libcpuid(3) Library Functions Manual libcpuid(3) NAME libcpuid - LibCPUID - LibCPUID provides CPU identification. SYNOPSIS Data Structures struct cpu_raw_data_t Contains just the raw CPUID data. struct cpu_raw_data_array_t Contains an array of raw CPUID data. struct cpu_sgx_t This contains information about SGX features of the processor Example usage: struct cpu_id_t This contains the recognized CPU features/info. struct system_id_t This contains the recognized features/info for all CPUs on the system. struct cpu_mark_t Internal structure, used in cpu_tsc_mark, cpu_tsc_unmark and cpu_clock_by_mark. struct cpu_epc_t The return value of cpuid_get_epc(). struct cpu_list_t a structure that holds a list of processor names Macros #define NUM_CPU_VENDORS NUM_CPU_VENDORS #define NUM_CPU_ARCHITECTURES NUM_CPU_ARCHITECTURES #define NUM_CPU_PURPOSES NUM_CPU_PURPOSES #define NUM_HYPERVISOR_VENDORS NUM_HYPERVISOR_VENDORS #define CPU_INVALID_VALUE 0x3fffffff Typedefs typedef void(* libcpuid_warn_fn_t) (const char *msg) Enumerations enum cpu_vendor_t { VENDOR_INTEL = 0, VENDOR_AMD, VENDOR_CYRIX, VENDOR_NEXGEN, VENDOR_TRANSMETA, VENDOR_UMC, VENDOR_CENTAUR, VENDOR_RISE, VENDOR_SIS, VENDOR_NSC, VENDOR_HYGON, NUM_CPU_VENDORS, VENDOR_UNKNOWN = -1 } CPU vendor, as guessed from the Vendor String. enum cpu_architecture_t { ARCHITECTURE_X86 = 0, ARCHITECTURE_ARM, NUM_CPU_ARCHITECTURES, ARCHITECTURE_UNKNOWN = -1 } CPU architecture. enum cpu_purpose_t { PURPOSE_GENERAL = 0, PURPOSE_PERFORMANCE, PURPOSE_EFFICIENCY, NUM_CPU_PURPOSES } CPU purpose. enum hypervisor_vendor_t { HYPERVISOR_NONE = 0, HYPERVISOR_BHYVE, HYPERVISOR_HYPERV, HYPERVISOR_KVM, HYPERVISOR_PARALLELS, HYPERVISOR_QEMU, HYPERVISOR_VIRTUALBOX, HYPERVISOR_VMWARE, HYPERVISOR_XEN, NUM_HYPERVISOR_VENDORS, HYPERVISOR_UNKNOWN = -1 } Hypervisor vendor, as guessed from the CPU_FEATURE_HYPERVISOR flag. enum cpu_feature_t { CPU_FEATURE_FPU = 0, CPU_FEATURE_VME, CPU_FEATURE_DE, CPU_FEATURE_PSE, CPU_FEATURE_TSC, CPU_FEATURE_MSR, CPU_FEATURE_PAE, CPU_FEATURE_MCE, CPU_FEATURE_CX8, CPU_FEATURE_APIC, CPU_FEATURE_MTRR, CPU_FEATURE_SEP, CPU_FEATURE_PGE, CPU_FEATURE_MCA, CPU_FEATURE_CMOV, CPU_FEATURE_PAT, CPU_FEATURE_PSE36, CPU_FEATURE_PN, CPU_FEATURE_CLFLUSH, CPU_FEATURE_DTS, CPU_FEATURE_ACPI, CPU_FEATURE_MMX, CPU_FEATURE_FXSR, CPU_FEATURE_SSE, CPU_FEATURE_SSE2, CPU_FEATURE_SS, CPU_FEATURE_HT, CPU_FEATURE_TM, CPU_FEATURE_IA64, CPU_FEATURE_PBE, CPU_FEATURE_PNI, CPU_FEATURE_PCLMUL, CPU_FEATURE_DTS64, CPU_FEATURE_MONITOR, CPU_FEATURE_DS_CPL, CPU_FEATURE_VMX, CPU_FEATURE_SMX, CPU_FEATURE_EST, CPU_FEATURE_TM2, CPU_FEATURE_SSSE3, CPU_FEATURE_CID, CPU_FEATURE_CX16, CPU_FEATURE_XTPR, CPU_FEATURE_PDCM, CPU_FEATURE_DCA, CPU_FEATURE_SSE4_1, CPU_FEATURE_SSE4_2, CPU_FEATURE_SYSCALL, CPU_FEATURE_XD, CPU_FEATURE_MOVBE, CPU_FEATURE_POPCNT, CPU_FEATURE_AES, CPU_FEATURE_XSAVE, CPU_FEATURE_OSXSAVE, CPU_FEATURE_AVX, CPU_FEATURE_MMXEXT, CPU_FEATURE_3DNOW, CPU_FEATURE_3DNOWEXT, CPU_FEATURE_NX, CPU_FEATURE_FXSR_OPT, CPU_FEATURE_RDTSCP, CPU_FEATURE_LM, CPU_FEATURE_LAHF_LM, CPU_FEATURE_CMP_LEGACY, CPU_FEATURE_SVM, CPU_FEATURE_ABM, CPU_FEATURE_MISALIGNSSE, CPU_FEATURE_SSE4A, CPU_FEATURE_3DNOWPREFETCH, CPU_FEATURE_OSVW, CPU_FEATURE_IBS, CPU_FEATURE_SSE5, CPU_FEATURE_SKINIT, CPU_FEATURE_WDT, CPU_FEATURE_TS, CPU_FEATURE_FID, CPU_FEATURE_VID, CPU_FEATURE_TTP, CPU_FEATURE_TM_AMD, CPU_FEATURE_STC, CPU_FEATURE_100MHZSTEPS, CPU_FEATURE_HWPSTATE, CPU_FEATURE_CONSTANT_TSC, CPU_FEATURE_XOP, CPU_FEATURE_FMA3, CPU_FEATURE_FMA4, CPU_FEATURE_TBM, CPU_FEATURE_F16C, CPU_FEATURE_RDRAND, CPU_FEATURE_X2APIC, CPU_FEATURE_CPB, CPU_FEATURE_APERFMPERF, CPU_FEATURE_PFI, CPU_FEATURE_PA, CPU_FEATURE_AVX2, CPU_FEATURE_BMI1, CPU_FEATURE_BMI2, CPU_FEATURE_HLE, CPU_FEATURE_RTM, CPU_FEATURE_AVX512F, CPU_FEATURE_AVX512DQ, CPU_FEATURE_AVX512PF, CPU_FEATURE_AVX512ER, CPU_FEATURE_AVX512CD, CPU_FEATURE_SHA_NI, CPU_FEATURE_AVX512BW, CPU_FEATURE_AVX512VL, CPU_FEATURE_SGX, CPU_FEATURE_RDSEED, CPU_FEATURE_ADX, CPU_FEATURE_AVX512VNNI, CPU_FEATURE_AVX512VBMI, CPU_FEATURE_AVX512VBMI2, CPU_FEATURE_HYPERVISOR, NUM_CPU_FEATURES } CPU feature identifiers. enum cpu_hint_t { CPU_HINT_SSE_SIZE_AUTH = 0, NUM_CPU_HINTS } CPU detection hints identifiers. enum cpu_sgx_feature_t { INTEL_SGX1, INTEL_SGX2, NUM_SGX_FEATURES } SGX features flags. enum cpu_error_t { ERR_OK = 0, ERR_NO_CPUID = -1, ERR_NO_RDTSC = -2, ERR_NO_MEM = -3, ERR_OPEN = -4, ERR_BADFMT = -5, ERR_NOT_IMP = -6, ERR_CPU_UNKN = -7, ERR_NO_RDMSR = -8, ERR_NO_DRIVER = -9, ERR_NO_PERMS = -10, ERR_EXTRACT = -11, ERR_HANDLE = -12, ERR_INVMSR = -13, ERR_INVCNB = -14, ERR_HANDLE_R = -15, ERR_INVRANGE = -16, ERR_NOT_FOUND = -17 } Describes common library error codes. enum cpu_msrinfo_request_t { INFO_MPERF, INFO_APERF, INFO_MIN_MULTIPLIER, INFO_CUR_MULTIPLIER, INFO_MAX_MULTIPLIER, INFO_TEMPERATURE, INFO_THROTTLING, INFO_VOLTAGE, INFO_BCLK, INFO_BUS_CLOCK } Functions int cpuid_get_total_cpus (void) Returns the total number of logical CPU threads (even if CPUID is not present). int cpuid_present (void) Checks if the CPUID instruction is supported. void cpu_exec_cpuid (uint32_t eax, uint32_t *regs) Executes the CPUID instruction. void cpu_exec_cpuid_ext (uint32_t *regs) Executes the CPUID instruction with the given input registers. int cpuid_get_raw_data (struct cpu_raw_data_t *data) Obtains the raw CPUID data from the current CPU. int cpuid_get_all_raw_data (struct cpu_raw_data_array_t *data) Obtains the raw CPUID data from all CPUs. int cpuid_serialize_raw_data (struct cpu_raw_data_t *data, const char *filename) Writes the raw CPUID data to a text file. int cpuid_serialize_all_raw_data (struct cpu_raw_data_array_t *data, const char *filename) Writes all the raw CPUID data to a text file. int cpuid_deserialize_raw_data (struct cpu_raw_data_t *data, const char *filename) Reads raw CPUID data from file. int cpuid_deserialize_all_raw_data (struct cpu_raw_data_array_t *data, const char *filename) Reads all raw CPUID data from file. int cpu_identify (struct cpu_raw_data_t *raw, struct cpu_id_t *data) Identifies the CPU. int cpu_identify_all (struct cpu_raw_data_array_t *raw_array, struct system_id_t *system) Identifies all the CPUs. int cpu_request_core_type (cpu_purpose_t purpose, struct cpu_raw_data_array_t *raw_array, struct cpu_id_t *data) Identifies a given CPU type. const char * cpu_architecture_str (cpu_architecture_t architecture) Returns the short textual representation of a CPU architecture. const char * cpu_purpose_str (cpu_purpose_t purpose) Returns the short textual representation of a CPU purpose. char * affinity_mask_str_r (cpu_affinity_mask_t *affinity_mask, char *buffer, uint32_t buffer_len) Returns textual representation of a CPU affinity mask (thread-safe) char * affinity_mask_str (cpu_affinity_mask_t *affinity_mask) Returns textual representation of a CPU affinity mask. const char * cpu_feature_str (cpu_feature_t feature) Returns the short textual representation of a CPU flag. const char * cpuid_error (void) Returns textual description of the last error. void cpu_rdtsc (uint64_t *result) Executes RDTSC. void cpu_tsc_mark (struct cpu_mark_t *mark) Store TSC and timing info. void cpu_tsc_unmark (struct cpu_mark_t *mark) Calculate TSC and timing difference. int cpu_clock_by_mark (struct cpu_mark_t *mark) Calculates the CPU clock. int cpu_clock_by_os (void) Returns the CPU clock, as reported by the OS. int cpu_clock_measure (int millis, int quad_check) Measure the CPU clock frequency. int cpu_clock_by_ic (int millis, int runs) Measure the CPU clock frequency using instruction-counting. int cpu_clock (void) Get the CPU clock frequency (all-in-one method) struct cpu_epc_t cpuid_get_epc (int index, const struct cpu_raw_data_t *raw) Fetches information about an EPC (Enclave Page Cache) area. const char * cpuid_lib_version (void) Returns the libcpuid version. libcpuid_warn_fn_t cpuid_set_warn_function (libcpuid_warn_fn_t warn_fun) Sets the warning print function. void cpuid_set_verbosiness_level (int level) Sets the verbosiness level. cpu_vendor_t cpuid_get_vendor (void) Obtains the CPU vendor from CPUID from the current CPU. hypervisor_vendor_t cpuid_get_hypervisor (struct cpu_raw_data_t *raw, struct cpu_id_t *data) Obtains the hypervisor vendor from CPUID from the current CPU. void cpuid_get_cpu_list (cpu_vendor_t vendor, struct cpu_list_t *list) Gets a list of all known CPU names from a specific vendor. void cpuid_free_cpu_list (struct cpu_list_t *list) Frees a CPU list. void cpuid_free_raw_data_array (struct cpu_raw_data_array_t *raw_array) Frees a raw array. void cpuid_free_system_id (struct system_id_t *system) Frees a system ID type. struct msr_driver_t * cpu_msr_driver_open (void) Starts/opens a driver, needed to read MSRs (Model Specific Registers) struct msr_driver_t * cpu_msr_driver_open_core (unsigned core_num) Similar to cpu_msr_driver_open, but accept one parameter. int cpu_rdmsr (struct msr_driver_t *handle, uint32_t msr_index, uint64_t *result) Reads a Model-Specific Register (MSR) int cpu_rdmsr_range (struct msr_driver_t *handle, uint32_t msr_index, uint8_t highbit, uint8_t lowbit, uint64_t *result) Similar to cpu_rdmsr, but extract a range of bits. int cpu_msrinfo (struct msr_driver_t *handle, cpu_msrinfo_request_t which) Reads extended CPU information from Model-Specific Registers. int msr_serialize_raw_data (struct msr_driver_t *handle, const char *filename) Writes the raw MSR data to a text file. int cpu_msr_driver_close (struct msr_driver_t *handle) Closes an open MSR driver. Detailed Description LibCPUID provides CPU identification. Enumeration Type Documentation enum cpu_architecture_t CPU architecture. Enumerator ARCHITECTURE_X86 x86 CPU ARCHITECTURE_ARM ARM CPU NUM_CPU_ARCHITECTURES Valid CPU architecture ids: 0..NUM_CPU_ARCHITECTURES - 1 enum cpu_error_t Describes common library error codes. Enumerator ERR_OK No error ERR_NO_CPUID CPUID instruction is not supported ERR_NO_RDTSC RDTSC instruction is not supported ERR_NO_MEM Memory allocation failed ERR_OPEN File open operation failed ERR_BADFMT Bad file format ERR_NOT_IMP Not implemented ERR_CPU_UNKN Unsupported processor ERR_NO_RDMSR RDMSR instruction is not supported ERR_NO_DRIVER RDMSR driver error (generic) ERR_NO_PERMS No permissions to install RDMSR driver ERR_EXTRACT Cannot extract RDMSR driver (read only media?) ERR_HANDLE Bad handle ERR_INVMSR Invalid MSR ERR_INVCNB Invalid core number ERR_HANDLE_R Error on handle read ERR_INVRANGE Invalid given range ERR_NOT_FOUND Requested type not found enum cpu_feature_t CPU feature identifiers. Usage: ... struct cpu_raw_data_t raw; struct cpu_id_t id; if (cpuid_get_raw_data(&raw) == 0 && cpu_identify(&raw, &id) == 0) { if (id.flags[CPU_FEATURE_SSE2]) { // The CPU has SSE2... ... } else { // no SSE2 } } else { // processor cannot be determined. } Enumerator CPU_FEATURE_FPU Floating point unit CPU_FEATURE_VME Virtual mode extension CPU_FEATURE_DE Debugging extension CPU_FEATURE_PSE Page size extension CPU_FEATURE_TSC Time-stamp counter CPU_FEATURE_MSR Model-specific regsisters, RDMSR/WRMSR supported CPU_FEATURE_PAE Physical address extension CPU_FEATURE_MCE Machine check exception CPU_FEATURE_CX8 CMPXCHG8B instruction supported CPU_FEATURE_APIC APIC support CPU_FEATURE_MTRR Memory type range registers CPU_FEATURE_SEP SYSENTER / SYSEXIT instructions supported CPU_FEATURE_PGE Page global enable CPU_FEATURE_MCA Machine check architecture CPU_FEATURE_CMOV CMOVxx instructions supported CPU_FEATURE_PAT Page attribute table CPU_FEATURE_PSE36 36-bit page address extension CPU_FEATURE_PN Processor serial # implemented (Intel P3 only) CPU_FEATURE_CLFLUSH CLFLUSH instruction supported CPU_FEATURE_DTS Debug store supported CPU_FEATURE_ACPI ACPI support (power states) CPU_FEATURE_MMX MMX instruction set supported CPU_FEATURE_FXSR FXSAVE / FXRSTOR supported CPU_FEATURE_SSE Streaming-SIMD Extensions (SSE) supported CPU_FEATURE_SSE2 SSE2 instructions supported CPU_FEATURE_SS Self-snoop CPU_FEATURE_HT Hyper-threading supported (but might be disabled) CPU_FEATURE_TM Thermal monitor CPU_FEATURE_IA64 IA64 supported (Itanium only) CPU_FEATURE_PBE Pending-break enable CPU_FEATURE_PNI PNI (SSE3) instructions supported CPU_FEATURE_PCLMUL PCLMULQDQ instruction supported CPU_FEATURE_DTS64 64-bit Debug store supported CPU_FEATURE_MONITOR MONITOR / MWAIT supported CPU_FEATURE_DS_CPL CPL Qualified Debug Store CPU_FEATURE_VMX Virtualization technology supported CPU_FEATURE_SMX Safer mode exceptions CPU_FEATURE_EST Enhanced SpeedStep CPU_FEATURE_TM2 Thermal monitor 2 CPU_FEATURE_SSSE3 SSSE3 instructionss supported (this is different from SSE3!) CPU_FEATURE_CID Context ID supported CPU_FEATURE_CX16 CMPXCHG16B instruction supported CPU_FEATURE_XTPR Send Task Priority Messages disable CPU_FEATURE_PDCM Performance capabilities MSR supported CPU_FEATURE_DCA Direct cache access supported CPU_FEATURE_SSE4_1 SSE 4.1 instructions supported CPU_FEATURE_SSE4_2 SSE 4.2 instructions supported CPU_FEATURE_SYSCALL SYSCALL / SYSRET instructions supported CPU_FEATURE_XD Execute disable bit supported CPU_FEATURE_MOVBE MOVBE instruction supported CPU_FEATURE_POPCNT POPCNT instruction supported CPU_FEATURE_AES AES* instructions supported CPU_FEATURE_XSAVE XSAVE/XRSTOR/etc instructions supported CPU_FEATURE_OSXSAVE non-privileged copy of OSXSAVE supported CPU_FEATURE_AVX Advanced vector extensions supported CPU_FEATURE_MMXEXT AMD MMX-extended instructions supported CPU_FEATURE_3DNOW AMD 3DNow! instructions supported CPU_FEATURE_3DNOWEXT AMD 3DNow! extended instructions supported CPU_FEATURE_NX No-execute bit supported CPU_FEATURE_FXSR_OPT FFXSR: FXSAVE and FXRSTOR optimizations CPU_FEATURE_RDTSCP RDTSCP instruction supported (AMD-only) CPU_FEATURE_LM Long mode (x86_64/EM64T) supported CPU_FEATURE_LAHF_LM LAHF/SAHF supported in 64-bit mode CPU_FEATURE_CMP_LEGACY core multi-processing legacy mode CPU_FEATURE_SVM AMD Secure virtual machine CPU_FEATURE_ABM LZCNT instruction support CPU_FEATURE_MISALIGNSSE Misaligned SSE supported CPU_FEATURE_SSE4A SSE 4a from AMD CPU_FEATURE_3DNOWPREFETCH PREFETCH/PREFETCHW support CPU_FEATURE_OSVW OS Visible Workaround (AMD) CPU_FEATURE_IBS Instruction-based sampling CPU_FEATURE_SSE5 SSE 5 instructions supported (deprecated, will never be 1) CPU_FEATURE_SKINIT SKINIT / STGI supported CPU_FEATURE_WDT Watchdog timer support CPU_FEATURE_TS Temperature sensor CPU_FEATURE_FID Frequency ID control CPU_FEATURE_VID Voltage ID control CPU_FEATURE_TTP THERMTRIP CPU_FEATURE_TM_AMD AMD-specified hardware thermal control CPU_FEATURE_STC Software thermal control CPU_FEATURE_100MHZSTEPS 100 MHz multiplier control CPU_FEATURE_HWPSTATE Hardware P-state control CPU_FEATURE_CONSTANT_TSC TSC ticks at constant rate CPU_FEATURE_XOP The XOP instruction set (same as the old CPU_FEATURE_SSE5) CPU_FEATURE_FMA3 The FMA3 instruction set CPU_FEATURE_FMA4 The FMA4 instruction set CPU_FEATURE_TBM Trailing bit manipulation instruction support CPU_FEATURE_F16C 16-bit FP convert instruction support CPU_FEATURE_RDRAND RdRand instruction CPU_FEATURE_X2APIC x2APIC, APIC_BASE.EXTD, MSRs 0000_0800h...0000_0BFFh 64-bit ICR (+030h but not +031h), no DFR (+00Eh), SELF_IPI (+040h) also see standard level 0000_000Bh CPU_FEATURE_CPB Core performance boost CPU_FEATURE_APERFMPERF MPERF/APERF MSRs support CPU_FEATURE_PFI Processor Feedback Interface support CPU_FEATURE_PA Processor accumulator CPU_FEATURE_AVX2 AVX2 instructions CPU_FEATURE_BMI1 BMI1 instructions CPU_FEATURE_BMI2 BMI2 instructions CPU_FEATURE_HLE Hardware Lock Elision prefixes CPU_FEATURE_RTM Restricted Transactional Memory instructions CPU_FEATURE_AVX512F AVX-512 Foundation CPU_FEATURE_AVX512DQ AVX-512 Double/Quad granular insns CPU_FEATURE_AVX512PF AVX-512 Prefetch CPU_FEATURE_AVX512ER AVX-512 Exponential/Reciprocal CPU_FEATURE_AVX512CD AVX-512 Conflict detection CPU_FEATURE_SHA_NI SHA-1/SHA-256 instructions CPU_FEATURE_AVX512BW AVX-512 Byte/Word granular insns CPU_FEATURE_AVX512VL AVX-512 128/256 vector length extensions CPU_FEATURE_SGX SGX extensions. Non-autoritative, check cpu_id_t::sgx::present to verify presence CPU_FEATURE_RDSEED RDSEED instruction CPU_FEATURE_ADX ADX extensions (arbitrary precision) CPU_FEATURE_AVX512VNNI AVX-512 Vector Neural Network Instructions CPU_FEATURE_AVX512VBMI AVX-512 Vector Bit ManipulationInstructions (version 1) CPU_FEATURE_AVX512VBMI2 AVX-512 Vector Bit ManipulationInstructions (version 2) CPU_FEATURE_HYPERVISOR Hypervisor present (always zero on physical CPUs) enum cpu_hint_t CPU detection hints identifiers. Usage: similar to the flags usage Enumerator CPU_HINT_SSE_SIZE_AUTH SSE unit size is authoritative (not only a Family/Model guesswork, but based on an actual CPUID bit) enum cpu_msrinfo_request_t Enumerator INFO_MPERF Maximum performance frequency clock. This is a counter, which increments as a proportion of the actual processor speed. INFO_APERF Actual performance frequency clock. This accumulates the core clock counts when the core is active. INFO_MIN_MULTIPLIER Minimum CPU:FSB ratio for this CPU, multiplied by 100. INFO_CUR_MULTIPLIER Current CPU:FSB ratio, multiplied by 100. e.g., a CPU:FSB value of 18.5 reads as '1850'. INFO_MAX_MULTIPLIER Maximum CPU:FSB ratio for this CPU, multiplied by 100. INFO_TEMPERATURE The current core temperature in Celsius. INFO_THROTTLING 1 if the current logical processor is throttling. 0 if it is running normally. INFO_VOLTAGE The current core voltage in Volt, multiplied by 100. INFO_BCLK See INFO_BUS_CLOCK. INFO_BUS_CLOCK The main bus clock in MHz, e.g., FSB/QPI/DMI/HT base clock, multiplied by 100. enum cpu_purpose_t CPU purpose. Enumerator PURPOSE_GENERAL general purpose CPU PURPOSE_PERFORMANCE performance CPU PURPOSE_EFFICIENCY efficiency CPU NUM_CPU_PURPOSES Valid CPU purpose ids: 0..NUM_CPU_PURPOSES - 1 enum cpu_sgx_feature_t SGX features flags. See also cpu_sgx_t Usage: ... struct cpu_raw_data_t raw; struct cpu_id_t id; if (cpuid_get_raw_data(&raw) == 0 && cpu_identify(&raw, &id) == 0 && id.sgx.present) { if (id.sgx.flags[INTEL_SGX1]) // The CPU has SGX1 instructions support... ... } else { // no SGX } } else { // processor cannot be determined. } Enumerator INTEL_SGX1 SGX1 instructions support INTEL_SGX2 SGX2 instructions support enum cpu_vendor_t CPU vendor, as guessed from the Vendor String. Enumerator VENDOR_INTEL Intel CPU VENDOR_AMD AMD CPU VENDOR_CYRIX Cyrix CPU VENDOR_NEXGEN NexGen CPU VENDOR_TRANSMETA Transmeta CPU VENDOR_UMC x86 CPU by UMC VENDOR_CENTAUR x86 CPU by IDT VENDOR_RISE x86 CPU by Rise Technology VENDOR_SIS x86 CPU by SiS VENDOR_NSC x86 CPU by National Semiconductor VENDOR_HYGON Hygon CPU NUM_CPU_VENDORS Valid CPU vendor ids: 0..NUM_CPU_VENDORS - 1 enum hypervisor_vendor_t Hypervisor vendor, as guessed from the CPU_FEATURE_HYPERVISOR flag. Enumerator HYPERVISOR_NONE no hypervisor HYPERVISOR_BHYVE FreeBSD bhyve hypervisor HYPERVISOR_HYPERV Microsoft Hyper-V or Windows Virtual PC hypervisor HYPERVISOR_KVM KVM hypervisor HYPERVISOR_PARALLELS Parallels hypervisor HYPERVISOR_QEMU QEMU hypervisor HYPERVISOR_VIRTUALBOX VirtualBox hypervisor HYPERVISOR_VMWARE VMware hypervisor HYPERVISOR_XEN Xen hypervisor NUM_HYPERVISOR_VENDORS Valid hypervisor vendor ids: 0..NUM_HYPERVISOR_VENDORS - 1 Function Documentation char * affinity_mask_str (cpu_affinity_mask_t * affinity_mask) Returns textual representation of a CPU affinity mask. Parameters affinity_mask - the affinity mask (in hexadecimal), whose textual representation is wanted. Note This function is not thread-safe Returns a string like '0000FFFF', '00FF0000', etc. char * affinity_mask_str_r (cpu_affinity_mask_t * affinity_mask, char * buffer, uint32_t buffer_len) Returns textual representation of a CPU affinity mask (thread-safe) Parameters affinity_mask - Input - the affinity mask (in hexadecimal), whose textual representation is wanted. buffer - Output - an allocated string where to store the textual representation, like '0000FFFF', '00FF0000', etc. buffer_len - Input - the size of buffer. Returns a pointer on buffer const char * cpu_architecture_str (cpu_architecture_t architecture) Returns the short textual representation of a CPU architecture. Parameters architecture - the architecture, whose textual representation is wanted. Returns a constant string like 'x86', 'ARM', etc. int cpu_clock (void) Get the CPU clock frequency (all-in-one method) This is an all-in-one method for getting the CPU clock frequency. It tries to use the OS for that. If the OS doesn't have this info, it uses cpu_clock_measure with 200ms time interval and quadruple checking. Returns the CPU clock frequency in MHz. If every possible method fails, the result is -1. int cpu_clock_by_ic (int millis, int runs) Measure the CPU clock frequency using instruction-counting. Parameters millis - how much time to allocate for each run, in milliseconds runs - how many runs to perform The function performs a busy-wait cycle using a known number of 'heavy' (SSE) instructions. These instructions run at (more or less guaranteed) 1 IPC rate, so by running a busy loop for a fixed amount of time, and measuring the amount of instructions done, the CPU clock is accurately measured. Of course, this function is still affected by the power-saving schemes, so the warnings as of cpu_clock_measure() still apply. However, this function is immune to problems with detection, related to the Intel Nehalem's 'Turbo' mode, where the internal clock is raised, but the RDTSC rate is unaffected. The function will run for about (millis * runs) milliseconds. You can make only a single busy-wait run (runs == 1); however, this can be affected by task scheduling (which will break the counting), so allowing more than one run is recommended. As run length is not imperative for accurate readings (e.g., 50ms is sufficient), you can afford a lot of short runs, e.g. 10 runs of 50ms or 20 runs of 25ms. Recommended values - millis = 50, runs = 4. For more robustness, increase the number of runs. NOTE: on Bulldozer and later CPUs, the busy-wait cycle runs at 1.4 IPC, thus the results are skewed. This is corrected internally by dividing the resulting value by 1.4. However, this only occurs if the thread is executed on a single CMT module - if there are other threads competing for resources, the results are unpredictable. Make sure you run cpu_clock_by_ic() on a CPU that is free from competing threads, or if there are such threads, they shouldn't exceed the number of modules. On a Bulldozer X8, that means 4 threads. Returns the CPU clock frequency in MHz (within some measurement error margin). If SSE is not supported, the result is -1. If the input parameters are incorrect, or some other internal fault is detected, the result is -2. int cpu_clock_by_mark (struct cpu_mark_t * mark) Calculates the CPU clock. Parameters mark - pointer to a cpu_mark_t structure, which has been initialized with cpu_tsc_mark and later `stopped' with cpu_tsc_unmark. Note For reliable results, the marked time interval should be at least about 10 ms. Returns the CPU clock frequency, in MHz. Due to measurement error, it will differ from the true value in a few least-significant bits. Accuracy depends on the timing interval - the more, the better. If the timing interval is insufficient, the result is -1. Also, see the comment on cpu_clock_measure for additional issues and pitfalls in using RDTSC for CPU frequency measurements. int cpu_clock_by_os (void) Returns the CPU clock, as reported by the OS. This function uses OS- specific functions to obtain the CPU clock. It may differ from the true clock for several reasons: i) The CPU might be in some power saving state, while the OS reports its full-power frequency, or vice-versa. ii) In some cases you can raise or lower the CPU frequency with overclocking utilities and the OS will not notice. Returns the CPU clock frequency in MHz. If the OS is not (yet) supported or lacks the necessary reporting machinery, the return value is -1 int cpu_clock_measure (int millis, int quad_check) Measure the CPU clock frequency. Parameters millis - How much time to waste in the busy-wait cycle. In millisecs. Useful values 10 - 1000 quad_check - Do a more thorough measurement if nonzero (see the explanation). The function performs a busy-wait cycle for the given time and calculates the CPU frequency by the difference of the TSC values. The accuracy of the calculation depends on the length of the busy-wait cycle: more is better, but 100ms should be enough for most purposes. While this will calculate the CPU frequency correctly in most cases, there are several reasons why it might be incorrect: i) RDTSC doesn't guarantee it will run at the same clock as the CPU. Apparently there aren't CPUs at the moment, but still, there's no guarantee. ii) The CPU might be in a low-frequency power saving mode, and the CPU might be switched to higher frequency at any time. If this happens during the measurement, the result can be anywhere between the low and high frequencies. Also, if you're interested in the high frequency value only, this function might return the low one instead. iii) On SMP systems exhibiting TSC drift (see cpu_rdtsc) the quad_check option will run four consecutive measurements and then return the average of the two most-consistent results. The total runtime of the function will still be `millis' - consider using a bit more time for the timing interval. Finally, for benchmarking / CPU intensive applications, the best strategy is to use the cpu_tsc_mark() / cpu_tsc_unmark() / cpu_clock_by_mark() method. Begin by mark()-ing about one second after application startup (allowing the power-saving manager to kick in and rise the frequency during that time), then unmark() just before application finishing. The result will most acurately represent at what frequency your app was running. Returns the CPU clock frequency in MHz (within some measurement error margin). If RDTSC is not supported, the result is -1. void cpu_exec_cpuid (uint32_t eax, uint32_t * regs) Executes the CPUID instruction. Parameters eax - the value of the EAX register when executing CPUID regs - the results will be stored here. regs[0] = EAX, regs[1] = EBX, ... Note CPUID will be executed with EAX set to the given value and EBX, ECX, EDX set to zero. void cpu_exec_cpuid_ext (uint32_t * regs) Executes the CPUID instruction with the given input registers. Note This is just a bit more generic version of cpu_exec_cpuid - it allows you to control all the registers. Parameters regs - Input/output. Prior to executing CPUID, EAX, EBX, ECX and EDX will be set to regs[0], regs[1], regs[2] and regs[3]. After CPUID, this array will contain the results. const char * cpu_feature_str (cpu_feature_t feature) Returns the short textual representation of a CPU flag. Parameters feature - the feature, whose textual representation is wanted. Returns a constant string like 'fpu', 'tsc', 'sse2', etc. Note the names of the returned flags are compatible with those from /proc/cpuinfo in Linux, with the exception of `tm_amd' int cpu_identify (struct cpu_raw_data_t * raw, struct cpu_id_t * data) Identifies the CPU. Parameters raw - Input - a pointer to the raw CPUID data, which is obtained either by cpuid_get_raw_data or cpuid_deserialize_raw_data. Can also be NULL, in which case the functions calls cpuid_get_raw_data itself. data - Output - the decoded CPU features/info is written here. Note The function will not fail, even if some of the information cannot be obtained. Even when the CPU is new and thus unknown to libcpuid, some generic info, such as 'AMD K9 family CPU' will be written to data.cpu_codename, and most other things, such as the CPU flags, cache sizes, etc. should be detected correctly anyway. However, the function CAN fail, if the CPU is completely alien to libcpuid. While cpu_identify() and cpuid_get_raw_data() are fast for most purposes, running them several thousand times per second can hamper performance significantly. Specifically, avoid writing 'cpu feature checker' wrapping function, which calls cpu_identify and returns the value of some flag, if that function is going to be called frequently. Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t int cpu_identify_all (struct cpu_raw_data_array_t * raw_array, struct system_id_t * system) Identifies all the CPUs. Parameters raw_array - Input - a pointer to the array of raw CPUID data, which is obtained either by cpuid_get_all_raw_data or cpuid_deserialize_all_raw_data. Can also be NULL, in which case the functions calls cpuid_get_all_raw_data itself. system - Output - the decoded CPU features/info is written here for each CPU type. Note The function is similar to cpu_identify. Refer to cpu_identify notes. As the memory is dynamically allocated, be sure to call cpuid_free_raw_data_array() and cpuid_free_system_id() after you're done with the data Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t int cpu_msr_driver_close (struct msr_driver_t * handle) Closes an open MSR driver. This function unloads the MSR driver opened by cpu_msr_driver_open and frees any resources associated with it. Parameters handle - a handle to the MSR reader driver, as created by cpu_msr_driver_open Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t struct msr_driver_t * cpu_msr_driver_open (void) Starts/opens a driver, needed to read MSRs (Model Specific Registers) On systems that support it, this function will create a temporary system driver, that has privileges to execute the RDMSR instruction. After the driver is created, you can read MSRs by calling cpu_rdmsr Returns a handle to the driver on success, and NULL on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t struct msr_driver_t * cpu_msr_driver_open_core (unsigned core_num) Similar to cpu_msr_driver_open, but accept one parameter. This function works on certain operating systems (GNU/Linux, FreeBSD) Parameters core_num specify the core number for MSR. The first core number is 0. The last core number is cpuid_get_total_cpus - 1. Returns a handle to the driver on success, and NULL on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t int cpu_msrinfo (struct msr_driver_t * handle, cpu_msrinfo_request_t which) Reads extended CPU information from Model-Specific Registers. Parameters handle - a handle to an open MSR driver, See also cpu_msr_driver_open Parameters which - which info field should be returned. A list of available information entities is listed in the cpu_msrinfo_request_t enum. Return values - if the requested information is available for the current processor model, the respective value is returned. if no information is available, or the CPU doesn't support the query, the special value CPU_INVALID_VALUE is returned Note This function is not MT-safe. If you intend to call it from multiple threads, guard it through a mutex or a similar primitive. const char * cpu_purpose_str (cpu_purpose_t purpose) Returns the short textual representation of a CPU purpose. Parameters purpose - the purpose, whose textual representation is wanted. Returns a constant string like 'general', 'performance', 'efficiency', etc. int cpu_rdmsr (struct msr_driver_t * handle, uint32_t msr_index, uint64_t * result) Reads a Model-Specific Register (MSR) If the CPU has MSRs (as indicated by the CPU_FEATURE_MSR flag), you can read a MSR with the given index by calling this function. There are several prerequisites you must do before reading MSRs: 1) You must ensure the CPU has RDMSR. Check the CPU_FEATURE_MSR flag in cpu_id_t::flags 2) You must ensure that the CPU implements the specific MSR you intend to read. 3) You must open a MSR-reader driver. RDMSR is a privileged instruction and needs ring-0 access in order to work. This temporary driver is created by calling cpu_msr_driver_open Parameters handle - a handle to the MSR reader driver, as created by cpu_msr_driver_open msr_index - the numeric ID of the MSR you want to read result - a pointer to a 64-bit integer, where the MSR value is stored Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t int cpu_rdmsr_range (struct msr_driver_t * handle, uint32_t msr_index, uint8_t highbit, uint8_t lowbit, uint64_t * result) Similar to cpu_rdmsr, but extract a range of bits. Parameters handle - a handle to the MSR reader driver, as created by cpu_msr_driver_open msr_index - the numeric ID of the MSR you want to read highbit - the high bit in range, must be inferior to 64 lowbit - the low bit in range, must be equal or superior to 0 result - a pointer to a 64-bit integer, where the MSR value is stored Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t void cpu_rdtsc (uint64_t * result) Executes RDTSC. The RDTSC (ReaD Time Stamp Counter) instruction gives access to an internal 64-bit counter, which usually increments at each clock cycle. This can be used for various timing routines, and as a very precise clock source. It is set to zero on system startup. Beware that may not increment at the same frequency as the CPU. Consecutive calls of RDTSC are, however, guaranteed to return monotonically- increasing values. Parameters result - a pointer to a 64-bit unsigned integer, where the TSC value will be stored Note If 100% compatibility is a concern, you must first check if the RDTSC instruction is present (if it is not, your program will crash with 'invalid opcode' exception). Only some very old processors (i486, early AMD K5 and some Cyrix CPUs) lack that instruction - they should have become exceedingly rare these days. To verify RDTSC presence, run cpu_identify() and check flags[CPU_FEATURE_TSC]. The monotonically increasing nature of the TSC may be violated on SMP systems, if their TSC clocks run at different rate. If the OS doesn't account for that, the TSC drift may become arbitrary large. int cpu_request_core_type (cpu_purpose_t purpose, struct cpu_raw_data_array_t * raw_array, struct cpu_id_t * data) Identifies a given CPU type. Parameters purpose - Input - a cpu_purpose_t to request raw_array - Optional input - a pointer to the array of raw CPUID data, which is obtained either by cpuid_get_all_raw_data or cpuid_deserialize_all_raw_data. Can also be NULL, in which case the functions calls cpuid_get_all_raw_data itself. data - Output - the decoded CPU features/info is written here. Returns zero if successful, and some negative number on error (like ERR_NOT_FOUND if CPU type not found). The error message can be obtained by calling cpuid_error. See also cpu_error_t void cpu_tsc_mark (struct cpu_mark_t * mark) Store TSC and timing info. This function stores the current TSC value and current time info from a precise OS-specific clock source in the cpu_mark_t structure. The sys_clock field contains time with microsecond resolution. The values can later be used to measure time intervals, number of clocks, FPU frequency, etc. See also cpu_rdtsc Parameters mark [out] - a pointer to a cpu_mark_t structure void cpu_tsc_unmark (struct cpu_mark_t * mark) Calculate TSC and timing difference. Parameters mark - input/output: a pointer to a cpu_mark_t structure, which has already been initialized by cpu_tsc_mark. The difference in TSC and time will be written here. This function calculates the TSC and time difference, by obtaining the current TSC and timing values and subtracting the contents of the `mark' structure from them. Results are written in the same structure. Example: ... struct cpu_mark_t mark; cpu_tsc_mark(&mark); foo(); cpu_tsc_unmark(&mark); printf("Foo finished. Executed in %llu cycles and %llu usecs\n", mark.tsc, mark.sys_clock); ... int cpuid_deserialize_all_raw_data (struct cpu_raw_data_array_t * data, const char * filename) Reads all raw CPUID data from file. Parameters data - a pointer to cpu_raw_data_array_t structure. The deserialized array data will be written here. filename - the path of the file, containing the serialized raw data. If empty, stdin will be used. Note This function may fail, if the file is created by different version of the library. Also, see the notes on cpuid_serialize_all_raw_data. As the memory is dynamically allocated, be sure to call cpuid_free_raw_data_array() after you're done with the data Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t int cpuid_deserialize_raw_data (struct cpu_raw_data_t * data, const char * filename) Reads raw CPUID data from file. Parameters data - a pointer to cpu_raw_data_t structure. The deserialized data will be written here. filename - the path of the file, containing the serialized raw data. If empty, stdin will be used. Note This function may fail, if the file is created by different version of the library. Also, see the notes on cpuid_serialize_raw_data. Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t const char * cpuid_error (void) Returns textual description of the last error. libcpuid stores an `errno'-style error status, whose description can be obtained with this function. Note This function is not thread-safe See also cpu_error_t void cpuid_free_cpu_list (struct cpu_list_t * list) Frees a CPU list. This function deletes all the memory associated with a CPU list, as obtained by cpuid_get_cpu_list() Parameters list - the list to be free()'d. void cpuid_free_raw_data_array (struct cpu_raw_data_array_t * raw_array) Frees a raw array. This function deletes all the memory associated with a raw array, as obtained by cpuid_get_all_raw_data(), cpuid_deserialize_all_raw_data() and cpu_identify_all() Parameters raw_array - the raw array to be free()'d. void cpuid_free_system_id (struct system_id_t * system) Frees a system ID type. This function deletes all the memory associated with a system ID, as obtained by cpu_identify_all() Parameters system - the system ID to be free()'d. int cpuid_get_all_raw_data (struct cpu_raw_data_array_t * data) Obtains the raw CPUID data from all CPUs. Parameters data - a pointer to cpu_raw_data_array_t structure Note As the memory is dynamically allocated, be sure to call cpuid_free_raw_data_array() after you're done with the data Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t void cpuid_get_cpu_list (cpu_vendor_t vendor, struct cpu_list_t * list) Gets a list of all known CPU names from a specific vendor. This function compiles a list of all known CPU (code)names (i.e. the possible values of cpu_id_t::cpu_codename) for the given vendor. There are about 100 entries for Intel and AMD, and a few for the other vendors. The list is written out in approximate chronological introduction order of the parts. Parameters vendor the vendor to be queried list [out] the resulting list will be written here. On failure, num_entries is set to zero and names to NULL. The error message can be obtained by calling cpuid_error. See also cpu_error_t NOTE: As the memory is dynamically allocated, be sure to call cpuid_free_cpu_list() after you're done with the data cpu_list_t struct cpu_epc_t cpuid_get_epc (int index, const struct cpu_raw_data_t * raw) Fetches information about an EPC (Enclave Page Cache) area. Parameters index - zero-based index, valid range [0..cpu_id_t.egx.num_epc_sections) raw - a pointer to fetched raw CPUID data. Needed only for testing, you can safely pass NULL here (if you pass a real structure, it will be used for fetching the leaf 12h data if index < 2; otherwise the real CPUID instruction will be used). Returns the requested data. If the CPU doesn't support SGX, or if index >= cpu_id_t.egx.num_epc_sections, both fields of the returned structure will be zeros. hypervisor_vendor_t cpuid_get_hypervisor (struct cpu_raw_data_t * raw, struct cpu_id_t * data) Obtains the hypervisor vendor from CPUID from the current CPU. Parameters raw - Optional input - a pointer to the raw CPUID data, which is obtained either by cpuid_get_raw_data or cpuid_deserialize_raw_data. Can also be NULL, in which case the functions calls cpuid_get_raw_data itself. data - Optional input - the decoded CPU features/info is written here. Can also be NULL, in which case the functions calls cpu_identify itself. Note If no hypervisor is detected, the hypervisor can be hidden in some cases. Refer to https://github.com/anrieff/libcpuid/issues/90#issuecomment-296568713. Returns HYPERVISOR_UNKNOWN if failed, HYPERVISOR_NONE if no hypervisor detected (or hidden), otherwise the hypervisor vendor type. See also hypervisor_vendor_t int cpuid_get_raw_data (struct cpu_raw_data_t * data) Obtains the raw CPUID data from the current CPU. Parameters data - a pointer to cpu_raw_data_t structure Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t int cpuid_get_total_cpus (void) Returns the total number of logical CPU threads (even if CPUID is not present). Under VM, this number (and total_logical_cpus, since they are fetched with the same code) may be nonsensical, i.e. might not equal NumPhysicalCPUs*NumCoresPerCPU*HyperThreading. This is because no matter how many logical threads the host machine has, you may limit them in the VM to any number you like. This is the number returned by cpuid_get_total_cpus(). Returns Number of logical CPU threads available. Equals the cpu_id_t::total_logical_cpus. cpu_vendor_t cpuid_get_vendor (void) Obtains the CPU vendor from CPUID from the current CPU. Note The result is cached. Returns VENDOR_UNKNOWN if failed, otherwise the CPU vendor type. See also cpu_vendor_t const char * cpuid_lib_version (void) Returns the libcpuid version. Returns the string representation of the libcpuid version, like '0.1.1' int cpuid_present (void) Checks if the CPUID instruction is supported. Return values 1 if CPUID is present 0 the CPU doesn't have CPUID. int cpuid_serialize_all_raw_data (struct cpu_raw_data_array_t * data, const char * filename) Writes all the raw CPUID data to a text file. Parameters data - a pointer to cpu_raw_data_array_t structure filename - the path of the file, where the serialized data for all CPUs should be written. If empty, stdout will be used. Note This is intended primarily for debugging. On some processor, which is not currently supported or not completely recognized by cpu_identify_all, one can still successfully get the raw data and write it to a file. libcpuid developers can later import this file and debug the detection code as if running on the actual hardware. The file is simple text format of 'something=value' pairs. Version info is also written, but the format is not intended to be neither backward- nor forward compatible. Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t int cpuid_serialize_raw_data (struct cpu_raw_data_t * data, const char * filename) Writes the raw CPUID data to a text file. Parameters data - a pointer to cpu_raw_data_t structure filename - the path of the file, where the serialized data should be written. If empty, stdout will be used. Note This is intended primarily for debugging. On some processor, which is not currently supported or not completely recognized by cpu_identify, one can still successfully get the raw data and write it to a file. libcpuid developers can later import this file and debug the detection code as if running on the actual hardware. The file is simple text format of 'something=value' pairs. Version info is also written, but the format is not intended to be neither backward- nor forward compatible. Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t void cpuid_set_verbosiness_level (int level) Sets the verbosiness level. When the verbosiness level is above zero, some functions might print diagnostic information about what are they doing. The higher the level is, the more detail is printed. Level zero is guaranteed to omit all such output. The output is written using the same machinery as the warnings, See also cpuid_set_warn_function() Parameters level the desired verbosiness level. Useful values 0..2 inclusive libcpuid_warn_fn_t cpuid_set_warn_function (libcpuid_warn_fn_t warn_fun) Sets the warning print function. In some cases, the internal libcpuid machinery would like to emit useful debug warnings. By default, these warnings are written to stderr. However, you can set a custom function that will receive those warnings. Parameters warn_fun - the warning function you want to set. If NULL, warnings are disabled. The function takes const char* argument. Returns the current warning function. You can use the return value to keep the previous warning function and restore it at your discretion. int msr_serialize_raw_data (struct msr_driver_t * handle, const char * filename) Writes the raw MSR data to a text file. Parameters handle - a handle to the MSR reader driver, as created by cpu_msr_driver_open filename - the path of the file, where the serialized data should be written. If empty, stdout will be used. Note This is intended primarily for debugging. On some processor, which is not currently supported or not completely recognized by cpu_identify, one can still successfully get the raw data and write it to a file. libcpuid developers can later import this file and debug the detection code as if running on the actual hardware. The file is simple text format of 'something=value' pairs. Version info is also written, but the format is not intended to be neither backward- nor forward compatible. Returns zero if successful, and some negative number on error. The error message can be obtained by calling cpuid_error. See also cpu_error_t Author Generated automatically by Doxygen for libcpuid from the source code. libcpuid libcpuid(3)