X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=src%2Fmisc.cpp;h=832a9ac18f93465baca9eb9c54de107453c71916;hb=HEAD;hp=0af20e10d58f2b014f050ddcec8c23bd29cfc763;hpb=9048ac00db12a9ac48bff9b9eb145b30ff88d984;p=stockfish diff --git a/src/misc.cpp b/src/misc.cpp index 0af20e10..7d504241 100644 --- a/src/misc.cpp +++ b/src/misc.cpp @@ -1,6 +1,6 @@ /* Stockfish, a UCI chess playing engine derived from Glaurung 2.1 - Copyright (C) 2004-2021 The Stockfish developers (see AUTHORS file) + Copyright (C) 2004-2024 The Stockfish developers (see AUTHORS file) Stockfish is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -16,638 +16,489 @@ along with this program. If not, see . */ -#ifdef _WIN32 -#if _WIN32_WINNT < 0x0601 -#undef _WIN32_WINNT -#define _WIN32_WINNT 0x0601 // Force to include needed API prototypes -#endif - -#ifndef NOMINMAX -#define NOMINMAX -#endif - -#include -// The needed Windows API for processor groups could be missed from old Windows -// versions, so instead of calling them directly (forcing the linker to resolve -// the calls at compile time), try to load them at runtime. To do this we need -// first to define the corresponding function pointers. -extern "C" { -typedef bool(*fun1_t)(LOGICAL_PROCESSOR_RELATIONSHIP, - PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX, PDWORD); -typedef bool(*fun2_t)(USHORT, PGROUP_AFFINITY); -typedef bool(*fun3_t)(HANDLE, CONST GROUP_AFFINITY*, PGROUP_AFFINITY); -} -#endif +#include "misc.h" +#include +#include +#include +#include #include #include #include +#include +#include +#include #include -#include -#include +#include -#if defined(__linux__) && !defined(__ANDROID__) -#include -#include -#endif - -#if defined(__APPLE__) || defined(__ANDROID__) || defined(__OpenBSD__) || (defined(__GLIBCXX__) && !defined(_GLIBCXX_HAVE_ALIGNED_ALLOC) && !defined(_WIN32)) || defined(__e2k__) -#define POSIXALIGNEDALLOC -#include -#endif - -#include "misc.h" -#include "thread.h" - -using namespace std; +#include "types.h" namespace Stockfish { namespace { -/// Version number. If Version is left empty, then compile date in the format -/// DD-MM-YY and show in engine_info. -const string Version = ""; +// Version number or dev. +constexpr std::string_view version = "dev"; -/// Our fancy logging facility. The trick here is to replace cin.rdbuf() and -/// cout.rdbuf() with two Tie objects that tie cin and cout to a file stream. We -/// can toggle the logging of std::cout and std:cin at runtime whilst preserving -/// usual I/O functionality, all without changing a single line of code! -/// Idea from http://groups.google.com/group/comp.lang.c++/msg/1d941c0f26ea0d81 +// Our fancy logging facility. The trick here is to replace cin.rdbuf() and +// cout.rdbuf() with two Tie objects that tie cin and cout to a file stream. We +// can toggle the logging of std::cout and std:cin at runtime whilst preserving +// usual I/O functionality, all without changing a single line of code! +// Idea from http://groups.google.com/group/comp.lang.c++/msg/1d941c0f26ea0d81 -struct Tie: public streambuf { // MSVC requires split streambuf for cin and cout +struct Tie: public std::streambuf { // MSVC requires split streambuf for cin and cout - Tie(streambuf* b, streambuf* l) : buf(b), logBuf(l) {} + Tie(std::streambuf* b, std::streambuf* l) : + buf(b), + logBuf(l) {} - int sync() override { return logBuf->pubsync(), buf->pubsync(); } - int overflow(int c) override { return log(buf->sputc((char)c), "<< "); } - int underflow() override { return buf->sgetc(); } - int uflow() override { return log(buf->sbumpc(), ">> "); } + int sync() override { return logBuf->pubsync(), buf->pubsync(); } + int overflow(int c) override { return log(buf->sputc(char(c)), "<< "); } + int underflow() override { return buf->sgetc(); } + int uflow() override { return log(buf->sbumpc(), ">> "); } - streambuf *buf, *logBuf; + std::streambuf *buf, *logBuf; - int log(int c, const char* prefix) { + int log(int c, const char* prefix) { - static int last = '\n'; // Single log file + static int last = '\n'; // Single log file - if (last == '\n') - logBuf->sputn(prefix, 3); + if (last == '\n') + logBuf->sputn(prefix, 3); - return last = logBuf->sputc((char)c); - } + return last = logBuf->sputc(char(c)); + } }; class Logger { - Logger() : in(cin.rdbuf(), file.rdbuf()), out(cout.rdbuf(), file.rdbuf()) {} - ~Logger() { start(""); } + Logger() : + in(std::cin.rdbuf(), file.rdbuf()), + out(std::cout.rdbuf(), file.rdbuf()) {} + ~Logger() { start(""); } - ofstream file; - Tie in, out; + std::ofstream file; + Tie in, out; -public: - static void start(const std::string& fname) { + public: + static void start(const std::string& fname) { - static Logger l; + static Logger l; - if (l.file.is_open()) - { - cout.rdbuf(l.out.buf); - cin.rdbuf(l.in.buf); - l.file.close(); - } - - if (!fname.empty()) - { - l.file.open(fname, ifstream::out); - - if (!l.file.is_open()) + if (l.file.is_open()) { - cerr << "Unable to open debug log file " << fname << endl; - exit(EXIT_FAILURE); + std::cout.rdbuf(l.out.buf); + std::cin.rdbuf(l.in.buf); + l.file.close(); } - cin.rdbuf(&l.in); - cout.rdbuf(&l.out); - } - } -}; + if (!fname.empty()) + { + l.file.open(fname, std::ifstream::out); -} // namespace + if (!l.file.is_open()) + { + std::cerr << "Unable to open debug log file " << fname << std::endl; + exit(EXIT_FAILURE); + } + std::cin.rdbuf(&l.in); + std::cout.rdbuf(&l.out); + } + } +}; -/// engine_info() returns the full name of the current Stockfish version. This -/// will be either "Stockfish DD-MM-YY" (where DD-MM-YY is the date when -/// the program was compiled) or "Stockfish ", depending on whether -/// Version is empty. +} // namespace + + +// Returns the full name of the current Stockfish version. +// +// For local dev compiles we try to append the commit SHA and +// commit date from git. If that fails only the local compilation +// date is set and "nogit" is specified: +// Stockfish dev-YYYYMMDD-SHA +// or +// Stockfish dev-YYYYMMDD-nogit +// +// For releases (non-dev builds) we only include the version number: +// Stockfish version +std::string engine_info(bool to_uci) { + std::stringstream ss; + ss << "Stockfish " << version << std::setfill('0'); + + if constexpr (version == "dev") + { + ss << "-"; +#ifdef GIT_DATE + ss << stringify(GIT_DATE); +#else + constexpr std::string_view months("Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec"); -string engine_info(bool to_uci) { + std::string month, day, year; + std::stringstream date(__DATE__); // From compiler, format is "Sep 21 2008" - const string months("Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec"); - string month, day, year; - stringstream ss, date(__DATE__); // From compiler, format is "Sep 21 2008" + date >> month >> day >> year; + ss << year << std::setw(2) << std::setfill('0') << (1 + months.find(month) / 4) + << std::setw(2) << std::setfill('0') << day; +#endif - ss << "Stockfish " << Version << setfill('0'); + ss << "-"; - if (Version.empty()) - { - date >> month >> day >> year; - ss << setw(2) << day << setw(2) << (1 + months.find(month) / 4) << year.substr(2); - } +#ifdef GIT_SHA + ss << stringify(GIT_SHA); +#else + ss << "nogit"; +#endif + ss << "-asn"; + } - ss << (to_uci ? "\nid author ": " by ") - << "the Stockfish developers (see AUTHORS file)"; + ss << (to_uci ? "\nid author " : " by ") << "the Stockfish developers (see AUTHORS file)"; - return ss.str(); + return ss.str(); } -/// compiler_info() returns a string trying to describe the compiler we use - +// Returns a string trying to describe the compiler we use std::string compiler_info() { - #define stringify2(x) #x - #define stringify(x) stringify2(x) - #define make_version_string(major, minor, patch) stringify(major) "." stringify(minor) "." stringify(patch) - -/// Predefined macros hell: -/// -/// __GNUC__ Compiler is gcc, Clang or Intel on Linux -/// __INTEL_COMPILER Compiler is Intel -/// _MSC_VER Compiler is MSVC or Intel on Windows -/// _WIN32 Building on Windows (any) -/// _WIN64 Building on Windows 64 bit - - std::string compiler = "\nCompiled by "; - - #ifdef __clang__ - compiler += "clang++ "; - compiler += make_version_string(__clang_major__, __clang_minor__, __clang_patchlevel__); - #elif __INTEL_COMPILER - compiler += "Intel compiler "; - compiler += "(version "; - compiler += stringify(__INTEL_COMPILER) " update " stringify(__INTEL_COMPILER_UPDATE); - compiler += ")"; - #elif _MSC_VER - compiler += "MSVC "; - compiler += "(version "; - compiler += stringify(_MSC_FULL_VER) "." stringify(_MSC_BUILD); - compiler += ")"; - #elif defined(__e2k__) && defined(__LCC__) +#define make_version_string(major, minor, patch) \ + stringify(major) "." stringify(minor) "." stringify(patch) + + // Predefined macros hell: + // + // __GNUC__ Compiler is GCC, Clang or ICX + // __clang__ Compiler is Clang or ICX + // __INTEL_LLVM_COMPILER Compiler is ICX + // _MSC_VER Compiler is MSVC + // _WIN32 Building on Windows (any) + // _WIN64 Building on Windows 64 bit + + std::string compiler = "\nCompiled by : "; + +#if defined(__INTEL_LLVM_COMPILER) + compiler += "ICX "; + compiler += stringify(__INTEL_LLVM_COMPILER); +#elif defined(__clang__) + compiler += "clang++ "; + compiler += make_version_string(__clang_major__, __clang_minor__, __clang_patchlevel__); +#elif _MSC_VER + compiler += "MSVC "; + compiler += "(version "; + compiler += stringify(_MSC_FULL_VER) "." stringify(_MSC_BUILD); + compiler += ")"; +#elif defined(__e2k__) && defined(__LCC__) #define dot_ver2(n) \ - compiler += (char)'.'; \ - compiler += (char)('0' + (n) / 10); \ - compiler += (char)('0' + (n) % 10); - - compiler += "MCST LCC "; - compiler += "(version "; - compiler += std::to_string(__LCC__ / 100); - dot_ver2(__LCC__ % 100) - dot_ver2(__LCC_MINOR__) - compiler += ")"; - #elif __GNUC__ - compiler += "g++ (GNUC) "; - compiler += make_version_string(__GNUC__, __GNUC_MINOR__, __GNUC_PATCHLEVEL__); - #else - compiler += "Unknown compiler "; - compiler += "(unknown version)"; - #endif - - #if defined(__APPLE__) - compiler += " on Apple"; - #elif defined(__CYGWIN__) - compiler += " on Cygwin"; - #elif defined(__MINGW64__) - compiler += " on MinGW64"; - #elif defined(__MINGW32__) - compiler += " on MinGW32"; - #elif defined(__ANDROID__) - compiler += " on Android"; - #elif defined(__linux__) - compiler += " on Linux"; - #elif defined(_WIN64) - compiler += " on Microsoft Windows 64-bit"; - #elif defined(_WIN32) - compiler += " on Microsoft Windows 32-bit"; - #else - compiler += " on unknown system"; - #endif - - compiler += "\nCompilation settings include: "; - compiler += (Is64Bit ? " 64bit" : " 32bit"); - #if defined(USE_VNNI) + compiler += char('.'); \ + compiler += char('0' + (n) / 10); \ + compiler += char('0' + (n) % 10); + + compiler += "MCST LCC "; + compiler += "(version "; + compiler += std::to_string(__LCC__ / 100); + dot_ver2(__LCC__ % 100) dot_ver2(__LCC_MINOR__) compiler += ")"; +#elif __GNUC__ + compiler += "g++ (GNUC) "; + compiler += make_version_string(__GNUC__, __GNUC_MINOR__, __GNUC_PATCHLEVEL__); +#else + compiler += "Unknown compiler "; + compiler += "(unknown version)"; +#endif + +#if defined(__APPLE__) + compiler += " on Apple"; +#elif defined(__CYGWIN__) + compiler += " on Cygwin"; +#elif defined(__MINGW64__) + compiler += " on MinGW64"; +#elif defined(__MINGW32__) + compiler += " on MinGW32"; +#elif defined(__ANDROID__) + compiler += " on Android"; +#elif defined(__linux__) + compiler += " on Linux"; +#elif defined(_WIN64) + compiler += " on Microsoft Windows 64-bit"; +#elif defined(_WIN32) + compiler += " on Microsoft Windows 32-bit"; +#else + compiler += " on unknown system"; +#endif + + compiler += "\nCompilation architecture : "; +#if defined(ARCH) + compiler += stringify(ARCH); +#else + compiler += "(undefined architecture)"; +#endif + + compiler += "\nCompilation settings : "; + compiler += (Is64Bit ? "64bit" : "32bit"); +#if defined(USE_VNNI) compiler += " VNNI"; - #endif - #if defined(USE_AVX512) +#endif +#if defined(USE_AVX512) compiler += " AVX512"; - #endif - compiler += (HasPext ? " BMI2" : ""); - #if defined(USE_AVX2) +#endif + compiler += (HasPext ? " BMI2" : ""); +#if defined(USE_AVX2) compiler += " AVX2"; - #endif - #if defined(USE_SSE41) +#endif +#if defined(USE_SSE41) compiler += " SSE41"; - #endif - #if defined(USE_SSSE3) +#endif +#if defined(USE_SSSE3) compiler += " SSSE3"; - #endif - #if defined(USE_SSE2) +#endif +#if defined(USE_SSE2) compiler += " SSE2"; - #endif - compiler += (HasPopCnt ? " POPCNT" : ""); - #if defined(USE_MMX) - compiler += " MMX"; - #endif - #if defined(USE_NEON) +#endif + compiler += (HasPopCnt ? " POPCNT" : ""); +#if defined(USE_NEON_DOTPROD) + compiler += " NEON_DOTPROD"; +#elif defined(USE_NEON) compiler += " NEON"; - #endif +#endif - #if !defined(NDEBUG) +#if !defined(NDEBUG) compiler += " DEBUG"; - #endif - - compiler += "\n__VERSION__ macro expands to: "; - #ifdef __VERSION__ - compiler += __VERSION__; - #else - compiler += "(undefined macro)"; - #endif - compiler += "\n"; - - return compiler; -} - - -/// Debug functions used mainly to collect run-time statistics -static std::atomic hits[2], means[2]; - -void dbg_hit_on(bool b) { ++hits[0]; if (b) ++hits[1]; } -void dbg_hit_on(bool c, bool b) { if (c) dbg_hit_on(b); } -void dbg_mean_of(int v) { ++means[0]; means[1] += v; } - -void dbg_print() { - - if (hits[0]) - cerr << "Total " << hits[0] << " Hits " << hits[1] - << " hit rate (%) " << 100 * hits[1] / hits[0] << endl; - - if (means[0]) - cerr << "Total " << means[0] << " Mean " - << (double)means[1] / means[0] << endl; -} - - -/// Used to serialize access to std::cout to avoid multiple threads writing at -/// the same time. - -std::ostream& operator<<(std::ostream& os, SyncCout sc) { - - static std::mutex m; - - if (sc == IO_LOCK) - m.lock(); - - if (sc == IO_UNLOCK) - m.unlock(); - - return os; -} - - -/// Trampoline helper to avoid moving Logger to misc.h -void start_logger(const std::string& fname) { Logger::start(fname); } - - -/// prefetch() preloads the given address in L1/L2 cache. This is a non-blocking -/// function that doesn't stall the CPU waiting for data to be loaded from memory, -/// which can be quite slow. -#ifdef NO_PREFETCH - -void prefetch(void*) {} - -#else - -void prefetch(void* addr) { - -# if defined(__INTEL_COMPILER) - // This hack prevents prefetches from being optimized away by - // Intel compiler. Both MSVC and gcc seem not be affected by this. - __asm__ (""); -# endif - -# if defined(__INTEL_COMPILER) || defined(_MSC_VER) - _mm_prefetch((char*)addr, _MM_HINT_T0); -# else - __builtin_prefetch(addr); -# endif -} - #endif - -/// std_aligned_alloc() is our wrapper for systems where the c++17 implementation -/// does not guarantee the availability of aligned_alloc(). Memory allocated with -/// std_aligned_alloc() must be freed with std_aligned_free(). - -void* std_aligned_alloc(size_t alignment, size_t size) { - -#if defined(POSIXALIGNEDALLOC) - void *mem; - return posix_memalign(&mem, alignment, size) ? nullptr : mem; -#elif defined(_WIN32) - return _mm_malloc(size, alignment); + compiler += "\nCompiler __VERSION__ macro : "; +#ifdef __VERSION__ + compiler += __VERSION__; #else - return std::aligned_alloc(alignment, size); + compiler += "(undefined macro)"; #endif -} -void std_aligned_free(void* ptr) { + compiler += "\n"; -#if defined(POSIXALIGNEDALLOC) - free(ptr); -#elif defined(_WIN32) - _mm_free(ptr); -#else - free(ptr); -#endif + return compiler; } -/// aligned_large_pages_alloc() will return suitably aligned memory, if possible using large pages. -#if defined(_WIN32) +// Debug functions used mainly to collect run-time statistics +constexpr int MaxDebugSlots = 32; -static void* aligned_large_pages_alloc_windows(size_t allocSize) { +namespace { - #if !defined(_WIN64) - (void)allocSize; // suppress unused-parameter compiler warning - return nullptr; - #else +template +struct DebugInfo { + std::atomic data[N] = {0}; - HANDLE hProcessToken { }; - LUID luid { }; - void* mem = nullptr; + constexpr std::atomic& operator[](int index) { return data[index]; } +}; - const size_t largePageSize = GetLargePageMinimum(); - if (!largePageSize) - return nullptr; +struct DebugExtremes: public DebugInfo<3> { + DebugExtremes() { + data[1] = std::numeric_limits::min(); + data[2] = std::numeric_limits::max(); + } +}; - // We need SeLockMemoryPrivilege, so try to enable it for the process - if (!OpenProcessToken(GetCurrentProcess(), TOKEN_ADJUST_PRIVILEGES | TOKEN_QUERY, &hProcessToken)) - return nullptr; +DebugInfo<2> hit[MaxDebugSlots]; +DebugInfo<2> mean[MaxDebugSlots]; +DebugInfo<3> stdev[MaxDebugSlots]; +DebugInfo<6> correl[MaxDebugSlots]; +DebugExtremes extremes[MaxDebugSlots]; - if (LookupPrivilegeValue(NULL, SE_LOCK_MEMORY_NAME, &luid)) - { - TOKEN_PRIVILEGES tp { }; - TOKEN_PRIVILEGES prevTp { }; - DWORD prevTpLen = 0; +} // namespace - tp.PrivilegeCount = 1; - tp.Privileges[0].Luid = luid; - tp.Privileges[0].Attributes = SE_PRIVILEGE_ENABLED; +void dbg_hit_on(bool cond, int slot) { - // Try to enable SeLockMemoryPrivilege. Note that even if AdjustTokenPrivileges() succeeds, - // we still need to query GetLastError() to ensure that the privileges were actually obtained. - if (AdjustTokenPrivileges( - hProcessToken, FALSE, &tp, sizeof(TOKEN_PRIVILEGES), &prevTp, &prevTpLen) && - GetLastError() == ERROR_SUCCESS) - { - // Round up size to full pages and allocate - allocSize = (allocSize + largePageSize - 1) & ~size_t(largePageSize - 1); - mem = VirtualAlloc( - NULL, allocSize, MEM_RESERVE | MEM_COMMIT | MEM_LARGE_PAGES, PAGE_READWRITE); + ++hit[slot][0]; + if (cond) + ++hit[slot][1]; +} - // Privilege no longer needed, restore previous state - AdjustTokenPrivileges(hProcessToken, FALSE, &prevTp, 0, NULL, NULL); - } - } +void dbg_mean_of(int64_t value, int slot) { - CloseHandle(hProcessToken); + ++mean[slot][0]; + mean[slot][1] += value; +} - return mem; +void dbg_stdev_of(int64_t value, int slot) { - #endif + ++stdev[slot][0]; + stdev[slot][1] += value; + stdev[slot][2] += value * value; } -void* aligned_large_pages_alloc(size_t allocSize) { - - // Try to allocate large pages - void* mem = aligned_large_pages_alloc_windows(allocSize); +void dbg_extremes_of(int64_t value, int slot) { + ++extremes[slot][0]; - // Fall back to regular, page aligned, allocation if necessary - if (!mem) - mem = VirtualAlloc(NULL, allocSize, MEM_RESERVE | MEM_COMMIT, PAGE_READWRITE); + int64_t current_max = extremes[slot][1].load(); + while (current_max < value && !extremes[slot][1].compare_exchange_weak(current_max, value)) + {} - return mem; + int64_t current_min = extremes[slot][2].load(); + while (current_min > value && !extremes[slot][2].compare_exchange_weak(current_min, value)) + {} } -#else - -void* aligned_large_pages_alloc(size_t allocSize) { - -#if defined(__linux__) - constexpr size_t alignment = 2 * 1024 * 1024; // assumed 2MB page size -#else - constexpr size_t alignment = 4096; // assumed small page size -#endif +void dbg_correl_of(int64_t value1, int64_t value2, int slot) { - // round up to multiples of alignment - size_t size = ((allocSize + alignment - 1) / alignment) * alignment; - void *mem = std_aligned_alloc(alignment, size); -#if defined(MADV_HUGEPAGE) - madvise(mem, size, MADV_HUGEPAGE); -#endif - return mem; + ++correl[slot][0]; + correl[slot][1] += value1; + correl[slot][2] += value1 * value1; + correl[slot][3] += value2; + correl[slot][4] += value2 * value2; + correl[slot][5] += value1 * value2; } -#endif - +void dbg_print() { -/// aligned_large_pages_free() will free the previously allocated ttmem + int64_t n; + auto E = [&n](int64_t x) { return double(x) / n; }; + auto sqr = [](double x) { return x * x; }; -#if defined(_WIN32) + for (int i = 0; i < MaxDebugSlots; ++i) + if ((n = hit[i][0])) + std::cerr << "Hit #" << i << ": Total " << n << " Hits " << hit[i][1] + << " Hit Rate (%) " << 100.0 * E(hit[i][1]) << std::endl; -void aligned_large_pages_free(void* mem) { + for (int i = 0; i < MaxDebugSlots; ++i) + if ((n = mean[i][0])) + { + std::cerr << "Mean #" << i << ": Total " << n << " Mean " << E(mean[i][1]) << std::endl; + } - if (mem && !VirtualFree(mem, 0, MEM_RELEASE)) - { - DWORD err = GetLastError(); - std::cerr << "Failed to free large page memory. Error code: 0x" - << std::hex << err - << std::dec << std::endl; - exit(EXIT_FAILURE); - } -} + for (int i = 0; i < MaxDebugSlots; ++i) + if ((n = stdev[i][0])) + { + double r = sqrt(E(stdev[i][2]) - sqr(E(stdev[i][1]))); + std::cerr << "Stdev #" << i << ": Total " << n << " Stdev " << r << std::endl; + } -#else + for (int i = 0; i < MaxDebugSlots; ++i) + if ((n = extremes[i][0])) + { + std::cerr << "Extremity #" << i << ": Total " << n << " Min " << extremes[i][2] + << " Max " << extremes[i][1] << std::endl; + } -void aligned_large_pages_free(void *mem) { - std_aligned_free(mem); + for (int i = 0; i < MaxDebugSlots; ++i) + if ((n = correl[i][0])) + { + double r = (E(correl[i][5]) - E(correl[i][1]) * E(correl[i][3])) + / (sqrt(E(correl[i][2]) - sqr(E(correl[i][1]))) + * sqrt(E(correl[i][4]) - sqr(E(correl[i][3])))); + std::cerr << "Correl. #" << i << ": Total " << n << " Coefficient " << r << std::endl; + } } -#endif +// Used to serialize access to std::cout +// to avoid multiple threads writing at the same time. +std::ostream& operator<<(std::ostream& os, SyncCout sc) { -namespace WinProcGroup { + static std::mutex m; -#ifndef _WIN32 + if (sc == IO_LOCK) + m.lock(); -void bindThisThread(size_t) {} + if (sc == IO_UNLOCK) + m.unlock(); -#else - -/// best_group() retrieves logical processor information using Windows specific -/// API and returns the best group id for the thread with index idx. Original -/// code from Texel by Peter Österlund. - -int best_group(size_t idx) { - - int threads = 0; - int groups = 0; - int cores = 0; - DWORD returnLength = 0; - DWORD byteOffset = 0; - - // Early exit if the needed API is not available at runtime - HMODULE k32 = GetModuleHandle("Kernel32.dll"); - auto fun1 = (fun1_t)(void(*)())GetProcAddress(k32, "GetLogicalProcessorInformationEx"); - if (!fun1) - return -1; - - // First call to get returnLength. We expect it to fail due to null buffer - if (fun1(RelationAll, nullptr, &returnLength)) - return -1; - - // Once we know returnLength, allocate the buffer - SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *buffer, *ptr; - ptr = buffer = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)malloc(returnLength); - - // Second call, now we expect to succeed - if (!fun1(RelationAll, buffer, &returnLength)) - { - free(buffer); - return -1; - } - - while (byteOffset < returnLength) - { - if (ptr->Relationship == RelationGroup) - groups += ptr->Group.MaximumGroupCount; - - else if (ptr->Relationship == RelationProcessorCore) - { - cores++; - threads += (ptr->Processor.Flags == LTP_PC_SMT) ? 2 : 1; - } - - assert(ptr->Size); - byteOffset += ptr->Size; - ptr = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)(((char*)ptr) + ptr->Size); - } - - free(buffer); - - std::vector core_groups; - - // Run as many threads as possible on the same group until core limit is - // reached, then move on filling the next group. - for (int n = 0; n < groups; n++) - for (int i = 0; i < cores / groups; i++) - core_groups.push_back(n); - - // In case a core has more than one logical processor (we assume 2) and we - // have still threads to allocate, then spread them evenly across available - // groups. - for (int t = 0; t < threads - cores; t++) - core_groups.push_back(t % groups); - - // If we still have more threads than the total number of logical processors - // then return -1 and let the OS to decide what to do. - return idx < core_groups.size() ? core_groups[idx] : -1; + return os; } +void sync_cout_start() { std::cout << IO_LOCK; } +void sync_cout_end() { std::cout << IO_UNLOCK; } -/// bindThisThread() set the group affinity of the current thread +// Trampoline helper to avoid moving Logger to misc.h +void start_logger(const std::string& fname) { Logger::start(fname); } -void bindThisThread(size_t idx) { - // Use only local variables to be thread-safe - int group = best_group(idx); +#ifdef NO_PREFETCH - if (group == -1) - return; +void prefetch(const void*) {} - // Early exit if the needed API are not available at runtime - HMODULE k32 = GetModuleHandle("Kernel32.dll"); - auto fun2 = (fun2_t)(void(*)())GetProcAddress(k32, "GetNumaNodeProcessorMaskEx"); - auto fun3 = (fun3_t)(void(*)())GetProcAddress(k32, "SetThreadGroupAffinity"); +#else - if (!fun2 || !fun3) - return; +void prefetch(const void* addr) { - GROUP_AFFINITY affinity; - if (fun2(group, &affinity)) - fun3(GetCurrentThread(), &affinity, nullptr); + #if defined(_MSC_VER) + _mm_prefetch((char const*) addr, _MM_HINT_T0); + #else + __builtin_prefetch(addr); + #endif } #endif -} // namespace WinProcGroup - #ifdef _WIN32 -#include -#define GETCWD _getcwd + #include + #define GETCWD _getcwd #else -#include -#define GETCWD getcwd + #include + #define GETCWD getcwd #endif -namespace CommandLine { +size_t str_to_size_t(const std::string& s) { + unsigned long long value = std::stoull(s); + if (value > std::numeric_limits::max()) + std::exit(EXIT_FAILURE); + return static_cast(value); +} -string argv0; // path+name of the executable binary, as given by argv[0] -string binaryDirectory; // path of the executable directory -string workingDirectory; // path of the working directory +std::optional read_file_to_string(const std::string& path) { + std::ifstream f(path, std::ios_base::binary); + if (!f) + return std::nullopt; + return std::string(std::istreambuf_iterator(f), std::istreambuf_iterator()); +} -void init(int argc, char* argv[]) { - (void)argc; - string pathSeparator; +void remove_whitespace(std::string& s) { + s.erase(std::remove_if(s.begin(), s.end(), [](char c) { return std::isspace(c); }), s.end()); +} - // extract the path+name of the executable binary - argv0 = argv[0]; +bool is_whitespace(const std::string& s) { + return std::all_of(s.begin(), s.end(), [](char c) { return std::isspace(c); }); +} + +std::string CommandLine::get_binary_directory(std::string argv0) { + std::string pathSeparator; #ifdef _WIN32 pathSeparator = "\\"; - #ifdef _MSC_VER + #ifdef _MSC_VER // Under windows argv[0] may not have the extension. Also _get_pgmptr() had - // issues in some windows 10 versions, so check returned values carefully. + // issues in some Windows 10 versions, so check returned values carefully. char* pgmptr = nullptr; if (!_get_pgmptr(&pgmptr) && pgmptr != nullptr && *pgmptr) argv0 = pgmptr; - #endif + #endif #else pathSeparator = "/"; #endif - // extract the working directory - workingDirectory = ""; - char buff[40000]; - char* cwd = GETCWD(buff, 40000); - if (cwd) - workingDirectory = cwd; + // Extract the working directory + auto workingDirectory = CommandLine::get_working_directory(); - // extract the binary directory path from argv0 - binaryDirectory = argv0; - size_t pos = binaryDirectory.find_last_of("\\/"); + // Extract the binary directory path from argv0 + auto binaryDirectory = argv0; + size_t pos = binaryDirectory.find_last_of("\\/"); if (pos == std::string::npos) binaryDirectory = "." + pathSeparator; else binaryDirectory.resize(pos + 1); - // pattern replacement: "./" at the start of path is replaced by the working directory + // Pattern replacement: "./" at the start of path is replaced by the working directory if (binaryDirectory.find("." + pathSeparator) == 0) binaryDirectory.replace(0, 1, workingDirectory); + + return binaryDirectory; } +std::string CommandLine::get_working_directory() { + std::string workingDirectory = ""; + char buff[40000]; + char* cwd = GETCWD(buff, 40000); + if (cwd) + workingDirectory = cwd; + + return workingDirectory; +} -} // namespace CommandLine -} // namespace Stockfish +} // namespace Stockfish