X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=src%2Fmisc.cpp;h=b46786dff08e0e98badfbbe1373a08d3979c8432;hb=b82d93ece484f833c994b40d9eddd959ba20ef92;hp=3fbdea35d94345db5b82ee96eaaffdf01bb28b96;hpb=571c2d6d8daf70de884c493b40cf0279e9b48c61;p=stockfish diff --git a/src/misc.cpp b/src/misc.cpp index 3fbdea35..b46786df 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) 2004-2020 The Stockfish developers (see AUTHORS file) + Copyright (C) 2004-2021 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 @@ -36,6 +36,8 @@ 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); +typedef bool(*fun4_t)(USHORT, PGROUP_AFFINITY, USHORT, PUSHORT); +typedef WORD(*fun5_t)(); } #endif @@ -51,7 +53,7 @@ typedef bool(*fun3_t)(HANDLE, CONST GROUP_AFFINITY*, PGROUP_AFFINITY); #include #endif -#if defined(__APPLE__) || defined(__ANDROID__) || defined(__OpenBSD__) || (defined(__GLIBCXX__) && !defined(_GLIBCXX_HAVE_ALIGNED_ALLOC) && !defined(_WIN32)) +#if defined(__APPLE__) || defined(__ANDROID__) || defined(__OpenBSD__) || (defined(__GLIBCXX__) && !defined(_GLIBCXX_HAVE_ALIGNED_ALLOC) && !defined(_WIN32)) || defined(__e2k__) #define POSIXALIGNEDALLOC #include #endif @@ -61,6 +63,8 @@ typedef bool(*fun3_t)(HANDLE, CONST GROUP_AFFINITY*, PGROUP_AFFINITY); using namespace std; +namespace Stockfish { + namespace { /// Version number. If Version is left empty, then compile date in the format @@ -108,7 +112,14 @@ public: static Logger l; - if (!fname.empty() && !l.file.is_open()) + 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); @@ -121,12 +132,6 @@ public: cin.rdbuf(&l.in); cout.rdbuf(&l.out); } - else if (fname.empty() && l.file.is_open()) - { - cout.rdbuf(l.out.buf); - cin.rdbuf(l.in.buf); - l.file.close(); - } } }; @@ -138,7 +143,7 @@ public: /// the program was compiled) or "Stockfish ", depending on whether /// Version is empty. -const string engine_info(bool to_uci) { +string engine_info(bool to_uci) { const string months("Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec"); string month, day, year; @@ -161,7 +166,7 @@ const string engine_info(bool to_uci) { /// compiler_info() returns a string trying to describe the compiler we use -const std::string compiler_info() { +std::string compiler_info() { #define stringify2(x) #x #define stringify(x) stringify2(x) @@ -190,6 +195,18 @@ const std::string compiler_info() { 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__); @@ -357,27 +374,16 @@ void std_aligned_free(void* ptr) { #endif } -/// aligned_ttmem_alloc() will return suitably aligned memory, if possible using large pages. -/// The returned pointer is the aligned one, while the mem argument is the one that needs -/// to be passed to free. With c++17 some of this functionality could be simplified. +/// aligned_large_pages_alloc() will return suitably aligned memory, if possible using large pages. -#if defined(__linux__) && !defined(__ANDROID__) - -void* aligned_ttmem_alloc(size_t allocSize, void*& mem) { - - constexpr size_t alignment = 2 * 1024 * 1024; // assumed 2MB page sizes - size_t size = ((allocSize + alignment - 1) / alignment) * alignment; // multiple of alignment - if (posix_memalign(&mem, alignment, size)) - mem = nullptr; -#if defined(MADV_HUGEPAGE) - madvise(mem, allocSize, MADV_HUGEPAGE); -#endif - return mem; -} +#if defined(_WIN32) -#elif defined(_WIN64) +static void* aligned_large_pages_alloc_windows(size_t allocSize) { -static void* aligned_ttmem_alloc_large_pages(size_t allocSize) { + #if !defined(_WIN64) + (void)allocSize; // suppress unused-parameter compiler warning + return nullptr; + #else HANDLE hProcessToken { }; LUID luid { }; @@ -420,25 +426,14 @@ static void* aligned_ttmem_alloc_large_pages(size_t allocSize) { CloseHandle(hProcessToken); return mem; -} -void* aligned_ttmem_alloc(size_t allocSize, void*& mem) { + #endif +} - static bool firstCall = true; +void* aligned_large_pages_alloc(size_t allocSize) { // Try to allocate large pages - mem = aligned_ttmem_alloc_large_pages(allocSize); - - // Suppress info strings on the first call. The first call occurs before 'uci' - // is received and in that case this output confuses some GUIs. - if (!firstCall) - { - if (mem) - sync_cout << "info string Hash table allocation: Windows large pages used." << sync_endl; - else - sync_cout << "info string Hash table allocation: Windows large pages not used." << sync_endl; - } - firstCall = false; + void* mem = aligned_large_pages_alloc_windows(allocSize); // Fall back to regular, page aligned, allocation if necessary if (!mem) @@ -449,37 +444,46 @@ void* aligned_ttmem_alloc(size_t allocSize, void*& mem) { #else -void* aligned_ttmem_alloc(size_t allocSize, void*& mem) { +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 - constexpr size_t alignment = 64; // assumed cache line size - size_t size = allocSize + alignment - 1; // allocate some extra space - mem = malloc(size); - void* ret = reinterpret_cast((uintptr_t(mem) + alignment - 1) & ~uintptr_t(alignment - 1)); - return ret; + // 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; } #endif -/// aligned_ttmem_free() will free the previously allocated ttmem +/// aligned_large_pages_free() will free the previously allocated ttmem -#if defined(_WIN64) +#if defined(_WIN32) -void aligned_ttmem_free(void* mem) { +void aligned_large_pages_free(void* mem) { if (mem && !VirtualFree(mem, 0, MEM_RELEASE)) { DWORD err = GetLastError(); - std::cerr << "Failed to free transposition table. Error code: 0x" << - std::hex << err << std::dec << std::endl; + std::cerr << "Failed to free large page memory. Error code: 0x" + << std::hex << err + << std::dec << std::endl; exit(EXIT_FAILURE); } } #else -void aligned_ttmem_free(void *mem) { - free(mem); +void aligned_large_pages_free(void *mem) { + std_aligned_free(mem); } #endif @@ -493,11 +497,11 @@ void bindThisThread(size_t) {} #else -/// best_group() retrieves logical processor information using Windows specific -/// API and returns the best group id for the thread with index idx. Original +/// best_node() retrieves logical processor information using Windows specific +/// API and returns the best node id for the thread with index idx. Original /// code from Texel by Peter Österlund. -int best_group(size_t idx) { +int best_node(size_t idx) { int threads = 0; int nodes = 0; @@ -511,7 +515,8 @@ int best_group(size_t idx) { if (!fun1) return -1; - // First call to get returnLength. We expect it to fail due to null buffer + // First call to GetLogicalProcessorInformationEx() to get returnLength. + // We expect the call to fail due to null buffer. if (fun1(RelationAll, nullptr, &returnLength)) return -1; @@ -519,7 +524,7 @@ int best_group(size_t idx) { SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *buffer, *ptr; ptr = buffer = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX*)malloc(returnLength); - // Second call, now we expect to succeed + // Second call to GetLogicalProcessorInformationEx(), now we expect to succeed if (!fun1(RelationAll, buffer, &returnLength)) { free(buffer); @@ -569,22 +574,38 @@ int best_group(size_t idx) { void bindThisThread(size_t idx) { // Use only local variables to be thread-safe - int group = best_group(idx); + int node = best_node(idx); - if (group == -1) + if (node == -1) return; // 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"); + auto fun4 = (fun4_t)(void(*)())GetProcAddress(k32, "GetNumaNodeProcessorMask2"); + auto fun5 = (fun5_t)(void(*)())GetProcAddress(k32, "GetMaximumProcessorGroupCount"); if (!fun2 || !fun3) return; - GROUP_AFFINITY affinity; - if (fun2(group, &affinity)) - fun3(GetCurrentThread(), &affinity, nullptr); + if (!fun4 || !fun5) + { + GROUP_AFFINITY affinity; + if (fun2(node, &affinity)) // GetNumaNodeProcessorMaskEx + fun3(GetCurrentThread(), &affinity, nullptr); // SetThreadGroupAffinity + } + else + { + // If a numa node has more than one processor group, we assume they are + // sized equal and we spread threads evenly across the groups. + USHORT elements, returnedElements; + elements = fun5(); // GetMaximumProcessorGroupCount + GROUP_AFFINITY *affinity = (GROUP_AFFINITY*)malloc(elements * sizeof(GROUP_AFFINITY)); + if (fun4(node, affinity, elements, &returnedElements)) // GetNumaNodeProcessorMask2 + fun3(GetCurrentThread(), &affinity[idx % returnedElements], nullptr); // SetThreadGroupAffinity + free(affinity); + } } #endif @@ -604,11 +625,10 @@ namespace CommandLine { 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 -string pathSeparator; // Separator for our current OS void init(int argc, char* argv[]) { (void)argc; - string separator; + string pathSeparator; // extract the path+name of the executable binary argv0 = argv[0]; @@ -648,3 +668,5 @@ void init(int argc, char* argv[]) { } // namespace CommandLine + +} // namespace Stockfish