
else {
cach = get_cache_info_general(cach, level);
}
}
else {
level = 0x8000001D;
if(cpu->maxExtendedLevels < level) {
printWarn("Can't read cache information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", level, cpu->maxExtendedLevels);
level = 0x80000006;
if(cpu->maxExtendedLevels < level) {
printWarn("Can't read cache information from cpuid using old method (needed extended level is 0x%.8X, max is 0x%.8X)", level, cpu->maxExtendedLevels);
return NULL;
}
printWarn("Fallback to old method using 0x%.8X and 0x%.8X", level-1, level);
cach = get_cache_info_amd_fallback(cach);
}
else {
cach = get_cache_info_general(cach, level);
}
}
// Sanity checks. If we read values greater than this, they can't be valid ones
// The values were chosen by me
if(cach->L1i->size > 64 * 1024) {
printBug("Invalid L1i size: %dKB", cach->L1i->size/1024);
}
if(cach->L1d->size > 64 * 1024) {
printBug("Invalid L1d size: %dKB", cach->L1d->size/1024);
}
if(cach->L2->exists) {
if(cach->L3->exists && cach->L2->size > 2 * 1048576) {
printBug("Invalid L2 size: %dMB", cach->L2->size/(1048576));
}
else if(cach->L2->size > 100 * 1048576) {
printBug("Invalid L2 size: %dMB", cach->L2->size/(1048576));
}
}
if(cach->L3->exists && cach->L3->size > 100 * 1048576) {
printBug("Invalid L3 size: %dMB", cach->L3->size/(1048576));
}
if(!cach->L2->exists) {
printBug("Could not find L2 cache");
}
return cach;
}
struct frequency* get_frequency_info(struct cpuInfo* cpu) {
struct frequency* freq = malloc(sizeof(struct frequency));
if(cpu->maxLevels < 0x00000016) {
#if defined (_WIN32) || defined (__APPLE__)
printWarn("Can't read frequency information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x00000016, cpu->maxLevels);
freq->base = UNKNOWN_FREQ;
freq->max = UNKNOWN_FREQ;
#else
printWarn("Can't read frequency information from cpuid (needed level is 0x%.8X, max is 0x%.8X). Using udev", 0x00000016, cpu->maxLevels);
freq->base = UNKNOWN_FREQ;
freq->max = get_max_freq_from_file(0, cpu->hv->present);
if(freq->max == 0) {
if(cpu->hv->present) {
printWarn("Read max CPU frequency and got 0 MHz");
}
else {
printBug("Read max CPU frequency and got 0 MHz");
}
freq->max = UNKNOWN_FREQ;
}
#endif
}
else {
uint32_t eax = 0x00000016;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
freq->base = eax;
freq->max = ebx;
if(freq->base == 0) {
if(cpu->hv->present) {
printWarn("Read base CPU frequency and got 0 MHz");
}
else {
printBug("Read base CPU frequency and got 0 MHz");
}
freq->base = UNKNOWN_FREQ;
}
if(freq->max == 0) {
if(cpu->hv->present) {
printWarn("Read max CPU frequency and got 0 MHz");
}
else {
printBug("Read max CPU frequency and got 0 MHz");
}
freq->max = UNKNOWN_FREQ;
}
}
return freq;
}
/*** STRING FUNCTIONS ***/
char* get_str_peak_performance(struct cpuInfo* cpu, struct topology* topo, int64_t freq) {
/***
PP = PeakPerformance
SP = SinglePrecision
PP(SP) =
N_CORES *
FREQUENCY *
2(Two vector units) *
2(If cpu has fma) *
16(If AVX512), 8(If AVX), 4(If SSE) *
***/
//7 for GFLOP/s and 6 for digits,eg 412.14
uint32_t size = 7+6+1+1;
assert(strlen(STRING_UNKNOWN)+1 <= size);
char* string = malloc(sizeof(char)*size);
//First check we have consistent data
if(freq == UNKNOWN_FREQ) {
snprintf(string,strlen(STRING_UNKNOWN)+1,STRING_UNKNOWN);
return string;
}
struct features* feat = cpu->feat;
double flops = topo->physical_cores * topo->sockets * (freq*1000000);
int vpus = get_number_of_vpus(cpu);
flops = flops * vpus;
if(feat->FMA3 || feat->FMA4)
flops = flops*2;
// Ice Lake has AVX512, but it has 1 VPU for AVX512, while
// it has 2 for AVX2. If this is a Ice Lake CPU, we are computing
// the peak performance supposing AVX2, not AVX512
if(feat->AVX512 && vpus_are_AVX512(cpu))
flops = flops*16;
else if(feat->AVX || feat->AVX2)
flops = flops*8;
else if(feat->SSE)
flops = flops*4;
// See https://sites.utexas.edu/jdm4372/2018/01/22/a-peculiar-
// throughput-limitation-on-intels-xeon-phi-x200-knights-landing/
if(is_knights_landing(cpu))
flops = flops * 6 / 7;
if(flops >= (double)1000000000000.0)
snprintf(string,size,"%.2f TFLOP/s",flops/1000000000000);
else if(flops >= 1000000000.0)
snprintf(string,size,"%.2f GFLOP/s",flops/1000000000);
else
snprintf(string,size,"%.2f MFLOP/s",flops/1000000);
return string;
}
// TODO: Refactoring
char* get_str_topology(struct cpuInfo* cpu, struct topology* topo, bool dual_socket) {
char* string;
if(topo->smt_supported > 1) {
//3 for digits, 21 for ' cores (SMT disabled)' which is the longest possible output
uint32_t size = 3+21+1;
string = malloc(sizeof(char)*size);
if(dual_socket) {
if(topo->smt_available > 1)
snprintf(string, size, "%d cores (%d threads)",topo->physical_cores * topo->sockets, topo->logical_cores * topo->sockets);
else {
if(cpu->cpu_vendor == CPU_VENDOR_AMD)
snprintf(string, size, "%d cores (SMT disabled)",topo->physical_cores * topo->sockets);
else
snprintf(string, size, "%d cores (HT disabled)",topo->physical_cores * topo->sockets);
}
}
else {
if(topo->smt_available > 1)
snprintf(string, size, "%d cores (%d threads)",topo->physical_cores,topo->logical_cores);
else {
if(cpu->cpu_vendor == CPU_VENDOR_AMD)
snprintf(string, size, "%d cores (SMT disabled)",topo->physical_cores);
else
snprintf(string, size, "%d cores (HT disabled)",topo->physical_cores);
}
}
}
else {
uint32_t size = 3+7+1;
string = malloc(sizeof(char)*size);
if(dual_socket)
snprintf(string, size, "%d cores",topo->physical_cores * topo->sockets);
else
snprintf(string, size, "%d cores",topo->physical_cores);
}
return string;
}
char* get_str_avx(struct cpuInfo* cpu) {
//If all AVX are available, it will use up to 15
char* string = malloc(sizeof(char)*17+1);
if(!cpu->feat->AVX)
snprintf(string,2+1,"No");
else if(!cpu->feat->AVX2)
snprintf(string,3+1,"AVX");
else if(!cpu->feat->AVX512)
snprintf(string,8+1,"AVX,AVX2");
else
snprintf(string,15+1,"AVX,AVX2,AVX512");
return string;
}
char* get_str_sse(struct cpuInfo* cpu) {
uint32_t last = 0;
uint32_t SSE_sl = 4;
uint32_t SSE2_sl = 5;
uint32_t SSE3_sl = 5;
uint32_t SSSE3_sl = 6;
uint32_t SSE4a_sl = 6;
uint32_t SSE4_1_sl = 7;
uint32_t SSE4_2_sl = 7;
char* string = malloc(sizeof(char)*SSE_sl+SSE2_sl+SSE3_sl+SSSE3_sl+SSE4a_sl+SSE4_1_sl+SSE4_2_sl+1);
if(cpu->feat->SSE) {
snprintf(string+last,SSE_sl+1,"SSE,");
last+=SSE_sl;
}
if(cpu->feat->SSE2) {
snprintf(string+last,SSE2_sl+1,"SSE2,");
last+=SSE2_sl;
}
if(cpu->feat->SSE3) {
snprintf(string+last,SSE3_sl+1,"SSE3,");
last+=SSE3_sl;
}
if(cpu->feat->SSSE3) {
snprintf(string+last,SSSE3_sl+1,"SSSE3,");
last+=SSSE3_sl;
}
if(cpu->feat->SSE4a) {
snprintf(string+last,SSE4a_sl+1,"SSE4a,");
last+=SSE4a_sl;
}
if(cpu->feat->SSE4_1) {
snprintf(string+last,SSE4_1_sl+1,"SSE4.1,");
last+=SSE4_1_sl;
}
if(cpu->feat->SSE4_2) {
snprintf(string+last,SSE4_2_sl+1,"SSE4.2,");
last+=SSE4_2_sl;
}
//Purge last comma
string[last-1] = '\0';
return string;
}
char* get_str_fma(struct cpuInfo* cpu) {
char* string = malloc(sizeof(char)*9+1);
if(!cpu->feat->FMA3)
snprintf(string,2+1,"No");
else if(!cpu->feat->FMA4)
snprintf(string,4+1,"FMA3");
else
snprintf(string,9+1,"FMA3,FMA4");
return string;
}
void print_debug(struct cpuInfo* cpu) {
uint32_t eax = 0x00000001;
uint32_t ebx = 0;
uint32_t ecx = 0;
uint32_t edx = 0;
cpuid(&eax, &ebx, &ecx, &edx);
printf("%s\n", cpu->cpu_name);
if(cpu->hv->present) printf("- Hypervisor: %s\n", cpu->hv->hv_name);
printf("- Max standard level: 0x%.8X\n", cpu->maxLevels);
printf("- Max extended level: 0x%.8X\n", cpu->maxExtendedLevels);
printf("- CPUID dump: 0x%.8X\n", eax);
free_cpuinfo_struct(cpu);
}
void free_topo_struct(struct topology* topo) {
free(topo->apic->cache_select_mask);
free(topo->apic->cache_id_apic);
free(topo->apic);
free(topo);
}

21
src/x86/cpuid.h Normal file
View File

@@ -0,0 +1,21 @@
#ifndef __CPUID__
#define __CPUID__
#include "../common/cpu.h"
struct cpuInfo* get_cpu_info();
struct cache* get_cache_info(struct cpuInfo* cpu);
struct frequency* get_frequency_info(struct cpuInfo* cpu);
struct topology* get_topology_info(struct cpuInfo* cpu, struct cache* cach);
char* get_str_avx(struct cpuInfo* cpu);
char* get_str_sse(struct cpuInfo* cpu);
char* get_str_fma(struct cpuInfo* cpu);
char* get_str_topology(struct cpuInfo* cpu, struct topology* topo, bool dual_socket);
char* get_str_peak_performance(struct cpuInfo* cpu, struct topology* topo, int64_t freq);
void print_debug(struct cpuInfo* cpu);
void free_topo_struct(struct topology* topo);
#endif

419
src/x86/uarch.c Normal file
View File

@@ -0,0 +1,419 @@
#include <stdbool.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "uarch.h"
#include "../common/global.h"
/*
* - cpuid codes are based on Todd Allen's cpuid program
* http://www.etallen.com/cpuid.html
* - This should be updated from time to time, to support newer CPUs. A good reference to look at:
* https://en.wikichip.org/
* http://instlatx64.atw.hu/
*/
// From Todd Allen:
//
// MSR_CPUID_table* is a table that appears in Intel document 325462, "Intel 64
// and IA-32 Architectures Software Developer's Manual Combined Volumes: 1, 2A,
// 2B, 2C, 2D, 3A, 3B, 3C, 3D, and 4" (the name changes from version to version
// as more volumes are added). The table moves around from version to version,
// but in version 071US, was in "Volume 4: Model-Specific Registers", Table 2-1:
// "CPUID Signature Values of DisplayFamily_DisplayModel".
// MRG* is a table that forms the bulk of Intel Microcode Revision Guidance (or
// Microcode Update Guidance). Its purpose is not to list CPUID values, but
// it does so, and sometimes lists values that appear nowhere else.
// LX* indicates features that I have seen no documentation for, but which are
// used by the Linux kernel (which is good evidence that they're correct).
// The "hook" to find these generally is an X86_FEATURE_* flag in:
// arch/x86/include/asm/cpufeatures.h
// For (synth) and (uarch synth) decoding, it often indicates
// family/model/stepping value which are documented nowhere else. These usually
// can be found in:
// arch/x86/include/asm/intel-family.h
typedef uint32_t MICROARCH;
// Data not available
#define NA -1
// Unknown manufacturing process
#define UNK -1
enum {
UARCH_UNKNOWN,
// INTEL //
UARCH_P5,
UARCH_P6,
UARCH_DOTHAN,
UARCH_YONAH,
UARCH_MEROM,
UARCH_PENYR,
UARCH_NEHALEM,
UARCH_WESTMERE,
UARCH_BONNELL,
UARCH_SALTWELL,
UARCH_SANDY_BRIDGE,
UARCH_SILVERMONT,
UARCH_IVY_BRIDGE,
UARCH_HASWELL,
UARCH_BROADWELL,
UARCH_AIRMONT,
UARCH_KABY_LAKE,
UARCH_COMET_LAKE,
UARCH_ROCKET_LAKE,
UARCH_AMBER_LAKE,
UARCH_WHISKEY_LAKE,
UARCH_SKYLAKE,
UARCH_CASCADE_LAKE,
UARCH_COOPER_LAKE,
UARCH_KNIGHTS_LANDING,
UARCH_KNIGHTS_MILL,
UARCH_GOLDMONT,
UARCH_PALM_COVE,
UARCH_SUNNY_COVE,
UARCH_GOLDMONT_PLUS,
UARCH_TREMONT,
UARCH_WILLOW_COVE,
UARCH_COFFE_LAKE,
UARCH_ITANIUM,
UARCH_KNIGHTS_FERRY,
UARCH_KNIGHTS_CORNER,
UARCH_WILLAMETTE,
UARCH_NORTHWOOD,
UARCH_PRESCOTT,
UARCH_CEDAR_MILL,
UARCH_ITANIUM2,
UARCH_ICE_LAKE,
// AMD //
UARCH_AM486,
UARCH_AM5X86,
UARCH_K6,
UARCH_K7,
UARCH_K8,
UARCH_K10,
UARCH_PUMA_2008,
UARCH_BOBCAT,
UARCH_BULLDOZER,
UARCH_PILEDRIVER,
UARCH_STEAMROLLER,
UARCH_EXCAVATOR,
UARCH_JAGUAR,
UARCH_PUMA_2014,
UARCH_ZEN,
UARCH_ZEN_PLUS,
UARCH_ZEN2,
UARCH_ZEN3
};
struct uarch {
MICROARCH uarch;
char* uarch_str;
int32_t process; // measured in nanometers
};
#define UARCH_START if (false) {}
#define CHECK_UARCH(arch, ef_, f_, em_, m_, s_, str, uarch, process) \
else if (ef_ == ef && f_ == f && (em_ == NA || em_ == em) && (m_ == NA || m_ == m) && (s_ == NA || s_ == s)) fill_uarch(arch, str, uarch, process);
#define UARCH_END else { printBug("Unknown microarchitecture detected: M=0x%.8X EM=0x%.8X F=0x%.8X EF=0x%.8X S=0x%.8X", m, em, f, ef, s); fill_uarch(arch, "Unknown", UARCH_UNKNOWN, 0); }
void fill_uarch(struct uarch* arch, char* str, MICROARCH u, uint32_t process) {
arch->uarch_str = malloc(sizeof(char) * (strlen(str)+1));
strcpy(arch->uarch_str, str);
arch->uarch = u;
arch->process= process;
}
// Inspired in Todd Allen's decode_uarch_intel
struct uarch* get_uarch_from_cpuid_intel(uint32_t ef, uint32_t f, uint32_t em, uint32_t m, int s) {
struct uarch* arch = malloc(sizeof(struct uarch));
// EF: Extended Family //
// F: Family //
// EM: Extended Model //
// M: Model //
// S: Stepping //
// ----------------------------------------------------------------------------- //
// EF F EM M S //
UARCH_START
CHECK_UARCH(arch, 0, 5, 0, 0, NA, "P5", UARCH_P5, 800)
CHECK_UARCH(arch, 0, 5, 0, 1, NA, "P5", UARCH_P5, 800)
CHECK_UARCH(arch, 0, 5, 0, 2, NA, "P5", UARCH_P5, UNK)
CHECK_UARCH(arch, 0, 5, 0, 3, NA, "P5", UARCH_P5, 600)
CHECK_UARCH(arch, 0, 5, 0, 4, NA, "P5 MMX", UARCH_P5, UNK)
CHECK_UARCH(arch, 0, 5, 0, 7, NA, "P5 MMX", UARCH_P5, UNK)
CHECK_UARCH(arch, 0, 5, 0, 8, NA, "P5 MMX", UARCH_P5, 250)
CHECK_UARCH(arch, 0, 5, 0, 9, NA, "P5 MMX", UARCH_P5, UNK)
CHECK_UARCH(arch, 0, 6, 0, 0, NA, "P6 Pentium II", UARCH_P6, UNK)
CHECK_UARCH(arch, 0, 6, 0, 1, NA, "P6 Pentium II", UARCH_P6, UNK) // process depends on core
CHECK_UARCH(arch, 0, 6, 0, 2, NA, "P6 Pentium II", UARCH_P6, UNK)
CHECK_UARCH(arch, 0, 6, 0, 3, NA, "P6 Pentium II", UARCH_P6, 350)
CHECK_UARCH(arch, 0, 6, 0, 4, NA, "P6 Pentium II", UARCH_P6, UNK)
CHECK_UARCH(arch, 0, 6, 0, 5, NA, "P6 Pentium II", UARCH_P6, 250)
CHECK_UARCH(arch, 0, 6, 0, 6, NA, "P6 Pentium II", UARCH_P6, UNK)
CHECK_UARCH(arch, 0, 6, 0, 7, NA, "P6 Pentium III", UARCH_P6, 250)
CHECK_UARCH(arch, 0, 6, 0, 8, NA, "P6 Pentium III", UARCH_P6, 180)
CHECK_UARCH(arch, 0, 6, 0, 9, NA, "P6 Pentium M", UARCH_P6, 130)
CHECK_UARCH(arch, 0, 6, 0, 10, NA, "P6 Pentium III", UARCH_P6, 180)
CHECK_UARCH(arch, 0, 6, 0, 11, NA, "P6 Pentium III", UARCH_P6, 130)
CHECK_UARCH(arch, 0, 6, 0, 13, NA, "Dothan", UARCH_DOTHAN, UNK) // process depends on core
CHECK_UARCH(arch, 0, 6, 0, 14, NA, "Yonah", UARCH_YONAH, 65)
CHECK_UARCH(arch, 0, 6, 0, 15, NA, "Merom", UARCH_MEROM, 65)
CHECK_UARCH(arch, 0, 6, 1, 5, NA, "Dothan", UARCH_DOTHAN, 90)
CHECK_UARCH(arch, 0, 6, 1, 6, NA, "Merom", UARCH_MEROM, 65)
CHECK_UARCH(arch, 0, 6, 1, 7, NA, "Penryn", UARCH_PENYR, 45)
CHECK_UARCH(arch, 0, 6, 1, 10, NA, "Nehalem", UARCH_NEHALEM, 45)
CHECK_UARCH(arch, 0, 6, 1, 12, NA, "Bonnell", UARCH_BONNELL, 45)
CHECK_UARCH(arch, 0, 6, 1, 13, NA, "Penryn", UARCH_PENYR, 45)
CHECK_UARCH(arch, 0, 6, 1, 14, NA, "Nehalem", UARCH_NEHALEM, 45)
CHECK_UARCH(arch, 0, 6, 1, 15, NA, "Nehalem", UARCH_NEHALEM, 45)
CHECK_UARCH(arch, 0, 6, 2, 5, NA, "Westmere", UARCH_WESTMERE, 32)
CHECK_UARCH(arch, 0, 6, 2 , 6, NA, "Bonnell", UARCH_BONNELL, 45)
CHECK_UARCH(arch, 0, 6, 2, 7, NA, "Saltwell", UARCH_SALTWELL, 32)
CHECK_UARCH(arch, 0, 6, 2, 10, NA, "Sandy Bridge", UARCH_SANDY_BRIDGE, 32)
CHECK_UARCH(arch, 0, 6, 2, 12, NA, "Westmere", UARCH_WESTMERE, 32)
CHECK_UARCH(arch, 0, 6, 2, 13, NA, "Sandy Bridge", UARCH_SANDY_BRIDGE, 32)
CHECK_UARCH(arch, 0, 6, 2, 14, NA, "Nehalem", UARCH_NEHALEM, 45)
CHECK_UARCH(arch, 0, 6, 2, 15, NA, "Westmere", UARCH_WESTMERE, 32)
CHECK_UARCH(arch, 0, 6, 3, 5, NA, "Saltwell", UARCH_SALTWELL, 14)
CHECK_UARCH(arch, 0, 6, 3, 6, NA, "Saltwell", UARCH_SALTWELL, 32)
CHECK_UARCH(arch, 0, 6, 3, 7, NA, "Silvermont", UARCH_SILVERMONT, 22)
CHECK_UARCH(arch, 0, 6, 3, 10, NA, "Ivy Bridge", UARCH_IVY_BRIDGE, 22)
CHECK_UARCH(arch, 0, 6, 3, 12, NA, "Haswell", UARCH_HASWELL, 22)
CHECK_UARCH(arch, 0, 6, 3, 13, NA, "Broadwell", UARCH_BROADWELL, 14)
CHECK_UARCH(arch, 0, 6, 3, 14, NA, "Ivy Bridge", UARCH_IVY_BRIDGE, 22)
CHECK_UARCH(arch, 0, 6, 3, 15, NA, "Haswell", UARCH_HASWELL, 22)
CHECK_UARCH(arch, 0, 6, 4, 5, NA, "Haswell", UARCH_HASWELL, 22)
CHECK_UARCH(arch, 0, 6, 4, 6, NA, "Haswell", UARCH_HASWELL, 22)
CHECK_UARCH(arch, 0, 6, 4, 7, NA, "Broadwell", UARCH_BROADWELL, 14)
CHECK_UARCH(arch, 0, 6, 4, 10, NA, "Silvermont", UARCH_SILVERMONT, 22) // no docs, but /proc/cpuinfo seen in wild
CHECK_UARCH(arch, 0, 6, 4, 12, NA, "Airmont", UARCH_AIRMONT, 14)
CHECK_UARCH(arch, 0, 6, 4, 13, NA, "Silvermont", UARCH_SILVERMONT, 22)
CHECK_UARCH(arch, 0, 6, 4, 14, 8, "Kaby Lake", UARCH_KABY_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 4, 14, NA, "Skylake", UARCH_SKYLAKE, 14)
CHECK_UARCH(arch, 0, 6, 4, 15, NA, "Broadwell", UARCH_BROADWELL, 14)
CHECK_UARCH(arch, 0, 6, 5, 5, 6, "Cascade Lake", UARCH_CASCADE_LAKE, 14) // no docs, but example from Greg Stewart
CHECK_UARCH(arch, 0, 6, 5, 5, 7, "Cascade Lake", UARCH_CASCADE_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 5, 5, 10, "Cooper Lake", UARCH_COOPER_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 5, 5, NA, "Skylake", UARCH_SKYLAKE, 14)
CHECK_UARCH(arch, 0, 6, 5, 6, NA, "Broadwell", UARCH_BROADWELL, 14)
CHECK_UARCH(arch, 0, 6, 5, 7, NA, "Knights Landing", UARCH_KNIGHTS_LANDING, 14)
CHECK_UARCH(arch, 0, 6, 5, 10, NA, "Silvermont", UARCH_SILVERMONT, 22) // no spec update; only MSR_CPUID_table* so far
CHECK_UARCH(arch, 0, 6, 5, 12, NA, "Goldmont", UARCH_GOLDMONT, 14)
CHECK_UARCH(arch, 0, 6, 5, 13, NA, "Silvermont", UARCH_SILVERMONT, 22) // no spec update; only MSR_CPUID_table* so far
CHECK_UARCH(arch, 0, 6, 5, 14, 8, "Kaby Lake", UARCH_KABY_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 5, 14, NA, "Skylake", UARCH_SKYLAKE, 14)
CHECK_UARCH(arch, 0, 6, 5, 15, NA, "Goldmont", UARCH_GOLDMONT, 14)
CHECK_UARCH(arch, 0, 6, 6, 6, NA, "Palm Cove", UARCH_PALM_COVE, 10) // no spec update; only MSR_CPUID_table* so far
CHECK_UARCH(arch, 0, 6, 6, 10, NA, "Sunny Cove", UARCH_SUNNY_COVE, 10) // no spec update; only MSR_CPUID_table* so far
CHECK_UARCH(arch, 0, 6, 6, 12, NA, "Sunny Cove", UARCH_SUNNY_COVE, 10) // no spec update; only MSR_CPUID_table* so far
CHECK_UARCH(arch, 0, 6, 7, 5, NA, "Airmont", UARCH_AIRMONT, 14) // no spec update; whispers & rumors
CHECK_UARCH(arch, 0, 6, 7, 10, NA, "Goldmont Plus", UARCH_GOLDMONT_PLUS, 14)
CHECK_UARCH(arch, 0, 6, 7, 13, NA, "Sunny Cove", UARCH_SUNNY_COVE, 10) // no spec update; only MSR_CPUID_table* so far
CHECK_UARCH(arch, 0, 6, 7, 14, NA, "Ice Lake", UARCH_ICE_LAKE, 10)
CHECK_UARCH(arch, 0, 6, 8, 5, NA, "Knights Mill", UARCH_KNIGHTS_MILL, 14) // no spec update; only MSR_CPUID_table* so far
CHECK_UARCH(arch, 0, 6, 8, 6, NA, "Tremont", UARCH_TREMONT, 10) // LX*
CHECK_UARCH(arch, 0, 6, 8, 10, NA, "Tremont", UARCH_TREMONT, 10) // no spec update; only geekbench.com example
CHECK_UARCH(arch, 0, 6, 8, 12, NA, "Willow Cove", UARCH_WILLOW_COVE, 10) // found only on en.wikichip.org
CHECK_UARCH(arch, 0, 6, 8, 13, NA, "Willow Cove", UARCH_WILLOW_COVE, 10) // LX*
CHECK_UARCH(arch, 0, 6, 8, 14, 9, "Amber Lake", UARCH_AMBER_LAKE, 14) // wikichip
CHECK_UARCH(arch, 0, 6, 8, 14, 10, "Kaby Lake", UARCH_KABY_LAKE, 14) // wikichip
CHECK_UARCH(arch, 0, 6, 8, 14, 11, "Whiskey Lake", UARCH_WHISKEY_LAKE, 14) // wikichip
CHECK_UARCH(arch, 0, 6, 8, 14, 12, "Comet Lake", UARCH_COMET_LAKE, 14) // wikichip
CHECK_UARCH(arch, 0, 6, 9, 6, NA, "Tremont", UARCH_TREMONT, 10) // LX*
CHECK_UARCH(arch, 0, 6, 9, 12, NA, "Tremont", UARCH_TREMONT, 10) // LX*
CHECK_UARCH(arch, 0, 6, 9, 13, NA, "Sunny Cove", UARCH_SUNNY_COVE, 10) // LX*
CHECK_UARCH(arch, 0, 6, 9, 14, 9, "Kaby Lake", UARCH_KABY_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 9, 14, 10, "Coffee Lake", UARCH_COFFE_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 9, 14, 11, "Coffee Lake", UARCH_COFFE_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 9, 14, 12, "Coffee Lake", UARCH_COFFE_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 9, 14, 13, "Coffee Lake", UARCH_COFFE_LAKE, 14)
CHECK_UARCH(arch, 0, 6, 10, 5, NA, "Comet Lake", UARCH_COMET_LAKE, 14) // wikichip
CHECK_UARCH(arch, 0, 6, 10, 6, NA, "Comet Lake", UARCH_COMET_LAKE, 14) // instlatx64.atw.hu (i7-10710U)
CHECK_UARCH(arch, 0, 6, 10, 7, NA, "Rocket Lake", UARCH_ROCKET_LAKE, 14) // instlatx64.atw.hu (i7-11700K)
CHECK_UARCH(arch, 0, 11, 0, 0, NA, "Knights Ferry", UARCH_KNIGHTS_FERRY, 45) // found only on en.wikichip.org
CHECK_UARCH(arch, 0, 11, 0, 1, NA, "Knights Corner", UARCH_KNIGHTS_CORNER, 22)
CHECK_UARCH(arch, 0, 15, 0, 0, NA, "Willamette", UARCH_WILLAMETTE, 180)
CHECK_UARCH(arch, 0, 15, 0, 1, NA, "Willamette", UARCH_WILLAMETTE, 180)
CHECK_UARCH(arch, 0, 15, 0, 2, NA, "Northwood", UARCH_NORTHWOOD, 130)
CHECK_UARCH(arch, 0, 15, 0, 3, NA, "Prescott", UARCH_PRESCOTT, 90)
CHECK_UARCH(arch, 0, 15, 0, 4, NA, "Prescott", UARCH_PRESCOTT, 90)
CHECK_UARCH(arch, 0, 15, 0, 6, NA, "Cedar Mill", UARCH_CEDAR_MILL, 65)
CHECK_UARCH(arch, 1, 15, 0, 0, NA, "Itanium2", UARCH_ITANIUM2, 180)
CHECK_UARCH(arch, 1, 15, 0, 1, NA, "Itanium2", UARCH_ITANIUM2, 130)
CHECK_UARCH(arch, 1, 15, 0, 2, NA, "Itanium2", UARCH_ITANIUM2, 130)
UARCH_END
return arch;
}
// iNApired in Todd Allen's decode_uarch_amd
struct uarch* get_uarch_from_cpuid_amd(uint32_t ef, uint32_t f, uint32_t em, uint32_t m, int s) {
struct uarch* arch = malloc(sizeof(struct uarch));
// EF: Extended Family //
// F: Family //
// EM: Extended Model //
// M: Model //
// S: Stepping //
// ----------------------------------------------------------------------------- //
// EF F EM M S //
UARCH_START
CHECK_UARCH(arch, 0, 4, 0, 3, NA, "Am486", UARCH_AM486, UNK)
CHECK_UARCH(arch, 0, 4, 0, 7, NA, "Am486", UARCH_AM486, UNK)
CHECK_UARCH(arch, 0, 4, 0, 8, NA, "Am486", UARCH_AM486, UNK)
CHECK_UARCH(arch, 0, 4, 0, 9, NA, "Am486", UARCH_AM486, UNK)
CHECK_UARCH(arch, 0, 4, NA, NA, NA, "Am5x86", UARCH_AM5X86, UNK)
CHECK_UARCH(arch, 0, 5, 0, 6, NA, "K6", UARCH_K6, 300)
CHECK_UARCH(arch, 0, 5, 0, 7, NA, "K6", UARCH_K6, 250) // *p from sandpile.org
CHECK_UARCH(arch, 0, 5, 0, 13, NA, "K6", UARCH_K6, 80) // *p from sandpile.org
CHECK_UARCH(arch, 0, 5, NA, NA, NA, "K6", UARCH_K6, UNK)
CHECK_UARCH(arch, 0, 6, 0, 1, NA, "K7", UARCH_K7, 250)
CHECK_UARCH(arch, 0, 6, 0, 2, NA, "K7", UARCH_K7, 180)
CHECK_UARCH(arch, 0, 6, NA, NA, NA, "K7", UARCH_K7, UNK)
CHECK_UARCH(arch, 0, 15, 0, 4, 8, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 0, 4, NA, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 0, 5, NA, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 0, 7, NA, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 0, 8, NA, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 0, 11, NA, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 0, 12, NA, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 0, 14, NA, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 0, 15, NA, "K8", UARCH_K8, 130)
CHECK_UARCH(arch, 0, 15, 1, 4, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 1, 5, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 1, 7, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 1, 8, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 1, 11, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 1, 12, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 1, 15, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 2, 1, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 2, 3, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 2, 4, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 2, 5, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 2, 7, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 2, 11, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 2, 12, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 2, 15, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 4, 1, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 4, 3, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 4, 8, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 4, 11, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 4, 12, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 4, 15, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 5, 13, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 5, 15, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 0, 15, 6, 8, NA, "K8", UARCH_K8, 65)
CHECK_UARCH(arch, 0, 15, 6, 11, NA, "K8", UARCH_K8, 65)
CHECK_UARCH(arch, 0, 15, 6, 12, NA, "K8", UARCH_K8, 65)
CHECK_UARCH(arch, 0, 15, 6, 15, NA, "K8", UARCH_K8, 65)
CHECK_UARCH(arch, 0, 15, 7, 12, NA, "K8", UARCH_K8, 65)
CHECK_UARCH(arch, 0, 15, 7, 15, NA, "K8", UARCH_K8, 65)
CHECK_UARCH(arch, 0, 15, 12, 1, NA, "K8", UARCH_K8, 90)
CHECK_UARCH(arch, 1, 15, 0, 0, NA, "K10", UARCH_K10, 65) // sandpile.org
CHECK_UARCH(arch, 1, 15, 0, 2, NA, "K10", UARCH_K10, 65)
CHECK_UARCH(arch, 1, 15, 0, 4, NA, "K10", UARCH_K10, 45)
CHECK_UARCH(arch, 1, 15, 0, 5, NA, "K10", UARCH_K10, 45)
CHECK_UARCH(arch, 1, 15, 0, 6, NA, "K10", UARCH_K10, 45)
CHECK_UARCH(arch, 1, 15, 0, 8, NA, "K10", UARCH_K10, 45)
CHECK_UARCH(arch, 1, 15, 0, 9, NA, "K10", UARCH_K10, 45)
CHECK_UARCH(arch, 1, 15, 0, 10, NA, "K10", UARCH_K10, 45)
CHECK_UARCH(arch, 2, 15, NA, NA, NA, "Puma 2008", UARCH_PUMA_2008, 65)
CHECK_UARCH(arch, 3, 15, NA, NA, NA, "K10", UARCH_K10, 32)
CHECK_UARCH(arch, 5, 15, NA, NA, NA, "Bobcat", UARCH_BOBCAT, 40)
CHECK_UARCH(arch, 6, 15, 0, 0, NA, "Bulldozer", UARCH_BULLDOZER, 32) // iNAtlatx64 engr sample
CHECK_UARCH(arch, 6, 15, 0, 1, NA, "Bulldozer", UARCH_BULLDOZER, 32)
CHECK_UARCH(arch, 6, 15, 0, 2, NA, "Piledriver", UARCH_PILEDRIVER, 32)
CHECK_UARCH(arch, 6, 15, 1, 0, NA, "Piledriver", UARCH_PILEDRIVER, 32)
CHECK_UARCH(arch, 6, 15, 1, 3, NA, "Piledriver", UARCH_PILEDRIVER, 32)
CHECK_UARCH(arch, 6, 15, 3, 0, NA, "Steamroller", UARCH_STEAMROLLER, 28)
CHECK_UARCH(arch, 6, 15, 3, 8, NA, "Steamroller", UARCH_STEAMROLLER, 28)
CHECK_UARCH(arch, 6, 15, 4, 0, NA, "Steamroller", UARCH_STEAMROLLER, 28) // Software Optimization Guide (15h) says it has the same iNAt latencies as (6,15),(3,x).
CHECK_UARCH(arch, 6, 15, 6, 0, NA, "Excavator", UARCH_EXCAVATOR, 28) // undocumented, but iNAtlatx64 samples
CHECK_UARCH(arch, 6, 15, 6, 5, NA, "Excavator", UARCH_EXCAVATOR, 28) // undocumented, but sample from Alexandros Couloumbis
CHECK_UARCH(arch, 6, 15, 7, 0, NA, "Excavator", UARCH_EXCAVATOR, 28)
CHECK_UARCH(arch, 7, 15, 0, 0, NA, "Jaguar", UARCH_JAGUAR, 28)
CHECK_UARCH(arch, 7, 15, 3, 0, NA, "Puma 2014", UARCH_PUMA_2014, 28)
CHECK_UARCH(arch, 8, 15, 0, 0, NA, "Zen", UARCH_ZEN, 14) // iNAtlatx64 engr sample
CHECK_UARCH(arch, 8, 15, 0, 1, NA, "Zen", UARCH_ZEN, 14)
CHECK_UARCH(arch, 8, 15, 0, 8, NA, "Zen+", UARCH_ZEN_PLUS, 12)
CHECK_UARCH(arch, 8, 15, 1, 1, NA, "Zen", UARCH_ZEN, 14) // found only on en.wikichip.org & iNAtlatx64 examples
CHECK_UARCH(arch, 8, 15, 1, 8, NA, "Zen+", UARCH_ZEN_PLUS, 12) // found only on en.wikichip.org
CHECK_UARCH(arch, 8, 15, 3, 1, NA, "Zen 2", UARCH_ZEN2, 7) // found only on en.wikichip.org
CHECK_UARCH(arch, 8, 15, 6, 0, NA, "Zen 2", UARCH_ZEN2, 7) // undocumented, geekbench.com example
CHECK_UARCH(arch, 8, 15, 7, 1, NA, "Zen 2", UARCH_ZEN2, 7) // undocumented, but samples from Steven Noonan
CHECK_UARCH(arch, 10, 15, NA, NA, NA, "Zen 3", UARCH_ZEN3, 7) // undocumented, LX*
UARCH_END
return arch;
}
struct uarch* get_uarch_from_cpuid(struct cpuInfo* cpu, uint32_t ef, uint32_t f, uint32_t em, uint32_t m, int s) {
if(cpu->cpu_vendor == CPU_VENDOR_INTEL)
return get_uarch_from_cpuid_intel(ef, f, em, m, s);
else
return get_uarch_from_cpuid_amd(ef, f, em, m, s);
}
bool vpus_are_AVX512(struct cpuInfo* cpu) {
return cpu->arch->uarch != UARCH_ICE_LAKE;
}
bool is_knights_landing(struct cpuInfo* cpu) {
return cpu->arch->uarch == UARCH_KNIGHTS_LANDING;
}
int get_number_of_vpus(struct cpuInfo* cpu) {
switch(cpu->arch->uarch) {
// Intel
case UARCH_HASWELL:
case UARCH_BROADWELL:
case UARCH_SKYLAKE:
case UARCH_CASCADE_LAKE:
case UARCH_KABY_LAKE:
case UARCH_COMET_LAKE:
case UARCH_ROCKET_LAKE:
case UARCH_AMBER_LAKE:
case UARCH_WHISKEY_LAKE:
case UARCH_COFFE_LAKE:
case UARCH_PALM_COVE:
case UARCH_KNIGHTS_LANDING:
case UARCH_KNIGHTS_MILL:
case UARCH_ICE_LAKE:
// AMD
case UARCH_ZEN2:
case UARCH_ZEN3:
return 2;
default:
return 1;
}
}
char* get_str_uarch(struct cpuInfo* cpu) {
return cpu->arch->uarch_str;
}
char* get_str_process(struct cpuInfo* cpu) {
char* str = malloc(sizeof(char) * (4+2+1));
uint32_t process = cpu->arch->process;
if(process > 100)
sprintf(str, "%.2fum", (double)process/100);
else
sprintf(str, "%dnm", process);
return str;
}
void free_uarch_struct(struct uarch* arch) {
free(arch->uarch_str);
free(arch);
}

18
src/x86/uarch.h Normal file
View File

@@ -0,0 +1,18 @@
#ifndef __UARCH__
#define __UARCH__
#include <stdint.h>
#include "cpuid.h"
struct uarch;
struct uarch* get_uarch_from_cpuid(struct cpuInfo* cpu, uint32_t ef, uint32_t f, uint32_t em, uint32_t m, int s);
bool vpus_are_AVX512(struct cpuInfo* cpu);
bool is_knights_landing(struct cpuInfo* cpu);
int get_number_of_vpus(struct cpuInfo* cpu);
char* get_str_uarch(struct cpuInfo* cpu);
char* get_str_process(struct cpuInfo* cpu);
void free_uarch_struct(struct uarch* arch);
#endif