Small fix

Makefile

@@ -1,19 +1,21 @@
 CXX=gcc
 
-CXXFLAGS=-Wall -Wextra -Werror -fstack-protector-all -pedantic -Wno-unused -std=c99
+CXXFLAGS=-Wall -Wextra -Werror -pedantic -fstack-protector-all -pedantic -std=c99
 SANITY_FLAGS=-Wfloat-equal -Wshadow -Wpointer-arith -Wstrict-overflow=5 -Wformat=2
 
 SRC_DIR=src/
-SOURCE=$(SRC_DIR)main.c $(SRC_DIR)standart.c $(SRC_DIR)extended.c $(SRC_DIR)cpuid.c $(SRC_DIR)printer.c $(SRC_DIR)args.c $(SRC_DIR)global.c
+SOURCE=$(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 $(SRC_DIR)uarch.c
-HEADERS=$(SRC_DIR)standart.h $(SRC_DIR)extended.h  $(SRC_DIR)cpuid.h $(SRC_DIR)printer.h $(SRC_DIR)ascii.h $(SRC_DIR)args.h $(SRC_DIR)global.h
+HEADERS=$(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 $(SRC_DIR)uarch.h
 
 ifneq ($(OS),Windows_NT)
 	SOURCE += $(SRC_DIR)udev.c
 	HEADERS += $(SRC_DIR)udev.h
+	OUTPUT=cpufetch	
+else
+	SANITY_FLAGS += -Wno-pedantic-ms-format
+	OUTPUT=cpufetch.exe
 endif
 
-OUTPUT=cpufetch
-
 all: $(OUTPUT)
 
 debug: CXXFLAGS += -g -O0

README.md

@@ -1,12 +1,28 @@
 # cpufetch
 
 Prints a fancy summary of the CPU with some advanced information
+![cpu1](i9.png)
 
 ### 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
+cpufetch currently supports x86 CPUs (both Intel and AMD CPUs)
+
+| Platform  | Intel                     | AMD                      | Notes             |
+|:---------:|:-------------------------:|:------------------------:|:-----------------:|
+| Linux     | :heavy_check_mark:        | :heavy_check_mark:       | Prefered platform |
+| Windows   | :heavy_check_mark:        | :heavy_check_mark:       | Some information may be missing. <br> No colors and worse CPU art |
+| macOS     | :heavy_exclamation_mark:  | :heavy_exclamation_mark: | Untested |
+
 
 ### Usage and installation
+#### Linux
+There is a cpufetch package available in Arch Linux ([cpufetch-git](https://aur.archlinux.org/packages/cpufetch-git)).
 
+If you are in other distro, you can build `cpufetch` from source (see below)
+
+#### 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.
+
+#### Building from source
 Just clone the repo and use `make` to compile it
 
 ```
@@ -16,38 +32,34 @@ make
 ./cpufetch
 ```
 
+The Makefile is designed to work on both Linux and Windows.
+
 ### Example
 
-This is the output of `cpufetch` in a i7-4790K
+Here are more examples of how `cpufetch` looks on different CPUs.
 
-![Example](/preview.png)
+![cpu2](epyc.png)
 
-### Output
+![cpu3](cascade_lake.png)
 
-Output is detailed as follows:
+### 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:
 
-| Field      | Description             | Possible Values  |
+```
-|:----------:|:-----------------------:|:-----------------:|
+./cpufetch --color intel (default color for Intel)
-| Name       | Name of the CPU   | Any valid CPU name |
+./cpufetch --color amd (default color for AND)
-| Frequency  | Max frequency of the CPU(in GHz) | X.XX(GHz or MHz)
+./cpufetch --color 239,90,45:210,200,200:100,200,45:0,200,200 (example)
-| 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 |
 
-`cpufetch` also prints a simple ascii art of the manufacturer logo.
+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.
 
 ### Implementation
 
-`cpufetch` makes use of two techniques to fetch data:
+`cpufetch` fetches all of the information using the `CPUID` x86 instruction. There are, however, some cases where the CPU does not support fetching some needed information. In this case, `cpufetch` will use `/sys/devices/system/cpu` in Linux as a fallback. If `cpufetch` is running on Windows and `CPUID` does not give all the data, `cpufetch` won't be able to show it. [I hope this can be fixed in the future](https://github.com/Dr-Noob/cpufetch/issues/30)
-* __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`
 
 ### Bugs or improvements
-Feel free to open a issue on the repo to report a issue or propose any improvement in the tool
+There are many open issues in github (see [issues](https://github.com/Dr-Noob/cpufetch/issues)). Feel free to open a new one report a issue or propose any improvement in `cpufetch`
+
+### Testers
+I would like to thank [Gonzalocl](https://github.com/Gonzalocl) and [OdnetninI](https://github.com/OdnetninI) for their help, running `cpufeth` in many different CPUs they have access to, which makes it easier to debug and check the correctness of `cpufetch`.

cpufetch.8

@@ -1,33 +1,57 @@
-.TH man 8 "22 Jun 2018" "0.32" "cpufetch man page"
+.TH man 8 "1 Sep 2020" "0.7" "cpufetch man page"
 .SH NAME
 cpufetch \- Prints a fancy summary of the CPU with some advanced information
 .SH SYNOPSIS
-cpufetch [--help] [--style STYLE]
+cpufetch [--version] [--help] [--levels] [--style fancy|retro|legacy] [--color intel|amd|'R,G,B:R,G,B:R,G,B:R,G,B']
 .SH DESCRIPTION
-cpufetch will print CPU information, for which will query cpuid instructions and udev directories on Linux. It should display:
+cpufetch will print CPU information, for which will query CPUID instructions and udev directories on Linux as a fallback method. Some of this features are:
 .IP \[bu] 2
 Name
 .IP \[bu]
 Frequency
 .IP \[bu]
-Number of cores(Physical and Logical)
+Number of cores (Physical and Logical)
 .IP \[bu]
-AVX,SSE,FMA,AES and SHA support
+Cache sizes
 .IP \[bu]
-L1,L2 and L3 size
+Theoretical peak performance in floating point operations per second (FLOP/s)
-.IP \[bu]
-Theoretical peak flops
 .SH OPTIONS
 .TP
+\fB\-\-style\fR \f[I][intel|amd|R,G,B:R,G,B:R,G,B:R,G,B]\f[]
+Set the color scheme. By default, cpufetch uses the system color scheme. This option lets the user use different colors to print the CPU art:
+.IP \[bu]
+\fB"intel"\fR: Use intel color scheme
+.IP \[bu]
+\fB"amd"\fR: Use amd color scheme
+.IP \[bu]
+\fBcustom\fR: If color do not match "intel" or "amd", a custom scheme can be specified: 4 colors must be given in RGB with the format: R,G,B:R,G,B:... 
+             These colors correspond to CPU art color (2 colors) and for the text colors (following 2) 
+.TP
+\fB\-\-style\fR \f[I]STYLE\f[]
+Specify the style of ascii logo:
+.IP \[bu] 
+\fB"fancy"\fR:  Default style
+.IP \[bu]
+\fB"retro"\fR:  Old cpufetch style
+.IP \[bu]
+\fB"legacy"\fR: Fallback style for terminals that does not support colors
+.TP
+\fB\-\-levels\fR
+Prints CPUID levels and CPU name
+.TP
+\fB\-\-verbose\fR
+Prints extra information (if available) about how cpufetch tried fetching information
+.TP
 \fB\-\-help\fR
 Prints help
 .TP
 \fB\-\-version\fR
 Prints cpufetch version
-.TP
-\fB\-\-style\fR
-Specify the color style of ascii logo
 .SH BUGS
-No known bugs. AMD CPUs may not be fully supported
+Bugs should be posted on: https://github.com/Dr-Noob/cpufetch/issues
+.SH NOTES
+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.
 .SH AUTHOR
 Dr-Noob (https://github.com/Dr-Noob)

src/apic.c

@@ -0,0 +1,352 @@
+#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;
+}
+
+// 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 > 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 get_topology_from_apic(struct cpuInfo* cpu, 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);
+  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 = cpu->maxLevels >= 0x0000000B;
+  
+  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);  
+  
+  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;
+    
+    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) {
+  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;  
+}

src/apic.h

@@ -0,0 +1,20 @@
+#ifndef __APIC__
+#define __APIC__
+
+#include <stdbool.h>
+#include "cpuid.h"
+
+struct apic {
+  uint32_t pkg_mask;
+  uint32_t pkg_mask_shift;
+  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(struct cpuInfo* cpu, struct topology* topo);
+uint32_t is_smt_enabled_amd(struct topology* topo);
+
+#endif

src/args.c

@@ -1,44 +1,51 @@
 #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_HELP     'h'
+#define ARG_CHAR_STYLE    0
-#define ARG_CHAR_LEVELS   'l'
+#define ARG_CHAR_COLOR    1
-#define ARG_CHAR_VERSION  'v'
+#define ARG_CHAR_HELP     2
-#define STYLE_STR_1 "default"
+#define ARG_CHAR_LEVELS   3
-#define STYLE_STR_2 "dark"
+#define ARG_CHAR_VERBOSE  4
-#define STYLE_STR_3 "none"
+#define ARG_CHAR_VERSION  5
+
+#define STYLE_STR_1      "fancy"
+#define STYLE_STR_2      "retro"
+#define STYLE_STR_3      "legacy"
+
+#define COLOR_STR_INTEL "intel"
+#define COLOR_STR_AMD   "amd"
 
 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 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;
-}
-
 STYLE get_style() {
   return args.style;
 }
 
+struct colors* get_colors() {
+  return args.colors;
+}
+
 bool show_help() {
   return args.help_flag;
 }
@@ -52,81 +59,192 @@ bool show_levels() {
 }
 
 bool verbose_enabled() {
-  return false;
+  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;
+  char* str_to_parse = NULL;
+  bool free_ptr;
+  
+  if(strcmp(optarg, 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, COLOR_STR_AMD) == 0) {
+    str_to_parse = malloc(sizeof(char) * 44);
+    strcpy(str_to_parse, COLOR_DEFAULT_AMD);  
+    free_ptr = true;
+  }
+  else {  
+    str_to_parse = optarg;
+    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;      
 }
 
 bool parse_args(int argc, char* argv[]) {
   int c;
-  int digit_optind = 0;
   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);
+  c = getopt_long(argc, argv, "", long_options, &option_index);
 
   while (c != -1) {
-     if(c == ARG_CHAR_STYLE) {
+     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) {
-         printf("ERROR: Style option specified more than once\n");
+         printErr("Style option specified more than once");
          return false;
        }
        args.style = parse_style(optarg);
        if(args.style == STYLE_INVALID) {
-          printf("ERROR: Invalid style '%s'\n",optarg);
+         printErr("Invalid style '%s'",optarg);
-          return false;
+         return false;
        }
      }
      else if(c == ARG_CHAR_HELP) {
        if(args.help_flag) {
-         printf("ERROR: Help option specified more than once\n");
+         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) {
-         printf("ERROR: Levels option specified more than once\n");
+         printErr("Levels option specified more than once");
          return false;
        }
        args.levels_flag  = true;
      }
      else if (c == ARG_CHAR_VERSION) {
        if(args.version_flag) {
-         printf("ERROR: Version option specified more than once\n");
+         printErr("Version option specified more than once");
          return false;
        }
        args.version_flag = true;
      }
      else if(c == '?') {
-       printf("WARNING: Invalid options\n");
+       printWarn("Invalid options");
        args.help_flag  = true;
        break;
      }
      else
-      printf("Bug at line number %d in file %s\n", __LINE__, __FILE__);
+      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) {
-    printf("WARNING: Invalid options\n");
+    printWarn("Invalid options");
     args.help_flag  = true;
   }
 
-  if((args.help_flag + args.version_flag + (args.style != STYLE_EMPTY)) > 1) {
+  if((args.help_flag + args.version_flag + color_flag) > 1) {
-    printf("WARNING: You should specify just one option\n");
+    printWarn("You should specify just one option");
     args.help_flag  = true;
   }
 

src/args.h

@@ -2,13 +2,30 @@
 #define __ARGS__
 
 #include <stdbool.h>
+#include <stdint.h>
+
+struct color {
+  int32_t R;
+  int32_t G;
+  int32_t B;
+};
+
+struct colors {
+  struct color* c1;
+  struct color* c2;
+  struct color* c3;
+  struct color* c4;
+};
+
 #include "printer.h"
 
 bool parse_args(int argc, char* argv[]);
-STYLE get_style();
 bool show_help();
 bool show_levels();
 bool show_version();
 bool verbose_enabled();
+void free_colors_struct(struct colors* cs);
+struct colors* get_colors();
+STYLE get_style();
 
 #endif

src/ascii.h

@@ -1,7 +1,7 @@
 #ifndef __ASCII__
 #define __ASCII__
 
-#define NUMBER_OF_LINES   20
+#define NUMBER_OF_LINES   19
 #define LINE_SIZE         62
 
 #define AMD_ASCII \
@@ -23,8 +23,7 @@
                                                               \
                                                               \
                                                               \
-                                                              \
+                                                             "
-                                                            "
 
 #define INTEL_ASCII \
 "                              ################                \
@@ -45,7 +44,6 @@
  ####                                        ####             \
    #####                               ##########             \
      ##########               ################                \
-         ###############################                      \
+         ###############################                     "
-                                                            "
 
 #endif

src/cpuid.c

@@ -1,10 +1,978 @@
-#include "cpuid.h"
+#ifdef _WIN32
+  #include <windows.h>
+#else
+  #include "udev.h"
+  #include <unistd.h>
+#endif
 
-void cpuid(uint32_t *eax, uint32_t *ebx, uint32_t *ecx, uint32_t *edx) {
+#include <stdio.h>
-        __asm volatile("cpuid"
+#include <stdlib.h>
-            : "=a" (*eax),
+#include <string.h>
-              "=b" (*ebx),
+#include <assert.h>
-              "=c" (*ecx),
+#include <stdbool.h>
-              "=d" (*edx)
+
-            : "0" (*eax), "2" (*ecx));
+#include "cpuid.h"
+#include "cpuid_asm.h"
+#include "global.h"
+#include "apic.h"
+#include "uarch.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 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, 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_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 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 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);        
+  }
+  
+  cpu->arch = get_cpu_uarch(cpu);
+
+  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 {
+              printBug("Found unified cache at level %d (expected == 2 or 3)", cache_level);
+              return false;
+            }
+            break;
+          
+          default: // Unknown Type Cache
+            printBug("Unknown Type Cache found at ID %d", 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 VENDOR_INTEL:
+      if (cpu->maxLevels >= 0x00000004) { 
+        get_topology_from_apic(cpu, 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 >= 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(struct cpuInfo* cpu) {
+  struct cache* cach = malloc(sizeof(struct cache));
+  init_cache_struct(cach);
+  
+  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;
+    }
+  }
+  
+  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 {
+            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;                  
+      }
+    }    
+    
+    i++;
+  } while (cache_type > 0);
+  
+  // 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);
+    return NULL;
+  }
+  if(cach->L1d->size > 64 * 1024) {
+    printBug("Invalid L1d size: %dKB", cach->L1d->size/1024);
+    return NULL;
+  }
+  if(cach->L2->exists) {
+    if(cach->L3->exists && cach->L2->size > 2 * 1048576) {
+      printBug("Invalid L2 size: %dMB", cach->L2->size/(1048576));
+      return NULL;
+    }
+    else if(cach->L2->size > 100 * 1048576) {
+      printBug("Invalid L2 size: %dMB", cach->L2->size/(1048576));
+      return NULL;
+    }
+  }
+  if(cach->L3->exists && cach->L3->size > 100 * 1048576) {
+    printBug("Invalid L3 size: %dMB", cach->L3->size/(1048576));
+    return NULL;
+  }
+  if(!cach->L2->exists) {
+    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;
+      freq->max = UNKNOWN_FREQ;
+    #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;
+      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->size);
+  printf("L1d=%dB\n",cach->L1d->size);
+  printf("L2=%dB\n",cach->L2->size);
+  printf("L3=%dB\n",cach->L3->size);
+}
+
+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_FREQ) {
+    snprintf(string,strlen(STRING_UNKNOWN)+1,STRING_UNKNOWN);
+    return string;
+  }
+
+  double flops = topo->physical_cores * topo->sockets * (freq*1000000);
+  int vpus = get_number_of_vpus(cpu);
+  
+  flops = flops * vpus; 
+
+  if(cpu->FMA3 || cpu->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(cpu->AVX512 && vpus_are_AVX512(cpu))
+    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)*17+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, 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 = (4+3+1);
+  assert(strlen(STRING_UNKNOWN)+1 <= size);
+  char* string = malloc(sizeof(char)*size);
+  if(freq->max == UNKNOWN_FREQ)
+    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) {
+  printf("%s\n", cpu->cpu_name);
+  printf("- Max standart level: 0x%.8X\n", cpu->maxLevels);
+  printf("- Max extended level: 0x%.8X\n", cpu->maxExtendedLevels);
+  
+  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);
+}
+
+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_cpuinfo_struct(struct cpuInfo* cpu) {
+  free_uarch_struct(cpu->arch);
+  free(cpu->cpu_name);
+  free(cpu);
 }

src/cpuid.h

@@ -3,6 +3,109 @@
 
 #include <stdint.h>
 
-void cpuid(uint32_t *eax, uint32_t *ebx, uint32_t *ecx, uint32_t *edx);
+#define VENDOR_EMPTY   0
+#define VENDOR_INTEL   1
+#define VENDOR_AMD     2
+#define VENDOR_INVALID 3
+
+#define UNKNOWN_FREQ -1
+
+typedef int32_t VENDOR;
+
+struct frequency;
+
+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 uarch* arch;
+};
+
+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 {
+  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;
+  struct cache* cach;
+};
+
+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, struct cache* cach);
+
+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);
+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);
+
+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);
+
+void free_cache_struct(struct cache* cach);
+void free_topo_struct(struct topology* topo);
+void free_freq_struct(struct frequency* freq);
+void free_cpuinfo_struct(struct cpuInfo* cpu);
+
+void debug_cpu_info(struct cpuInfo* cpu);
+void debug_cache(struct cache* cach);
+void debug_frequency(struct frequency* freq);
 
 #endif

src/cpuid_asm.c

@@ -0,0 +1,10 @@
+#include "cpuid_asm.h"
+
+void cpuid(uint32_t *eax, uint32_t *ebx, uint32_t *ecx, uint32_t *edx) {
+        __asm volatile("cpuid"
+            : "=a" (*eax),
+              "=b" (*ebx),
+              "=c" (*ecx),
+              "=d" (*edx)
+            : "0" (*eax), "2" (*ecx));
+}

src/cpuid_asm.h

@@ -0,0 +1,8 @@
+#ifndef __CPUID_ASM__
+#define __CPUID_ASM__
+
+#include <stdint.h>
+
+void cpuid(uint32_t *eax, uint32_t *ebx, uint32_t *ecx, uint32_t *edx);
+
+#endif

src/extended.c

@@ -1,96 +0,0 @@
-#include <stdio.h>
-#include <string.h>
-#include "extended.h"
-
-char* get_str_cpu_name() {
-  uint32_t eax = 0;
-  uint32_t ebx = 0;
-  uint32_t ecx = 0;
-  uint32_t edx = 0;
-  
-  char *name = malloc(sizeof(char)*64);
-  memset(name, 0, 64);
-
-  //First, check we can use extended
-  eax = 0x80000000;
-  cpuid(&eax, &ebx, &ecx, &edx);
-  if(eax < 0x80000001) {
-    char* none = malloc(sizeof(char)*64);
-    sprintf(none,"Unknown");
-    return none;
-  }
-
-  //We can, fetch name
-  eax = 0x80000002;
-  cpuid(&eax, &ebx, &ecx, &edx);
-
-  name[__COUNTER__] = eax       & MASK;
-  name[__COUNTER__] = (eax>>8)  & MASK;
-  name[__COUNTER__] = (eax>>16) & MASK;
-  name[__COUNTER__] = (eax>>24) & MASK;
-  name[__COUNTER__] = ebx       & MASK;
-  name[__COUNTER__] = (ebx>>8)  & MASK;
-  name[__COUNTER__] = (ebx>>16) & MASK;
-  name[__COUNTER__] = (ebx>>24) & MASK;
-  name[__COUNTER__] = ecx       & MASK;
-  name[__COUNTER__] = (ecx>>8)  & MASK;
-  name[__COUNTER__] = (ecx>>16) & MASK;
-  name[__COUNTER__] = (ecx>>24) & MASK;
-  name[__COUNTER__] = edx       & MASK;
-  name[__COUNTER__] = (edx>>8)  & MASK;
-  name[__COUNTER__] = (edx>>16) & MASK;
-  name[__COUNTER__] = (edx>>24) & MASK;
-
-  eax = 0x80000003;
-  cpuid(&eax, &ebx, &ecx, &edx);
-
-  name[__COUNTER__] = eax       & MASK;
-  name[__COUNTER__] = (eax>>8)  & MASK;
-  name[__COUNTER__] = (eax>>16) & MASK;
-  name[__COUNTER__] = (eax>>24) & MASK;
-  name[__COUNTER__] = ebx       & MASK;
-  name[__COUNTER__] = (ebx>>8)  & MASK;
-  name[__COUNTER__] = (ebx>>16) & MASK;
-  name[__COUNTER__] = (ebx>>24) & MASK;
-  name[__COUNTER__] = ecx       & MASK;
-  name[__COUNTER__] = (ecx>>8)  & MASK;
-  name[__COUNTER__] = (ecx>>16) & MASK;
-  name[__COUNTER__] = (ecx>>24) & MASK;
-  name[__COUNTER__] = edx       & MASK;
-  name[__COUNTER__] = (edx>>8)  & MASK;
-  name[__COUNTER__] = (edx>>16) & MASK;
-  name[__COUNTER__] = (edx>>24) & MASK;
-
-  eax = 0x80000004;
-  cpuid(&eax, &ebx, &ecx, &edx);
-
-  name[__COUNTER__] = eax       & MASK;
-  name[__COUNTER__] = (eax>>8)  & MASK;
-  name[__COUNTER__] = (eax>>16) & MASK;
-  name[__COUNTER__] = (eax>>24) & MASK;
-  name[__COUNTER__] = ebx       & MASK;
-  name[__COUNTER__] = (ebx>>8)  & MASK;
-  name[__COUNTER__] = (ebx>>16) & MASK;
-  name[__COUNTER__] = (ebx>>24) & MASK;
-  name[__COUNTER__] = ecx       & MASK;
-  name[__COUNTER__] = (ecx>>8)  & MASK;
-  name[__COUNTER__] = (ecx>>16) & MASK;
-  name[__COUNTER__] = (ecx>>24) & MASK;
-  name[__COUNTER__] = edx       & MASK;
-  name[__COUNTER__] = (edx>>8)  & MASK;
-  name[__COUNTER__] = (edx>>16) & MASK;
-  name[__COUNTER__] = (edx>>24) & MASK;
-
-  name[__COUNTER__] = '\0';
-
-  //Remove unused characters  
-  char *str = name;
-  char *dest = name;
-  while (*str != '\0') {
-    while (*str == ' ' && *(str + 1) == ' ') str++;
-    *dest++ = *str++;
-  }
-  *dest = '\0';
-  
-  return name;
-}

src/extended.h

@@ -1,11 +0,0 @@
-#ifndef __EXTENDED__
-#define __EXTENDED__
-
-#define MASK 0xFF
-#include "cpuid.h"
-#include <stdint.h>
-#include <stdlib.h>
-
-char* get_str_cpu_name();
-
-#endif

src/global.c

@@ -51,7 +51,7 @@ 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");
+  fprintf(stderr,"Please, create a new issue with this error message and your CPU model in https://github.com/Dr-Noob/cpufetch/issues\n");
 }
 
 void set_log_level(bool verbose) {

src/main.c

@@ -3,40 +3,36 @@
 
 #include "args.h"
 #include "printer.h"
-#include "standart.h"
+#include "cpuid.h"
-#include "extended.h"
 #include "global.h"
 
-/***
+static const char* VERSION = "0.7";
-SAMPLE OUTPUT
-
-Name:        Intel Core i7-4790K
-Frequency:   4.0 GHz
-NºCores:     4 cores(8 threads)
-AXV:         AVX,AVX2
-SSE:         SSE,SSE2,SSE4.1,SSE4.2
-FMA:         FMA3
-AES:         Yes
-SHA:         No
-L1 Size:     32KB(Data)32KB(Instructions)
-L2 Size:     512KB
-L3 Size:     8MB
-Peak FLOPS:  512 GFLOP/s(in simple precision)
-
-***/
-
-static const char* VERSION = "0.410";
 
 void print_help(char *argv[]) {
-  printf("Usage: %s [--version] [--help] [--style STYLE]\n\
+  printf("Usage: %s [--version] [--help] [--levels] [--style \"fancy\"|\"retro\"|\"legacy\"] [--color \"intel\"|\"amd\"|'R,G,B:R,G,B:R,G,B:R,G,B']\n\n\
 Options: \n\
-  --style    Set logo style color\n\
+  --color       Set the color scheme. By default, cpufetch uses the system color scheme. This option \n\
-    default:   Default style color\n\
+                lets the user use different colors to print the CPU art:  \n\
-    dark:      Dark style color\n\
+   * \"intel\":   Use intel default color scheme \n\
-    none:      Don't use colors\n\
+   * \"amd\":     Use amd default color scheme \n\
-  --help     Prints this help and exit\n\
+   * custom:    If color do not match \"intel\" or \"amd\", a custom scheme can be specified: \n\
-  --levels   Prints CPU model and cpuid levels (debug purposes)\n\
+                4 colors must be given in RGB with the format: R,G,B:R,G,B:... \n\
-  --version  Prints cpufetch version and exit\n",
+                These colors correspond to CPU art color (2 colors) and for the text colors (following 2) \n\
+                For example: --color 239,90,45:210,200,200:100,200,45:0,200,200 \n\n\
+  --style       Set the style of CPU art: \n\
+    * \"fancy\":  Default style      \n\
+    * \"retro\":  Old cpufetch style \n\
+    * \"legacy\": Fallback style for terminals that does not support colors                       \n\n\
+  --levels      Prints CPU model and cpuid levels (debug purposes)\n\n\
+  --verbose     Prints extra information (if available) about how cpufetch tried fetching information\n\n\
+  --help        Prints this help and exit\n\n\
+  --version     Prints cpufetch version and exit\n\n\
+                                                   \n\
+NOTES: \n\
+  - Bugs or improvements should be submitted to: github.com/Dr-Noob/cpufetch/issues        \n\
+  - Peak performance information is NOT accurate. cpufetch computes peak performance using the max   \n\
+    frequency. However, to properly compute peak performance, you need to know the frequency of the  \n\
+    CPU running AVX code, which is not be fetched by cpufetch since it depends on each specific CPU. \n",
   argv[0]);
 }
 
@@ -63,75 +59,27 @@ int main(int argc, char* argv[]) {
   struct cpuInfo* cpu = get_cpu_info();
   if(cpu == NULL)
     return EXIT_FAILURE;
-  char* cpuName = get_str_cpu_name();
   
   if(show_levels()) {
     print_version();
-    print_levels(cpu, cpuName);
+    print_levels(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);
+  struct cache* cach = get_cache_info(cpu);
+  if(cach == NULL)
+    return EXIT_FAILURE;
+  
+  struct topology* topo = get_topology_info(cpu, cach);
   if(topo == NULL)
     return EXIT_FAILURE;
   
-  struct ascii* art = set_ascii(get_cpu_vendor(cpu),get_style());
+  if(print_cpufetch(cpu, cach, freq, topo, get_style(), get_colors()))  
-  if(art == NULL)
+    return EXIT_SUCCESS;
+  else
     return EXIT_FAILURE;
-  
-  char* maxFrequency = get_str_freq(freq);
-  char* nCores = get_str_topology(topo);
-  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* l1 = get_str_l1(cach);
-  char* l2 = get_str_l2(cach);
-  char* l3 = get_str_l3(cach);
-  char* pp = get_str_peak_performance(cpu,topo,get_freq(freq));
-
-  setAttribute(art,ATTRIBUTE_NAME,cpuName);
-  setAttribute(art,ATTRIBUTE_FREQUENCY,maxFrequency);
-  setAttribute(art,ATTRIBUTE_NCORES,nCores);
-  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_L1,l1);
-  setAttribute(art,ATTRIBUTE_L2,l2);
-  setAttribute(art,ATTRIBUTE_L3,l3);
-  setAttribute(art,ATTRIBUTE_PEAK,pp);
-
-  print_ascii(art);
-
-  free(cpuName);
-  free(maxFrequency);
-  free(nCores);
-  free(avx);
-  free(sse);
-  free(fma);
-  free(aes);
-  free(sha);
-  free(l1);
-  free(l2);
-  free(l3);
-  free(pp);
-
-  free(cpu);
-  free(art);
-  free_cache_struct(cach);
-  free_topo_struct(topo);
-  free_freq_struct(freq);
-
-  return EXIT_SUCCESS;
 }

src/printer.c

@@ -6,112 +6,207 @@
 #include "printer.h"
 #include "ascii.h"
 #include "global.h"
+#include "cpuid.h"
+#include "uarch.h"
 
-#define COL_NONE            ""
+#define COL_NONE          ""
-#define COL_INTEL_DEFAULT_1 "\x1b[36;1m"
+#define COL_INTEL_FANCY_1 "\x1b[46;1m"
-#define COL_INTEL_DEFAULT_2 "\x1b[37;1m"
+#define COL_INTEL_FANCY_2 "\x1b[47;1m"
-#define COL_INTEL_DARK_1    "\x1b[34;1m"
+#define COL_INTEL_FANCY_3 "\x1b[36;1m"
-#define COL_INTEL_DARK_2    "\x1b[30m"
+#define COL_INTEL_FANCY_4 "\x1b[37;1m"
-#define COL_AMD_DEFAULT_1   "\x1b[37;1m"
+#define COL_INTEL_RETRO_1 "\x1b[36;1m"
-#define COL_AMD_DEFAULT_2   "\x1b[31;1m"
+#define COL_INTEL_RETRO_2 "\x1b[37;1m"
-#define COL_AMD_DARK_1      "\x1b[30;1m"
+#define COL_AMD_FANCY_1   "\x1b[47;1m"
-#define COL_AMD_DARK_2      "\x1b[32;1m"
+#define COL_AMD_FANCY_2   "\x1b[42;1m"
-#define RESET               "\x1b[0m"
+#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_NAME          "Name:"
-#define TITLE_FREQUENCY "Frequency:  "
+#define TITLE_FREQUENCY     "Max Frequency:"
-#define TITLE_NCORES    "N.Cores:    "
+#define TITLE_SOCKETS       "Sockets:" 
-#define TITLE_AVX       "AVX:        "
+#define TITLE_NCORES        "Cores:" 
-#define TITLE_SSE       "SSE:        "
+#define TITLE_NCORES_DUAL   "Cores (Total):"
-#define TITLE_FMA       "FMA:        "
+#define TITLE_AVX           "AVX:"
-#define TITLE_AES       "AES:        "
+#define TITLE_SSE           "SSE:"
-#define TITLE_SHA       "SHA:        "
+#define TITLE_FMA           "FMA:"
-#define TITLE_L1        "L1 Size:    "
+#define TITLE_AES           "AES:"
-#define TITLE_L2        "L2 Size:    "
+#define TITLE_SHA           "SHA:"
-#define TITLE_L3        "L3 Size:    "
+#define TITLE_L1i           "L1i Size:"
-#define TITLE_PEAK      "Peak FLOPS: "
+#define TITLE_L1d           "L1d Size:"
+#define TITLE_L2            "L2 Size:"
+#define TITLE_L3            "L3 Size:"
+#define TITLE_PEAK          "Peak Performance:"
+#define TITLE_UARCH         "Microarchitecture:"
+#define TITLE_TECHNOLOGY    "Technology:"
 
-/*** CENTER TEXT ***/
+#define MAX_ATTRIBUTE_COUNT      14
-#define LINES_SPACE_UP   4
+#define ATTRIBUTE_NAME            0
-#define LINES_SPACE_DOWN 4
+#define ATTRIBUTE_UARCH           1
+#define ATTRIBUTE_TECHNOLOGY      2
+#define ATTRIBUTE_FREQUENCY       3
+#define ATTRIBUTE_SOCKETS         4
+#define ATTRIBUTE_NCORES          5
+#define ATTRIBUTE_NCORES_DUAL     6
+#define ATTRIBUTE_AVX             7
+#define ATTRIBUTE_FMA             8
+#define ATTRIBUTE_L1i             9
+#define ATTRIBUTE_L1d            10
+#define ATTRIBUTE_L2             11
+#define ATTRIBUTE_L3             12
+#define ATTRIBUTE_PEAK           13
 
-static const char* ATTRIBUTE_FIELDS [ATTRIBUTE_COUNT] =  {  TITLE_NAME, TITLE_FREQUENCY,
+static const char* ATTRIBUTE_FIELDS [MAX_ATTRIBUTE_COUNT] = { TITLE_NAME, TITLE_UARCH, TITLE_TECHNOLOGY,
-                                                            TITLE_NCORES, TITLE_AVX, TITLE_SSE,
+                                                              TITLE_FREQUENCY, TITLE_SOCKETS,
-                                                            TITLE_FMA, TITLE_AES, TITLE_SHA,
+                                                              TITLE_NCORES, TITLE_NCORES_DUAL, 
-                                                            TITLE_L1, TITLE_L2, TITLE_L3,
+                                                              TITLE_AVX,
-                                                            TITLE_PEAK };
+                                                              TITLE_FMA, TITLE_L1i, TITLE_L1d, TITLE_L2, TITLE_L3,
+                                                              TITLE_PEAK, 
+                                                               };
 
-static const int ATTRIBUTE_LIST[ATTRIBUTE_COUNT] =  { ATTRIBUTE_NAME, ATTRIBUTE_FREQUENCY,
+static const int ATTRIBUTE_LIST[MAX_ATTRIBUTE_COUNT] =  { ATTRIBUTE_NAME, ATTRIBUTE_UARCH, ATTRIBUTE_TECHNOLOGY,
-                                                        ATTRIBUTE_NCORES, ATTRIBUTE_AVX, ATTRIBUTE_SSE,
+                                                        ATTRIBUTE_FREQUENCY, ATTRIBUTE_SOCKETS,
-                                                        ATTRIBUTE_FMA, ATTRIBUTE_AES, ATTRIBUTE_SHA,
+                                                        ATTRIBUTE_NCORES, ATTRIBUTE_NCORES_DUAL, ATTRIBUTE_AVX,
-                                                        ATTRIBUTE_L1, ATTRIBUTE_L2, ATTRIBUTE_L3,
+                                                        ATTRIBUTE_FMA, 
+                                                        ATTRIBUTE_L1i, ATTRIBUTE_L1d, ATTRIBUTE_L2, ATTRIBUTE_L3,
                                                         ATTRIBUTE_PEAK };
 
 struct ascii {
   char art[NUMBER_OF_LINES][LINE_SIZE];
-  char color1[10];
+  char color1_ascii[100];
-  char color2[10];
+  char color2_ascii[100];
-  char reset[10];
+  char color1_text[100];
-  char* atributes[ATTRIBUTE_COUNT];
+  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) {
-  int i = 0;
+  art->attributes[type] = value;  
-  while(i < ATTRIBUTE_COUNT && type != ATTRIBUTE_LIST[i])
+  art->n_attributes_set++;
-    i++;
-  if(i != ATTRIBUTE_COUNT)
-    art->atributes[i] = value;
-  else
-    printBug("Setting attribute failed because it was not found");
 }
 
-struct ascii* set_ascii(VENDOR cpuVendor, STYLE style) {
+char* rgb_to_ansi(struct color* c, bool background, bool bold) {
-  /*** Check that number of lines of ascii art matches the number
+  char* str = malloc(sizeof(char) * 100);
-  of spaces plus the number of lines filled with text ***/
+  if(background) {
-  if(LINES_SPACE_UP+LINES_SPACE_DOWN+ATTRIBUTE_COUNT != NUMBER_OF_LINES) {
+    snprintf(str, 44, "\x1b[48;2;%.3d;%.3d;%.3dm", c->R, c->G, c->B);
-    printBug("Number of lines do not match (%d vs %d)",LINES_SPACE_UP+LINES_SPACE_DOWN+ATTRIBUTE_COUNT,NUMBER_OF_LINES);
+  }
-    return NULL;
+  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);
   }
   
-  char *COL_DEFAULT_1, *COL_DEFAULT_2, *COL_DARK_1, *COL_DARK_2;
+  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_DEFAULT_1 = COL_INTEL_DEFAULT_1;
+    COL_FANCY_1 = COL_INTEL_FANCY_1;
-    COL_DEFAULT_2 = COL_INTEL_DEFAULT_2;
+    COL_FANCY_2 = COL_INTEL_FANCY_2;
-    COL_DARK_1 = COL_INTEL_DARK_1;
+    COL_FANCY_3 = COL_INTEL_FANCY_3;
-    COL_DARK_2 = COL_INTEL_DARK_2;
+    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_DEFAULT_1 = COL_AMD_DEFAULT_1;
+    COL_FANCY_1 = COL_AMD_FANCY_1;
-    COL_DEFAULT_2 = COL_AMD_DEFAULT_2;
+    COL_FANCY_2 = COL_AMD_FANCY_2;
-    COL_DARK_1 = COL_AMD_DARK_1;
+    COL_FANCY_3 = COL_AMD_FANCY_3;
-    COL_DARK_2 = COL_AMD_DARK_2;    
+    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_NONE:
+    case STYLE_LEGACY:
-        strcpy(art->color1,COL_NONE);
+      strcpy(art->color1_ascii,COL_NONE);
-        strcpy(art->color2,COL_NONE);
+      strcpy(art->color2_ascii,COL_NONE);  
-        break; 
+      strcpy(art->color1_text,COL_NONE);  
-    case STYLE_EMPTY:
+      strcpy(art->color2_text,COL_NONE);  
-      #ifdef _WIN32
+      art->reset[0] = '\0';
-        strcpy(art->color1,COL_NONE);
-        strcpy(art->color2,COL_NONE);  
-        art->reset[0] = '\0';
-        break;
-      #endif
-    case STYLE_DEFAULT:
-      strcpy(art->color1,COL_DEFAULT_1);
-      strcpy(art->color2,COL_DEFAULT_2);
       break;
-    case STYLE_DARK:
+    case STYLE_FANCY:
-      strcpy(art->color1,COL_DARK_1);
+      if(cs != NULL) {
-      strcpy(art->color2,COL_DARK_2);
+        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;    
@@ -126,62 +221,168 @@ struct ascii* set_ascii(VENDOR cpuVendor, STYLE style) {
   return art;
 }
 
-void print_ascii_intel(struct ascii* 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(int n=0;n<NUMBER_OF_LINES;n++) {
+  printf("\n");
+  for(uint32_t n=0;n<NUMBER_OF_LINES;n++) {
 
-    /*** PRINT ASCII-ART ***/
     for(int i=0;i<LINE_SIZE;i++) {
       if(flag) {
         if(art->art[n][i] == ' ') {
           flag = false;
-          printf("%c",art->art[n][i]);
+          printf("%s%c%s", art->color2_ascii, art->ascii_chars[1], art->reset);          
         }
         else
-          printf("%s%c%s", art->color1, art->art[n][i], art->reset);
+          printf("%s%c%s",  art->color1_ascii, art->ascii_chars[0], art->reset);
       }
       else {
-        if(art->art[n][i] != ' ') {
+        if(art->art[n][i] != ' ' && art->art[n][i] != '\0') {
           flag = true;
-          printf("%s%c%s", art->color2, art->art[n][i], art->reset);
+          printf("%c",' ');
         }
         else
-          printf("%c",art->art[n][i]);
+          printf("%c",' ');
       }
     }
     
-    /*** PRINT ATTRIBUTE ***/
+    if(n > space_up-1 && n < NUMBER_OF_LINES-space_down) {      
-    if(n>LINES_SPACE_UP-1 && n<NUMBER_OF_LINES-LINES_SPACE_DOWN)
+      attr_to_print = get_next_attribute(art, attr_to_print);
-      printf("%s%s%s%s%s\n",art->color1,ATTRIBUTE_FIELDS[n-LINES_SPACE_UP],art->color2,art->atributes[n-LINES_SPACE_UP],art->reset);
+      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");
   }
+  printf("\n");
 }
 
-void print_ascii_amd(struct ascii* art) {
+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(int n=0;n<NUMBER_OF_LINES;n++) {
+  printf("\n");
-    /*** PRINT ASCII-ART ***/
+  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, art->art[n][i], art->reset);
+        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, art->art[n][i], art->reset);
+        printf("%s%c%s", art->color2_ascii, art->ascii_chars[1], art->reset);
       else
         printf("%c",art->art[n][i]);
     }
 
-    /*** PRINT ATTRIBUTE ***/
+    if(n > space_up-1 && n < NUMBER_OF_LINES-space_down) {
-    if(n>LINES_SPACE_UP-1 && n<NUMBER_OF_LINES-LINES_SPACE_DOWN)
+      attr_to_print = get_next_attribute(art, attr_to_print);
-      printf("%s%s%s%s%s\n",art->color1,ATTRIBUTE_FIELDS[n-LINES_SPACE_UP],art->color2,art->atributes[n-LINES_SPACE_UP], art->reset);
+      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");
   }
+  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);
+    print_ascii_intel(art, longest_attribute);
   else
-    print_ascii_amd(art);
+    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* uarch = get_str_uarch(cpu);
+  char* manufacturing_process = get_str_process(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* fma = get_str_fma(cpu);  
+  char* l1i = get_str_l1i(topo->cach);
+  char* l1d = get_str_l1d(topo->cach);
+  char* l2 = get_str_l2(topo->cach);
+  char* l3 = get_str_l3(topo->cach);
+  char* pp = get_str_peak_performance(cpu,topo,get_freq(freq));
+
+  setAttribute(art,ATTRIBUTE_NAME,cpu_name);
+  setAttribute(art,ATTRIBUTE_UARCH,uarch);
+  setAttribute(art,ATTRIBUTE_TECHNOLOGY,manufacturing_process);
+  setAttribute(art,ATTRIBUTE_FREQUENCY,max_frequency);
+  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);  
+  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(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);
+  
+  if(cs != NULL) free_colors_struct(cs);
+  free_cache_struct(cach);
+  free_topo_struct(topo);
+  free_freq_struct(freq);   
+  free_cpuinfo_struct(cpu);    
+  
+  return true;
 }

src/printer.h

@@ -1,37 +1,22 @@
 #ifndef __PRINTER__
 #define __PRINTER__
 
-#include "standart.h"
-#include "ascii.h"
-
-#define ATTRIBUTE_COUNT    12
-#define ATTRIBUTE_NAME      0
-#define ATTRIBUTE_FREQUENCY 1
-#define ATTRIBUTE_NCORES    2
-#define ATTRIBUTE_AVX       3
-#define ATTRIBUTE_SSE       4
-#define ATTRIBUTE_FMA       5
-#define ATTRIBUTE_AES       6
-#define ATTRIBUTE_SHA       7
-#define ATTRIBUTE_L1        8
-#define ATTRIBUTE_L2        9
-#define ATTRIBUTE_L3        10
-#define ATTRIBUTE_PEAK      11
-
 typedef int STYLE;
+
+#include "args.h"
+#include "cpuid.h"
+
 #define STYLES_COUNT 3
 
-#define STYLE_EMPTY   -2
+#define STYLE_INVALID -2
-#define STYLE_INVALID -1
+#define STYLE_EMPTY   -1
-#define STYLE_DEFAULT  0
+#define STYLE_FANCY    0
-#define STYLE_DARK     1
+#define STYLE_RETRO    1
-#define STYLE_NONE     2
+#define STYLE_LEGACY   2
 
-struct ascii;
+#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"
 
-static const int STYLES_CODE_LIST [STYLES_COUNT] = {STYLE_DEFAULT, STYLE_DARK};
+bool print_cpufetch(struct cpuInfo* cpu, struct cache* cach, struct frequency* freq, struct topology* topo, STYLE s, struct colors* cs);
-struct ascii* set_ascii(VENDOR cpuVendor, STYLE style);
-void print_ascii(struct ascii* art);
-void setAttribute(struct ascii* art, int type, char* value);
 
 #endif

src/standart.c

@@ -1,736 +0,0 @@
-#include <stdio.h>
-#include <stdlib.h>
-#include <unistd.h>
-#include <string.h>
-#include <assert.h>
-#include <stdbool.h>
-
-#ifdef _WIN32
-#include <windows.h>
-#else
-#include <unistd.h>
-#include "udev.h"
-#endif
-
-#include "standart.h"
-#include "cpuid.h"
-#include "global.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 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;
-  
-  //  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;
-};
-
-struct topology {
-  int64_t total_cores;
-  uint32_t physical_cores;
-  uint32_t logical_cores;
-  uint32_t smt;  
-  uint32_t sockets;  
-  bool ht;
-};
-
-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 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;
-}
-
-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);        
-  }
-
-  return cpu;
-}
-
-struct topology* get_topology_info(struct cpuInfo* cpu) {
-  struct topology* topo = malloc(sizeof(struct topology));
-  uint32_t eax = 0;
-  uint32_t ebx = 0;
-  uint32_t ecx = 0;
-  uint32_t edx = 0;
-  int32_t type;
-  
-  if (cpu->maxLevels >= 0x00000001) {
-    eax = 0x00000001;
-    cpuid(&eax, &ebx, &ecx, &edx);
-    topo->ht = edx & (1 << 28);
-  }
-  else {
-    printWarn("Can't read HT information from cpuid (needed level is 0x%.8X, max is 0x%.8X). Assuming HT is disabled", 0x00000001, cpu->maxLevels); 
-    topo->ht = false;
-  }
-    
-  switch(cpu->cpu_vendor) {
-    case VENDOR_INTEL:  
-      if (cpu->maxLevels >= 0x0000000B) {
-        //TODO: This idea only works with no NUMA systems  
-        eax = 0x0000000B;
-        ecx = 0x00000000;
-        cpuid(&eax, &ebx, &ecx, &edx);
-        type = (ecx >> 8) & 0xFF;
-        if (type != 1) {
-          printBug("Unexpected type in cpuid 0x0000000B (expected 1, got %d)", type);     
-          return NULL;
-        }        
-        topo->smt = ebx & 0xFFFF;                
-  
-        eax = 0x0000000B;
-        ecx = 0x00000001;
-        cpuid(&eax, &ebx, &ecx, &edx); 
-        type = (ecx >> 8) & 0xFF;
-        if (type < 2) {       
-          printBug("Unexpected type in cpuid 0x0000000B (expected < 2, got %d)", type);     
-          return NULL;
-        }
-        topo->logical_cores = ebx & 0xFFFF;
-        topo->physical_cores = topo->logical_cores / topo->smt;
-      }
-      else {
-        printWarn("Can't read topology information from cpuid (needed level is 0x%.8X, max is 0x%.8X)", 0x0000000B, cpu->maxLevels); 
-        topo->physical_cores = 1;
-        topo->logical_cores = 1;
-        topo->smt = 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 = ((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->maxLevels); 
-          topo->smt = 1;    
-        }
-        topo->physical_cores = topo->logical_cores / topo->smt;
-      }
-      else {
-        printWarn("Can't read topology information from cpuid (needed extended level is 0x%.8X, max is 0x%.8X)", 0x80000008, cpu->maxLevels); 
-        topo->physical_cores = 1;
-        topo->logical_cores = 1;
-        topo->smt = 1;    
-      }
-      break;
-    default:
-      printBug("Cant get topology because VENDOR is empty");
-      return NULL;
-  }
-  
-  // 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!
-  #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  
-  topo->sockets = topo->total_cores / topo->smt / topo->physical_cores; // Idea borrowed from lscpu
-  
-  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;
-      int32_t cache_is_self_initializing = (eax >>= 3) & 0x1; // does not need SW initialization
-      int32_t cache_is_fully_associative = (eax >>= 1) & 0x1;
-      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\n", cach->L1i/1024);
-    return NULL;
-  }
-  if(cach->L1d > 64 * 1024) {
-    printBug("Invalid L1d size: %dKB\n", cach->L1d/1024);
-    return NULL;
-  }
-  if(cach->L2 != UNKNOWN && cach->L2 > 2 * 1048576) {
-    printBug("Invalid L2 size: %dMB\n", cach->L2/(1048576));
-    return NULL;
-  }
-  if(cach->L3 != UNKNOWN && cach->L3 > 100 * 1048576) {
-    printBug("Invalid L3 size: %dMB\n", cach->L3/(1048576));
-    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;
-}
-
-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;
-}
-
-char* get_str_topology(struct topology* topo) {
-  char* string;
-  if(topo->smt > 1) {
-    //3 for digits, 8 for ' cores (', 3 for digits, 9 for ' threads)'
-    uint32_t size = 3+8+3+9+1;
-    string = malloc(sizeof(char)*size);
-    snprintf(string, size, "%d cores (%d threads)",topo->physical_cores,topo->logical_cores);
-  }
-  else {
-    uint32_t size = 3+7+1;
-    string = malloc(sizeof(char)*size);
-    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)*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;
-}
-
-// String functions
-char* get_str_l1(struct cache* cach) {
-  // 2*2 for digits, 4 for two 'KB' and 6 for '(D)' and '(I)'
-  uint32_t size = (2*2+4+6+1);
-  int32_t sanity_ret;
-  char* string = malloc(sizeof(char)*size);
-  sanity_ret = snprintf(string,size,"%d"STRING_KILOBYTES"(D)%d"STRING_KILOBYTES"(I)",cach->L1d/1024,cach->L1i/1024);
-  assert(sanity_ret > 0);
-  return string;
-}
-
-char* get_str_l2(struct cache* cach) {
-  if(cach->L2 == UNKNOWN) {
-    char* string = malloc(sizeof(char) * 5);
-    snprintf(string, 5, STRING_NONE);
-    return string;
-  }
-  else {
-    int32_t sanity_ret;  
-    char* string;
-    if(cach->L2/1024 >= 1024) {
-      //1 for digit, 2 for 'MB'
-      uint32_t size = (1+2+1);    
-      string = malloc(sizeof(char)*size);
-      sanity_ret = snprintf(string,size,"%d"STRING_MEGABYTES,cach->L2/(1048576));    
-    }
-    else {
-      //4 for digits, 2 for 'KB'
-      uint32_t size = (4+2+1);    
-      string = malloc(sizeof(char)*size);
-      sanity_ret = snprintf(string,size,"%d"STRING_KILOBYTES,cach->L2/1024);  
-    }    
-    assert(sanity_ret > 0);    
-    return string;
-  }
-}
-
-char* get_str_l3(struct cache* cach) {
-  if(cach->L3 == UNKNOWN) {
-    char* string = malloc(sizeof(char) * 5);
-    snprintf(string, 5, STRING_NONE);
-    return string;
-  }
-  else {
-    int32_t sanity_ret;  
-    char* string;
-    if(cach->L3/1024 >= 1024) {
-      //1 for digit, 2 for 'MB'
-      uint32_t size = (1+2+1);    
-      string = malloc(sizeof(char)*size);
-      sanity_ret = snprintf(string,size,"%d"STRING_MEGABYTES,cach->L3/(1048576));    
-    }
-    else {
-      //4 for digits, 2 for 'KB'
-      uint32_t size = (4+2+1);    
-      string = malloc(sizeof(char)*size);
-      sanity_ret = snprintf(string,size,"%d"STRING_KILOBYTES,cach->L3/1024);  
-    }    
-    assert(sanity_ret > 0);    
-    return string;
-  }
-}
-
-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);
-}

src/standart.h

@@ -1,55 +0,0 @@
-#ifndef __01h__
-#define __01h__
-
-#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;
-
-typedef int32_t VENDOR;
-
-struct cpuInfo* get_cpu_info();
-VENDOR get_cpu_vendor(struct cpuInfo* cpu);
-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_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_l1(struct cache* cach);
-char* get_str_l2(struct cache* cach);
-char* get_str_l3(struct cache* cach);
-
-char* get_str_freq(struct frequency* freq);
-
-char* get_str_topology(struct topology* topo);
-
-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

src/uarch.c

@@ -0,0 +1,397 @@
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "uarch.h"
+#include "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/
+ */
+
+// 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
+
+#define UARCH_UNKNOWN         0x000
+#define UARCH_P5              0x001
+#define UARCH_P6              0x002
+#define UARCH_DOTHAN          0x003
+#define UARCH_YONAH           0x004
+#define UARCH_MEROM           0x005
+#define UARCH_PENYR           0x006
+#define UARCH_NEHALEM         0x007
+#define UARCH_WESTMERE        0x008
+#define UARCH_BONNELL         0x009
+#define UARCH_SALTWELL        0x010
+#define UARCH_SANDY_BRIDGE    0x011
+#define UARCH_SILVERMONT      0x012
+#define UARCH_IVY_BRIDGE      0x013
+#define UARCH_HASWELL         0x014
+#define UARCH_BROADWELL       0x015
+#define UARCH_AIRMONT         0x016
+#define UARCH_KABY_LAKE       0x017
+#define UARCH_SKYLAKE         0x018
+#define UARCH_CASCADE_LAKE    0x019
+#define UARCH_COOPER_LAKE     0x020
+#define UARCH_KNIGHTS_LANDING 0x021
+#define UARCH_KNIGHTS_MILL    0x022
+#define UARCH_GOLDMONT        0x023
+#define UARCH_PALM_COVE       0x024
+#define UARCH_SUNNY_COVE      0x025
+#define UARCH_GOLDMONT_PLUS   0x026
+#define UARCH_TREMONT         0x027
+#define UARCH_WILLOW_COVE     0x028
+#define UARCH_COFFE_LAKE      0x029
+#define UARCH_ITANIUM         0x030
+#define UARCH_KNIGHTS_FERRY   0x031
+#define UARCH_KNIGHTS_CORNER  0x032
+#define UARCH_WILLAMETTE      0x033
+#define UARCH_NORTHWOOD       0x034
+#define UARCH_PRESCOTT        0x035
+#define UARCH_CEDAR_MILL      0x036
+#define UARCH_ITANIUM2        0x037
+#define UARCH_ICE_LAKE        0x038
+
+#define UARCH_AM486           0x038
+#define UARCH_AM5X86          0x039
+#define UARCH_K6              0x040
+#define UARCH_K7              0x041
+#define UARCH_K8              0x042
+#define UARCH_K10             0x043
+#define UARCH_PUMA_2008       0x044
+#define UARCH_BOBCAT          0x045
+#define UARCH_BULLDOZER       0x046
+#define UARCH_PILEDRIVER      0x047
+#define UARCH_STEAMROLLER     0x048
+#define UARCH_EXCAVATOR       0x049
+#define UARCH_JAGUAR          0x050
+#define UARCH_PUMA_2014       0x051
+#define UARCH_ZEN             0x052
+#define UARCH_ZEN_PLUS        0x053
+#define UARCH_ZEN2            0x054
+#define UARCH_ZEN3            0x055
+
+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;
+}
+
+// iNApired 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, NA, "Kaby Lake",       UARCH_KABY_LAKE,        14)
+  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, "Kaby Lake",       UARCH_KABY_LAKE,        14) // LX*
+  CHECK_UARCH(arch, 0,  6, 10,  6, NA, "Kaby Lake",       UARCH_KABY_LAKE,        14) // no spec update; only iNAtlatx64 example
+  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 == 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;    
+}
+
+int get_number_of_vpus(struct cpuInfo* cpu) {
+  if(cpu->cpu_vendor == VENDOR_AMD)
+    return 1;
+  
+  switch(cpu->arch->uarch) {
+      case UARCH_HASWELL:
+      case UARCH_BROADWELL:
+          
+      case UARCH_SKYLAKE:
+      case UARCH_CASCADE_LAKE:                    
+      case UARCH_KABY_LAKE:
+      case UARCH_COFFE_LAKE:
+      case UARCH_PALM_COVE:    
+      
+      case UARCH_KNIGHTS_LANDING:
+      case UARCH_KNIGHTS_MILL:
+          
+      case UARCH_ICE_LAKE:      
+        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);
+}

src/uarch.h

@@ -0,0 +1,17 @@
+#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);
+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

src/udev.c

@@ -6,7 +6,7 @@
 #include <errno.h>
 
 #include "global.h"
-#include "standart.h"
+#include "cpuid.h"
 
 #define _PATH_SYS_SYSTEM    "/sys/devices/system"
 #define _PATH_SYS_CPU       _PATH_SYS_SYSTEM"/cpu"
@@ -24,7 +24,7 @@ long get_freq_from_file(char* path) {
   if(fd == -1) {
     perror("open");
     printBug("Could not open '%s'", path);
-    return UNKNOWN;
+    return UNKNOWN_FREQ;
   }
 
   //File exists, read it
@@ -45,7 +45,7 @@ long get_freq_from_file(char* path) {
     perror("strtol");
     printBug("Failed parsing '%s' file. Read data was: '%s'", path, buf);
     free(buf);
-    return UNKNOWN;
+    return UNKNOWN_FREQ;
   }
   
   // We will be getting the frequency in KHz
@@ -53,7 +53,7 @@ long get_freq_from_file(char* path) {
   // 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;
+    return UNKNOWN_FREQ;
   }
   
   free(buf);