Distribute and run LLMs with a single file.
C++ cross-platformggufllama-cpplocal-ailocal-inferencelocal-llmopen-source-aisingle-file-executablespeech-to-text
common.cpp: detect physical cores at runtime under cosmocc (fix #976) (#977)
Davide Eynardb2d21fd21507
diff --git a/llama.cpp.patches/patches/common_common.cpp.patch b/llama.cpp.patches/patches/common_common.cpp.patchindex dfbbcbfd..7e5a3001 100644--- a/llama.cpp.patches/patches/common_common.cpp.patch+++ b/llama.cpp.patches/patches/common_common.cpp.patch@@ -1,7 +1,58 @@diff --git a/common/common.cpp b/common/common.cpp--- a/llama.cpp/common/common.cpp+++ b/llama.cpp/common/common.cpp-@@ -972,6 +972,16 @@ std::string fs_get_cache_directory() {+@@ -33,6 +33,10 @@+ #include <sys/sysctl.h>+ #endif+++#if defined(COSMOCC)++#include <cosmo.h> // for IsLinux() and sysctlbyname() declarations++#endif+++ #if defined(_WIN32)+ #define WIN32_LEAN_AND_MEAN+ #ifndef NOMINMAX+@@ -71,7 +75,38 @@ common_time_meas::~common_time_meas() {+ //++ int32_t common_cpu_get_num_physical_cores() {+-#ifdef __linux__++#if defined(COSMOCC)++ // cosmocc cross-compiles for both Linux and macOS at once, so neither++ // __linux__ nor __APPLE__ is defined here; detect the host OS at runtime.++ if (IsLinux()) {++ std::unordered_set<std::string> siblings;++ for (uint32_t cpu=0; cpu < UINT32_MAX; ++cpu) {++ std::ifstream thread_siblings("/sys/devices/system/cpu/cpu"++ + std::to_string(cpu) + "/topology/thread_siblings");++ if (!thread_siblings.is_open()) {++ break; // no more cpus++ }++ std::string line;++ if (std::getline(thread_siblings, line)) {++ siblings.insert(line);++ }++ }++ if (!siblings.empty()) {++ return static_cast<int32_t>(siblings.size());++ }++ } else {++ int32_t num_physical_cores;++ size_t len = sizeof(num_physical_cores);++ int result = sysctlbyname("hw.perflevel0.physicalcpu", &num_physical_cores, &len, NULL, 0);++ if (result == 0) {++ return num_physical_cores;++ }++ result = sysctlbyname("hw.physicalcpu", &num_physical_cores, &len, NULL, 0);++ if (result == 0) {++ return num_physical_cores;++ }++ }++#elif defined(__linux__)+ // enumerate the set of thread siblings, num entries is num cores+ std::unordered_set<std::string> siblings;+ for (uint32_t cpu=0; cpu < UINT32_MAX; ++cpu) {+@@ -972,6 +1007,16 @@ std::string fs_get_cache_directory() {cache_directory = std::getenv("HOME") + std::string("/Library/Caches/");#elif defined(_WIN32)cache_directory = std::getenv("LOCALAPPDATA");@@ -18,7 +69,7 @@ diff --git a/common/common.cpp b/common/common.cpp#elif defined(__EMSCRIPTEN__)GGML_ABORT("not implemented on this platform");#else-@@ -1148,10 +1158,31 @@ common_init_result::common_init_result(common_params & params) :+@@ -1148,10 +1193,31 @@ common_init_result::common_init_result(common_params & params) :if (params.fit_params) {LOG_INF("%s: fitting params to device memory, for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on\n", __func__);