From 8404aeb56a73ab24f9b295111de3b37a37f0b841 Mon Sep 17 00:00:00 2001 From: Alexandre Ganea Date: Thu, 13 Feb 2020 22:49:57 -0500 Subject: [PATCH] [Support] On Windows, ensure hardware_concurrency() extends to all CPU sockets and all NUMA groups The goal of this patch is to maximize CPU utilization on multi-socket or high core count systems, so that parallel computations such as LLD/ThinLTO can use all hardware threads in the system. Before this patch, on Windows, a maximum of 64 hardware threads could be used at most, in some cases dispatched only on one CPU socket. == Background == Windows doesn't have a flat cpu_set_t like Linux. Instead, it projects hardware CPUs (or NUMA nodes) to applications through a concept of "processor groups". A "processor" is the smallest unit of execution on a CPU, that is, an hyper-thread if SMT is active; a core otherwise. There's a limit of 32-bit processors on older 32-bit versions of Windows, which later was raised to 64-processors with 64-bit versions of Windows. This limit comes from the affinity mask, which historically is represented by the sizeof(void*). Consequently, the concept of "processor groups" was introduced for dealing with systems with more than 64 hyper-threads. By default, the Windows OS assigns only one "processor group" to each starting application, in a round-robin manner. If the application wants to use more processors, it needs to programmatically enable it, by assigning threads to other "processor groups". This also means that affinity cannot cross "processor group" boundaries; one can only specify a "preferred" group on start-up, but the application is free to allocate more groups if it wants to. This creates a peculiar situation, where newer CPUs like the AMD EPYC 7702P (64-cores, 128-hyperthreads) are projected by the OS as two (2) "processor groups". This means that by default, an application can only use half of the cores. This situation could only get worse in the years to come, as dies with more cores will appear on the market. == The problem == The heavyweight_hardware_concurrency() API was introduced so that only *one hardware thread per core* was used. Once that API returns, that original intention is lost, only the number of threads is retained. Consider a situation, on Windows, where the system has 2 CPU sockets, 18 cores each, each core having 2 hyper-threads, for a total of 72 hyper-threads. Both heavyweight_hardware_concurrency() and hardware_concurrency() currently return 36, because on Windows they are simply wrappers over std::thread::hardware_concurrency() -- which can only return processors from the current "processor group". == The changes in this patch == To solve this situation, we capture (and retain) the initial intention until the point of usage, through a new ThreadPoolStrategy class. The number of threads to use is deferred as late as possible, until the moment where the std::threads are created (ThreadPool in the case of ThinLTO). When using hardware_concurrency(), setting ThreadCount to 0 now means to use all the possible hardware CPU (SMT) threads. Providing a ThreadCount above to the maximum number of threads will have no effect, the maximum will be used instead. The heavyweight_hardware_concurrency() is similar to hardware_concurrency(), except that only one thread per hardware *core* will be used. When LLVM_ENABLE_THREADS is OFF, the threading APIs will always return 1, to ensure any caller loops will be exercised at least once. Differential Revision: https://reviews.llvm.org/D71775 --- clang-tools-extra/clang-doc/tool/ClangDocMain.cpp | 3 +- clang-tools-extra/clangd/TUScheduler.cpp | 8 +- clang-tools-extra/clangd/index/Background.cpp | 5 +- clang-tools-extra/clangd/index/Background.h | 2 +- clang-tools-extra/clangd/index/BackgroundRebuild.h | 4 +- clang/lib/Tooling/AllTUsExecution.cpp | 3 +- .../DependencyScanningFilesystem.cpp | 3 +- clang/tools/clang-scan-deps/ClangScanDeps.cpp | 14 +- lld/ELF/SyntheticSections.cpp | 8 +- llvm/include/llvm/LTO/LTO.h | 3 +- llvm/include/llvm/Support/ThreadPool.h | 17 ++- llvm/include/llvm/Support/Threading.h | 69 +++++++-- llvm/lib/CodeGen/ParallelCG.cpp | 2 +- llvm/lib/DWARFLinker/DWARFLinker.cpp | 2 +- llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp | 2 +- llvm/lib/ExecutionEngine/Orc/LLJIT.cpp | 3 +- llvm/lib/LTO/LTO.cpp | 6 +- llvm/lib/LTO/LTOBackend.cpp | 3 +- llvm/lib/LTO/ThinLTOCodeGenerator.cpp | 6 +- llvm/lib/Support/Host.cpp | 7 +- llvm/lib/Support/Parallel.cpp | 14 +- llvm/lib/Support/ThreadPool.cpp | 23 ++- llvm/lib/Support/Threading.cpp | 46 +++--- llvm/lib/Support/Unix/Threading.inc | 24 +++ llvm/lib/Support/Windows/Threading.inc | 162 +++++++++++++++++++++ llvm/tools/dsymutil/dsymutil.cpp | 9 +- llvm/tools/gold/gold-plugin.cpp | 4 +- llvm/tools/llvm-cov/CodeCoverage.cpp | 6 +- llvm/tools/llvm-cov/CoverageExporterJson.cpp | 8 +- llvm/tools/llvm-cov/CoverageReport.cpp | 7 +- llvm/tools/llvm-lto2/llvm-lto2.cpp | 4 +- llvm/tools/llvm-profdata/llvm-profdata.cpp | 9 +- llvm/unittests/Support/Host.cpp | 3 +- llvm/unittests/Support/TaskQueueTest.cpp | 6 +- llvm/unittests/Support/ThreadPool.cpp | 48 +++++- llvm/unittests/Support/Threading.cpp | 3 +- mlir/lib/Pass/Pass.cpp | 3 +- 37 files changed, 406 insertions(+), 143 deletions(-) diff --git a/clang-tools-extra/clang-doc/tool/ClangDocMain.cpp b/clang-tools-extra/clang-doc/tool/ClangDocMain.cpp index eae0d84..cfc8ff6 100644 --- a/clang-tools-extra/clang-doc/tool/ClangDocMain.cpp +++ b/clang-tools-extra/clang-doc/tool/ClangDocMain.cpp @@ -268,8 +268,7 @@ int main(int argc, const char **argv) { Error = false; llvm::sys::Mutex IndexMutex; // ExecutorConcurrency is a flag exposed by AllTUsExecution.h - llvm::ThreadPool Pool(ExecutorConcurrency == 0 ? llvm::hardware_concurrency() - : ExecutorConcurrency); + llvm::ThreadPool Pool(llvm::hardware_concurrency(ExecutorConcurrency)); for (auto &Group : USRToBitcode) { Pool.async([&]() { std::vector> Infos; diff --git a/clang-tools-extra/clangd/TUScheduler.cpp b/clang-tools-extra/clangd/TUScheduler.cpp index 5a1caa9..f59c19e 100644 --- a/clang-tools-extra/clangd/TUScheduler.cpp +++ b/clang-tools-extra/clangd/TUScheduler.cpp @@ -842,13 +842,7 @@ std::string renderTUAction(const TUAction &Action) { } // namespace unsigned getDefaultAsyncThreadsCount() { - unsigned HardwareConcurrency = llvm::heavyweight_hardware_concurrency(); - // heavyweight_hardware_concurrency may fall back to hardware_concurrency. - // C++ standard says that hardware_concurrency() may return 0; fallback to 1 - // worker thread in that case. - if (HardwareConcurrency == 0) - return 1; - return HardwareConcurrency; + return llvm::heavyweight_hardware_concurrency().compute_thread_count(); } FileStatus TUStatus::render(PathRef File) const { diff --git a/clang-tools-extra/clangd/index/Background.cpp b/clang-tools-extra/clangd/index/Background.cpp index ab80113..c254123 100644 --- a/clang-tools-extra/clangd/index/Background.cpp +++ b/clang-tools-extra/clangd/index/Background.cpp @@ -148,9 +148,10 @@ BackgroundIndex::BackgroundIndex( CDB.watch([&](const std::vector &ChangedFiles) { enqueue(ChangedFiles); })) { - assert(ThreadPoolSize > 0 && "Thread pool size can't be zero."); + assert(Rebuilder.TUsBeforeFirstBuild > 0 && + "Thread pool size can't be zero."); assert(this->IndexStorageFactory && "Storage factory can not be null!"); - for (unsigned I = 0; I < ThreadPoolSize; ++I) { + for (unsigned I = 0; I < Rebuilder.TUsBeforeFirstBuild; ++I) { ThreadPool.runAsync("background-worker-" + llvm::Twine(I + 1), [this] { WithContext Ctx(this->BackgroundContext.clone()); Queue.work([&] { Rebuilder.idle(); }); diff --git a/clang-tools-extra/clangd/index/Background.h b/clang-tools-extra/clangd/index/Background.h index b11008d..2ae11c7 100644 --- a/clang-tools-extra/clangd/index/Background.h +++ b/clang-tools-extra/clangd/index/Background.h @@ -135,7 +135,7 @@ public: Context BackgroundContext, const FileSystemProvider &, const GlobalCompilationDatabase &CDB, BackgroundIndexStorage::Factory IndexStorageFactory, - size_t ThreadPoolSize = llvm::heavyweight_hardware_concurrency(), + size_t ThreadPoolSize = 0, // 0 = use all hardware threads std::function OnProgress = nullptr); ~BackgroundIndex(); // Blocks while the current task finishes. diff --git a/clang-tools-extra/clangd/index/BackgroundRebuild.h b/clang-tools-extra/clangd/index/BackgroundRebuild.h index d74c28b..295f705 100644 --- a/clang-tools-extra/clangd/index/BackgroundRebuild.h +++ b/clang-tools-extra/clangd/index/BackgroundRebuild.h @@ -49,7 +49,9 @@ class BackgroundIndexRebuilder { public: BackgroundIndexRebuilder(SwapIndex *Target, FileSymbols *Source, unsigned Threads) - : TUsBeforeFirstBuild(Threads), Target(Target), Source(Source) {} + : TUsBeforeFirstBuild(llvm::heavyweight_hardware_concurrency(Threads) + .compute_thread_count()), + Target(Target), Source(Source) {} // Called to indicate a TU has been indexed. // May rebuild, if enough TUs have been indexed. diff --git a/clang/lib/Tooling/AllTUsExecution.cpp b/clang/lib/Tooling/AllTUsExecution.cpp index d85075f..777857a 100644 --- a/clang/lib/Tooling/AllTUsExecution.cpp +++ b/clang/lib/Tooling/AllTUsExecution.cpp @@ -114,8 +114,7 @@ llvm::Error AllTUsToolExecutor::execute( auto &Action = Actions.front(); { - llvm::ThreadPool Pool(ThreadCount == 0 ? llvm::hardware_concurrency() - : ThreadCount); + llvm::ThreadPool Pool(llvm::hardware_concurrency(ThreadCount)); for (std::string File : Files) { Pool.async( [&](std::string Path) { diff --git a/clang/lib/Tooling/DependencyScanning/DependencyScanningFilesystem.cpp b/clang/lib/Tooling/DependencyScanning/DependencyScanningFilesystem.cpp index b4d5a29..b1b87e7 100644 --- a/clang/lib/Tooling/DependencyScanning/DependencyScanningFilesystem.cpp +++ b/clang/lib/Tooling/DependencyScanning/DependencyScanningFilesystem.cpp @@ -106,7 +106,8 @@ DependencyScanningFilesystemSharedCache:: // sharding gives a performance edge by reducing the lock contention. // FIXME: A better heuristic might also consider the OS to account for // the different cost of lock contention on different OSes. - NumShards = std::max(2u, llvm::hardware_concurrency() / 4); + NumShards = + std::max(2u, llvm::hardware_concurrency().compute_thread_count() / 4); CacheShards = std::make_unique(NumShards); } diff --git a/clang/tools/clang-scan-deps/ClangScanDeps.cpp b/clang/tools/clang-scan-deps/ClangScanDeps.cpp index c499048..9105c61 100644 --- a/clang/tools/clang-scan-deps/ClangScanDeps.cpp +++ b/clang/tools/clang-scan-deps/ClangScanDeps.cpp @@ -485,15 +485,9 @@ int main(int argc, const char **argv) { DependencyScanningService Service(ScanMode, Format, ReuseFileManager, SkipExcludedPPRanges); -#if LLVM_ENABLE_THREADS - unsigned NumWorkers = - NumThreads == 0 ? llvm::hardware_concurrency() : NumThreads; -#else - unsigned NumWorkers = 1; -#endif - llvm::ThreadPool Pool(NumWorkers); + llvm::ThreadPool Pool(llvm::hardware_concurrency(NumThreads)); std::vector> WorkerTools; - for (unsigned I = 0; I < NumWorkers; ++I) + for (unsigned I = 0; I < Pool.getThreadCount(); ++I) WorkerTools.push_back(std::make_unique(Service)); std::vector Inputs; @@ -508,9 +502,9 @@ int main(int argc, const char **argv) { if (Verbose) { llvm::outs() << "Running clang-scan-deps on " << Inputs.size() - << " files using " << NumWorkers << " workers\n"; + << " files using " << Pool.getThreadCount() << " workers\n"; } - for (unsigned I = 0; I < NumWorkers; ++I) { + for (unsigned I = 0; I < Pool.getThreadCount(); ++I) { Pool.async([I, &Lock, &Index, &Inputs, &HadErrors, &FD, &WorkerTools, &DependencyOS, &Errs]() { llvm::StringSet<> AlreadySeenModules; diff --git a/lld/ELF/SyntheticSections.cpp b/lld/ELF/SyntheticSections.cpp index bd01bc0..3a0251c 100644 --- a/lld/ELF/SyntheticSections.cpp +++ b/lld/ELF/SyntheticSections.cpp @@ -2747,8 +2747,8 @@ createSymbols(ArrayRef> nameAttrs, size_t numShards = 32; size_t concurrency = 1; if (threadsEnabled) - concurrency = - std::min(PowerOf2Floor(hardware_concurrency()), numShards); + concurrency = std::min( + hardware_concurrency().compute_thread_count(), numShards); // A sharded map to uniquify symbols by name. std::vector> map(numShards); @@ -3191,8 +3191,8 @@ void MergeNoTailSection::finalizeContents() { // operations in the following tight loop. size_t concurrency = 1; if (threadsEnabled) - concurrency = - std::min(PowerOf2Floor(hardware_concurrency()), numShards); + concurrency = std::min( + hardware_concurrency().compute_thread_count(), numShards); // Add section pieces to the builders. parallelForEachN(0, concurrency, [&](size_t threadId) { diff --git a/llvm/include/llvm/LTO/LTO.h b/llvm/include/llvm/LTO/LTO.h index e864e05..df28533 100644 --- a/llvm/include/llvm/LTO/LTO.h +++ b/llvm/include/llvm/LTO/LTO.h @@ -227,7 +227,8 @@ using ThinBackend = std::function( AddStreamFn AddStream, NativeObjectCache Cache)>; /// This ThinBackend runs the individual backend jobs in-process. -ThinBackend createInProcessThinBackend(unsigned ParallelismLevel); +/// The default value means to use one job per hardware core (not hyper-thread). +ThinBackend createInProcessThinBackend(unsigned ParallelismLevel = 0); /// This ThinBackend writes individual module indexes to files, instead of /// running the individual backend jobs. This backend is for distributed builds diff --git a/llvm/include/llvm/Support/ThreadPool.h b/llvm/include/llvm/Support/ThreadPool.h index 4bcbaa3..2036f46 100644 --- a/llvm/include/llvm/Support/ThreadPool.h +++ b/llvm/include/llvm/Support/ThreadPool.h @@ -13,7 +13,9 @@ #ifndef LLVM_SUPPORT_THREAD_POOL_H #define LLVM_SUPPORT_THREAD_POOL_H +#include "llvm/ADT/BitVector.h" #include "llvm/Config/llvm-config.h" +#include "llvm/Support/Threading.h" #include "llvm/Support/thread.h" #include @@ -38,12 +40,11 @@ public: using TaskTy = std::function; using PackagedTaskTy = std::packaged_task; - /// Construct a pool with the number of threads found by - /// hardware_concurrency(). - ThreadPool(); - - /// Construct a pool of \p ThreadCount threads - ThreadPool(unsigned ThreadCount); + /// Construct a pool using the hardware strategy \p S for mapping hardware + /// execution resources (threads, cores, CPUs) + /// Defaults to using the maximum execution resources in the system, but + /// excluding any resources contained in the affinity mask. + ThreadPool(ThreadPoolStrategy S = hardware_concurrency()); /// Blocking destructor: the pool will wait for all the threads to complete. ~ThreadPool(); @@ -68,6 +69,8 @@ public: /// It is an error to try to add new tasks while blocking on this call. void wait(); + unsigned getThreadCount() const { return ThreadCount; } + private: /// Asynchronous submission of a task to the pool. The returned future can be /// used to wait for the task to finish and is *non-blocking* on destruction. @@ -94,6 +97,8 @@ private: /// Signal for the destruction of the pool, asking thread to exit. bool EnableFlag; #endif + + unsigned ThreadCount; }; } diff --git a/llvm/include/llvm/Support/Threading.h b/llvm/include/llvm/Support/Threading.h index bacab8f..d3d4a37 100644 --- a/llvm/include/llvm/Support/Threading.h +++ b/llvm/include/llvm/Support/Threading.h @@ -14,6 +14,7 @@ #ifndef LLVM_SUPPORT_THREADING_H #define LLVM_SUPPORT_THREADING_H +#include "llvm/ADT/BitVector.h" #include "llvm/ADT/FunctionExtras.h" #include "llvm/ADT/SmallVector.h" #include "llvm/Config/llvm-config.h" // for LLVM_ON_UNIX @@ -143,20 +144,52 @@ void llvm_execute_on_thread_async( #endif } - /// Get the amount of currency to use for tasks requiring significant - /// memory or other resources. Currently based on physical cores, if - /// available for the host system, otherwise falls back to - /// thread::hardware_concurrency(). - /// Returns 1 when LLVM is configured with LLVM_ENABLE_THREADS=OFF - unsigned heavyweight_hardware_concurrency(); - - /// Get the number of threads that the current program can execute - /// concurrently. On some systems std::thread::hardware_concurrency() returns - /// the total number of cores, without taking affinity into consideration. - /// Returns 1 when LLVM is configured with LLVM_ENABLE_THREADS=OFF. - /// Fallback to std::thread::hardware_concurrency() if sched_getaffinity is - /// not available. - unsigned hardware_concurrency(); + /// This tells how a thread pool will be used + class ThreadPoolStrategy { + public: + // The default value (0) means all available threads should be used, + // excluding affinity mask. If set, this value only represents a suggested + // high bound, the runtime might choose a lower value (not higher). + unsigned ThreadsRequested = 0; + + // If SMT is active, use hyper threads. If false, there will be only one + // std::thread per core. + bool UseHyperThreads = true; + + /// Retrieves the max available threads for the current strategy. This + /// accounts for affinity masks and takes advantage of all CPU sockets. + unsigned compute_thread_count() const; + + /// Assign the current thread to an ideal hardware CPU or NUMA node. In a + /// multi-socket system, this ensures threads are assigned to all CPU + /// sockets. \p ThreadPoolNum represents a number bounded by [0, + /// compute_thread_count()). + void apply_thread_strategy(unsigned ThreadPoolNum) const; + }; + + /// Returns a thread strategy for tasks requiring significant memory or other + /// resources. To be used for workloads where hardware_concurrency() proves to + /// be less efficient. Avoid this strategy if doing lots of I/O. Currently + /// based on physical cores, if available for the host system, otherwise falls + /// back to hardware_concurrency(). Returns 1 when LLVM is configured with + /// LLVM_ENABLE_THREADS = OFF + inline ThreadPoolStrategy + heavyweight_hardware_concurrency(unsigned ThreadCount = 0) { + ThreadPoolStrategy S; + S.UseHyperThreads = false; + S.ThreadsRequested = ThreadCount; + return S; + } + + /// Returns a default thread strategy where all available hardware ressources + /// are to be used, except for those initially excluded by an affinity mask. + /// This function takes affinity into consideration. Returns 1 when LLVM is + /// configured with LLVM_ENABLE_THREADS=OFF. + inline ThreadPoolStrategy hardware_concurrency(unsigned ThreadCount = 0) { + ThreadPoolStrategy S; + S.ThreadsRequested = ThreadCount; + return S; + } /// Return the current thread id, as used in various OS system calls. /// Note that not all platforms guarantee that the value returned will be @@ -184,6 +217,14 @@ void llvm_execute_on_thread_async( /// the operation succeeded or failed is returned. void get_thread_name(SmallVectorImpl &Name); + /// Returns a mask that represents on which hardware thread, core, CPU, NUMA + /// group, the calling thread can be executed. On Windows, threads cannot + /// cross CPU boundaries. + llvm::BitVector get_thread_affinity_mask(); + + /// Returns how many physical CPUs or NUMA groups the system has. + unsigned get_cpus(); + enum class ThreadPriority { Background = 0, Default = 1, diff --git a/llvm/lib/CodeGen/ParallelCG.cpp b/llvm/lib/CodeGen/ParallelCG.cpp index 7dbd830..c19ed1f 100644 --- a/llvm/lib/CodeGen/ParallelCG.cpp +++ b/llvm/lib/CodeGen/ParallelCG.cpp @@ -51,7 +51,7 @@ std::unique_ptr llvm::splitCodeGen( // Create ThreadPool in nested scope so that threads will be joined // on destruction. { - ThreadPool CodegenThreadPool(OSs.size()); + ThreadPool CodegenThreadPool(hardware_concurrency(OSs.size())); int ThreadCount = 0; SplitModule( diff --git a/llvm/lib/DWARFLinker/DWARFLinker.cpp b/llvm/lib/DWARFLinker/DWARFLinker.cpp index 03919c8..715ad24 100644 --- a/llvm/lib/DWARFLinker/DWARFLinker.cpp +++ b/llvm/lib/DWARFLinker/DWARFLinker.cpp @@ -2446,7 +2446,7 @@ bool DWARFLinker::link() { } EmitLambda(); } else { - ThreadPool Pool(2); + ThreadPool Pool(hardware_concurrency(2)); Pool.async(AnalyzeAll); Pool.async(CloneAll); Pool.wait(); diff --git a/llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp b/llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp index e01b6b6..c3bf71f 100644 --- a/llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp +++ b/llvm/lib/DebugInfo/GSYM/DwarfTransformer.cpp @@ -445,7 +445,7 @@ Error DwarfTransformer::convert(uint32_t NumThreads) { // Now parse all DIEs in case we have cross compile unit references in a // thread pool. - ThreadPool pool(NumThreads); + ThreadPool pool(hardware_concurrency(NumThreads)); for (const auto &CU : DICtx.compile_units()) pool.async([&CU]() { CU->getUnitDIE(false /*CUDieOnly*/); }); pool.wait(); diff --git a/llvm/lib/ExecutionEngine/Orc/LLJIT.cpp b/llvm/lib/ExecutionEngine/Orc/LLJIT.cpp index f81e584..4218ca4 100644 --- a/llvm/lib/ExecutionEngine/Orc/LLJIT.cpp +++ b/llvm/lib/ExecutionEngine/Orc/LLJIT.cpp @@ -157,7 +157,8 @@ LLJIT::LLJIT(LLJITBuilderState &S, Error &Err) if (S.NumCompileThreads > 0) { TransformLayer->setCloneToNewContextOnEmit(true); - CompileThreads = std::make_unique(S.NumCompileThreads); + CompileThreads = + std::make_unique(hardware_concurrency(S.NumCompileThreads)); ES->setDispatchMaterialization( [this](JITDylib &JD, std::unique_ptr MU) { // FIXME: Switch to move capture once we have c++14. diff --git a/llvm/lib/LTO/LTO.cpp b/llvm/lib/LTO/LTO.cpp index 1d23c6b..f8affcb 100644 --- a/llvm/lib/LTO/LTO.cpp +++ b/llvm/lib/LTO/LTO.cpp @@ -477,8 +477,7 @@ LTO::RegularLTOState::RegularLTOState(unsigned ParallelCodeGenParallelismLevel, LTO::ThinLTOState::ThinLTOState(ThinBackend Backend) : Backend(Backend), CombinedIndex(/*HaveGVs*/ false) { if (!Backend) - this->Backend = - createInProcessThinBackend(llvm::heavyweight_hardware_concurrency()); + this->Backend = createInProcessThinBackend(); } LTO::LTO(Config Conf, ThinBackend Backend, @@ -1095,7 +1094,8 @@ public: const StringMap &ModuleToDefinedGVSummaries, AddStreamFn AddStream, NativeObjectCache Cache) : ThinBackendProc(Conf, CombinedIndex, ModuleToDefinedGVSummaries), - BackendThreadPool(ThinLTOParallelismLevel), + BackendThreadPool( + heavyweight_hardware_concurrency(ThinLTOParallelismLevel)), AddStream(std::move(AddStream)), Cache(std::move(Cache)) { for (auto &Name : CombinedIndex.cfiFunctionDefs()) CfiFunctionDefs.insert( diff --git a/llvm/lib/LTO/LTOBackend.cpp b/llvm/lib/LTO/LTOBackend.cpp index b854715..ec57744 100644 --- a/llvm/lib/LTO/LTOBackend.cpp +++ b/llvm/lib/LTO/LTOBackend.cpp @@ -375,7 +375,8 @@ void codegen(const Config &Conf, TargetMachine *TM, AddStreamFn AddStream, void splitCodeGen(const Config &C, TargetMachine *TM, AddStreamFn AddStream, unsigned ParallelCodeGenParallelismLevel, std::unique_ptr Mod) { - ThreadPool CodegenThreadPool(ParallelCodeGenParallelismLevel); + ThreadPool CodegenThreadPool( + heavyweight_hardware_concurrency(ParallelCodeGenParallelismLevel)); unsigned ThreadCount = 0; const Target *T = &TM->getTarget(); diff --git a/llvm/lib/LTO/ThinLTOCodeGenerator.cpp b/llvm/lib/LTO/ThinLTOCodeGenerator.cpp index a4f2702..152f0af 100644 --- a/llvm/lib/LTO/ThinLTOCodeGenerator.cpp +++ b/llvm/lib/LTO/ThinLTOCodeGenerator.cpp @@ -80,8 +80,8 @@ extern cl::opt RemarksFormat; namespace { -static cl::opt - ThreadCount("threads", cl::init(llvm::heavyweight_hardware_concurrency())); +// Default to using one job per hardware core in the system +static cl::opt ThreadCount("threads", cl::init(0)); // Simple helper to save temporary files for debug. static void saveTempBitcode(const Module &TheModule, StringRef TempDir, @@ -1042,7 +1042,7 @@ void ThinLTOCodeGenerator::run() { // Parallel optimizer + codegen { - ThreadPool Pool(ThreadCount); + ThreadPool Pool(heavyweight_hardware_concurrency(ThreadCount)); for (auto IndexCount : ModulesOrdering) { auto &Mod = Modules[IndexCount]; Pool.async([&](int count) { diff --git a/llvm/lib/Support/Host.cpp b/llvm/lib/Support/Host.cpp index ef38c1c..7e772b2 100644 --- a/llvm/lib/Support/Host.cpp +++ b/llvm/lib/Support/Host.cpp @@ -1266,7 +1266,7 @@ StringRef sys::getHostCPUName() { return "generic"; } // On Linux, the number of physical cores can be computed from /proc/cpuinfo, // using the number of unique physical/core id pairs. The following // implementation reads the /proc/cpuinfo format on an x86_64 system. -static int computeHostNumPhysicalCores() { +int computeHostNumPhysicalCores() { // Read /proc/cpuinfo as a stream (until EOF reached). It cannot be // mmapped because it appears to have 0 size. llvm::ErrorOr> Text = @@ -1312,7 +1312,7 @@ static int computeHostNumPhysicalCores() { #include // Gets the number of *physical cores* on the machine. -static int computeHostNumPhysicalCores() { +int computeHostNumPhysicalCores() { uint32_t count; size_t len = sizeof(count); sysctlbyname("hw.physicalcpu", &count, &len, NULL, 0); @@ -1326,6 +1326,9 @@ static int computeHostNumPhysicalCores() { } return count; } +#elif defined(_WIN32) +// Defined in llvm/lib/Support/Windows/Threading.inc +int computeHostNumPhysicalCores(); #else // On other systems, return -1 to indicate unknown. static int computeHostNumPhysicalCores() { return -1; } diff --git a/llvm/lib/Support/Parallel.cpp b/llvm/lib/Support/Parallel.cpp index 523665d..0272a53 100644 --- a/llvm/lib/Support/Parallel.cpp +++ b/llvm/lib/Support/Parallel.cpp @@ -39,20 +39,21 @@ public: /// in filo order. class ThreadPoolExecutor : public Executor { public: - explicit ThreadPoolExecutor(unsigned ThreadCount = hardware_concurrency()) { + explicit ThreadPoolExecutor(ThreadPoolStrategy S = hardware_concurrency()) { + unsigned ThreadCount = S.compute_thread_count(); // Spawn all but one of the threads in another thread as spawning threads // can take a while. Threads.reserve(ThreadCount); Threads.resize(1); std::lock_guard Lock(Mutex); - Threads[0] = std::thread([&, ThreadCount] { - for (unsigned i = 1; i < ThreadCount; ++i) { - Threads.emplace_back([=] { work(); }); + Threads[0] = std::thread([this, ThreadCount, S] { + for (unsigned I = 1; I < ThreadCount; ++I) { + Threads.emplace_back([=] { work(S, I); }); if (Stop) break; } ThreadsCreated.set_value(); - work(); + work(S, 0); }); } @@ -90,7 +91,8 @@ public: } private: - void work() { + void work(ThreadPoolStrategy S, unsigned ThreadID) { + S.apply_thread_strategy(ThreadID); while (true) { std::unique_lock Lock(Mutex); Cond.wait(Lock, [&] { return Stop || !WorkStack.empty(); }); diff --git a/llvm/lib/Support/ThreadPool.cpp b/llvm/lib/Support/ThreadPool.cpp index 40982d7..5aa5815 100644 --- a/llvm/lib/Support/ThreadPool.cpp +++ b/llvm/lib/Support/ThreadPool.cpp @@ -20,16 +20,15 @@ using namespace llvm; #if LLVM_ENABLE_THREADS -// Default to hardware_concurrency -ThreadPool::ThreadPool() : ThreadPool(hardware_concurrency()) {} - -ThreadPool::ThreadPool(unsigned ThreadCount) - : ActiveThreads(0), EnableFlag(true) { +ThreadPool::ThreadPool(ThreadPoolStrategy S) + : ActiveThreads(0), EnableFlag(true), + ThreadCount(S.compute_thread_count()) { // Create ThreadCount threads that will loop forever, wait on QueueCondition // for tasks to be queued or the Pool to be destroyed. Threads.reserve(ThreadCount); for (unsigned ThreadID = 0; ThreadID < ThreadCount; ++ThreadID) { - Threads.emplace_back([&] { + Threads.emplace_back([S, ThreadID, this] { + S.apply_thread_strategy(ThreadID); while (true) { PackagedTaskTy Task; { @@ -108,12 +107,10 @@ ThreadPool::~ThreadPool() { #else // LLVM_ENABLE_THREADS Disabled -ThreadPool::ThreadPool() : ThreadPool(0) {} - // No threads are launched, issue a warning if ThreadCount is not 0 -ThreadPool::ThreadPool(unsigned ThreadCount) - : ActiveThreads(0) { - if (ThreadCount) { +ThreadPool::ThreadPool(ThreadPoolStrategy S) + : ActiveThreads(0), ThreadCount(S.compute_thread_count()) { + if (ThreadCount != 1) { errs() << "Warning: request a ThreadPool with " << ThreadCount << " threads, but LLVM_ENABLE_THREADS has been turned off\n"; } @@ -138,8 +135,6 @@ std::shared_future ThreadPool::asyncImpl(TaskTy Task) { return Future; } -ThreadPool::~ThreadPool() { - wait(); -} +ThreadPool::~ThreadPool() { wait(); } #endif diff --git a/llvm/lib/Support/Threading.cpp b/llvm/lib/Support/Threading.cpp index 48750cef..de5adad 100644 --- a/llvm/lib/Support/Threading.cpp +++ b/llvm/lib/Support/Threading.cpp @@ -45,10 +45,6 @@ void llvm::llvm_execute_on_thread(void (*Fn)(void *), void *UserData, Fn(UserData); } -unsigned llvm::heavyweight_hardware_concurrency() { return 1; } - -unsigned llvm::hardware_concurrency() { return 1; } - uint64_t llvm::get_threadid() { return 0; } uint32_t llvm::get_max_thread_name_length() { return 0; } @@ -57,6 +53,13 @@ void llvm::set_thread_name(const Twine &Name) {} void llvm::get_thread_name(SmallVectorImpl &Name) { Name.clear(); } +llvm::BitVector llvm::get_thread_affinity_mask() { return {}; } + +unsigned llvm::ThreadPoolStrategy::compute_thread_count() const { + // When threads are disabled, ensure clients will loop at least once. + return 1; +} + #if LLVM_ENABLE_THREADS == 0 void llvm::llvm_execute_on_thread_async( llvm::unique_function Func, @@ -78,30 +81,19 @@ void llvm::llvm_execute_on_thread_async( #else -#include -unsigned llvm::heavyweight_hardware_concurrency() { - // Since we can't get here unless LLVM_ENABLE_THREADS == 1, it is safe to use - // `std::thread` directly instead of `llvm::thread` (and indeed, doing so - // allows us to not define `thread` in the llvm namespace, which conflicts - // with some platforms such as FreeBSD whose headers also define a struct - // called `thread` in the global namespace which can cause ambiguity due to - // ADL. - int NumPhysical = sys::getHostNumPhysicalCores(); - if (NumPhysical == -1) - return std::thread::hardware_concurrency(); - return NumPhysical; -} +int computeHostNumHardwareThreads(); -unsigned llvm::hardware_concurrency() { -#if defined(HAVE_SCHED_GETAFFINITY) && defined(HAVE_CPU_COUNT) - cpu_set_t Set; - if (sched_getaffinity(0, sizeof(Set), &Set)) - return CPU_COUNT(&Set); -#endif - // Guard against std::thread::hardware_concurrency() returning 0. - if (unsigned Val = std::thread::hardware_concurrency()) - return Val; - return 1; +unsigned llvm::ThreadPoolStrategy::compute_thread_count() const { + int MaxThreadCount = UseHyperThreads ? computeHostNumHardwareThreads() + : sys::getHostNumPhysicalCores(); + if (MaxThreadCount <= 0) + MaxThreadCount = 1; + + // No need to create more threads than there are hardware threads, it would + // uselessly induce more context-switching and cache eviction. + if (!ThreadsRequested || ThreadsRequested > (unsigned)MaxThreadCount) + return MaxThreadCount; + return ThreadsRequested; } namespace { diff --git a/llvm/lib/Support/Unix/Threading.inc b/llvm/lib/Support/Unix/Threading.inc index afb887f..8cacaa8 100644 --- a/llvm/lib/Support/Unix/Threading.inc +++ b/llvm/lib/Support/Unix/Threading.inc @@ -267,3 +267,27 @@ SetThreadPriorityResult llvm::set_thread_priority(ThreadPriority Priority) { #endif return SetThreadPriorityResult::FAILURE; } + +#include + +int computeHostNumHardwareThreads() { +#if defined(HAVE_SCHED_GETAFFINITY) && defined(HAVE_CPU_COUNT) + cpu_set_t Set; + if (sched_getaffinity(0, sizeof(Set), &Set)) + return CPU_COUNT(&Set); +#endif + // Guard against std::thread::hardware_concurrency() returning 0. + if (unsigned Val = std::thread::hardware_concurrency()) + return Val; + return 1; +} + +void llvm::ThreadPoolStrategy::apply_thread_strategy( + unsigned ThreadPoolNum) const {} + +llvm::BitVector llvm::get_thread_affinity_mask() { + // FIXME: Implement + llvm_unreachable("Not implemented!"); +} + +unsigned llvm::get_cpus() { return 1; } diff --git a/llvm/lib/Support/Windows/Threading.inc b/llvm/lib/Support/Windows/Threading.inc index 9456efa..eb92296 100644 --- a/llvm/lib/Support/Windows/Threading.inc +++ b/llvm/lib/Support/Windows/Threading.inc @@ -16,6 +16,8 @@ #include "WindowsSupport.h" #include +#include + // Windows will at times define MemoryFence. #ifdef MemoryFence #undef MemoryFence @@ -122,3 +124,163 @@ SetThreadPriorityResult llvm::set_thread_priority(ThreadPriority Priority) { ? SetThreadPriorityResult::SUCCESS : SetThreadPriorityResult::FAILURE; } + +struct ProcessorGroup { + unsigned ID; + unsigned AllThreads; + unsigned UsableThreads; + unsigned ThreadsPerCore; + uint64_t Affinity; +}; + +template +static bool IterateProcInfo(LOGICAL_PROCESSOR_RELATIONSHIP Relationship, F Fn) { + DWORD Len = 0; + BOOL R = ::GetLogicalProcessorInformationEx(Relationship, NULL, &Len); + if (R || GetLastError() != ERROR_INSUFFICIENT_BUFFER) { + return false; + } + auto *Info = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *)calloc(1, Len); + R = ::GetLogicalProcessorInformationEx(Relationship, Info, &Len); + if (R) { + auto *End = + (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *)((uint8_t *)Info + Len); + for (auto *Curr = Info; Curr < End; + Curr = (SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *)((uint8_t *)Curr + + Curr->Size)) { + if (Curr->Relationship != Relationship) + continue; + Fn(Curr); + } + } + free(Info); + return true; +} + +static ArrayRef getProcessorGroups() { + auto computeGroups = []() { + SmallVector Groups; + + auto HandleGroup = [&](SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *ProcInfo) { + GROUP_RELATIONSHIP &El = ProcInfo->Group; + for (unsigned J = 0; J < El.ActiveGroupCount; ++J) { + ProcessorGroup G; + G.ID = Groups.size(); + G.AllThreads = El.GroupInfo[J].MaximumProcessorCount; + G.UsableThreads = El.GroupInfo[J].ActiveProcessorCount; + assert(G.UsableThreads <= 64); + G.Affinity = El.GroupInfo[J].ActiveProcessorMask; + Groups.push_back(G); + } + }; + + if (!IterateProcInfo(RelationGroup, HandleGroup)) + return std::vector(); + + auto HandleProc = [&](SYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX *ProcInfo) { + PROCESSOR_RELATIONSHIP &El = ProcInfo->Processor; + assert(El.GroupCount == 1); + unsigned NumHyperThreads = 1; + // If the flag is set, each core supports more than one hyper-thread. + if (El.Flags & LTP_PC_SMT) + NumHyperThreads = std::bitset<64>(El.GroupMask[0].Mask).count(); + unsigned I = El.GroupMask[0].Group; + Groups[I].ThreadsPerCore = NumHyperThreads; + }; + + if (!IterateProcInfo(RelationProcessorCore, HandleProc)) + return std::vector(); + + // If there's an affinity mask set on one of the CPUs, then assume the user + // wants to constrain the current process to only a single CPU. + for (auto &G : Groups) { + if (G.UsableThreads != G.AllThreads) { + ProcessorGroup NewG{G}; + Groups.clear(); + Groups.push_back(NewG); + break; + } + } + + return std::vector(Groups.begin(), Groups.end()); + }; + static auto Groups = computeGroups(); + return ArrayRef(Groups); +} + +template +static unsigned aggregate(R &&Range, UnaryPredicate P) { + unsigned I{}; + for (const auto &It : Range) + I += P(It); + return I; +} + +// for sys::getHostNumPhysicalCores +int computeHostNumPhysicalCores() { + static unsigned Cores = + aggregate(getProcessorGroups(), [](const ProcessorGroup &G) { + return G.UsableThreads / G.ThreadsPerCore; + }); + return Cores; +} + +int computeHostNumHardwareThreads() { + static unsigned Threads = + aggregate(getProcessorGroups(), + [](const ProcessorGroup &G) { return G.UsableThreads; }); + return Threads; +} + +// Assign the current thread to a more appropriate CPU socket or CPU group +void llvm::ThreadPoolStrategy::apply_thread_strategy( + unsigned ThreadPoolNum) const { + ArrayRef Groups = getProcessorGroups(); + + assert(ThreadPoolNum < compute_thread_count() && + "The thread index is not within thread strategy's range!"); + + // In this mode, the ThreadNumber represents the core number, not the + // hyper-thread number. Assumes all NUMA groups have the same amount of + // hyper-threads. + if (!UseHyperThreads) + ThreadPoolNum *= Groups[0].ThreadsPerCore; + + unsigned ThreadRangeStart = 0; + for (unsigned I = 0; I < Groups.size(); ++I) { + const ProcessorGroup &G = Groups[I]; + if (ThreadPoolNum >= ThreadRangeStart && + ThreadPoolNum < ThreadRangeStart + G.UsableThreads) { + + GROUP_AFFINITY Affinity{}; + Affinity.Group = G.ID; + Affinity.Mask = G.Affinity; + SetThreadGroupAffinity(GetCurrentThread(), &Affinity, nullptr); + } + ThreadRangeStart += G.UsableThreads; + } +} + +llvm::BitVector llvm::get_thread_affinity_mask() { + GROUP_AFFINITY Affinity{}; + GetThreadGroupAffinity(GetCurrentThread(), &Affinity); + + static unsigned All = + aggregate(getProcessorGroups(), + [](const ProcessorGroup &G) { return G.AllThreads; }); + + unsigned StartOffset = + aggregate(getProcessorGroups(), [&](const ProcessorGroup &G) { + return G.ID < Affinity.Group ? G.AllThreads : 0; + }); + + llvm::BitVector V; + V.resize(All); + for (unsigned I = 0; I < sizeof(KAFFINITY) * 8; ++I) { + if ((Affinity.Mask >> I) & 1) + V.set(StartOffset + I); + } + return V; +} + +unsigned llvm::get_cpus() { return getProcessorGroups().size(); } diff --git a/llvm/tools/dsymutil/dsymutil.cpp b/llvm/tools/dsymutil/dsymutil.cpp index 6b26081..54adeaa 100644 --- a/llvm/tools/dsymutil/dsymutil.cpp +++ b/llvm/tools/dsymutil/dsymutil.cpp @@ -258,7 +258,7 @@ static Expected getOptions(opt::InputArgList &Args) { if (opt::Arg *NumThreads = Args.getLastArg(OPT_threads)) Options.LinkOpts.Threads = atoi(NumThreads->getValue()); else - Options.LinkOpts.Threads = thread::hardware_concurrency(); + Options.LinkOpts.Threads = 0; // Use all available hardware threads if (Options.DumpDebugMap || Options.LinkOpts.Verbose) Options.LinkOpts.Threads = 1; @@ -541,9 +541,10 @@ int main(int argc, char **argv) { // Shared a single binary holder for all the link steps. BinaryHolder BinHolder; - unsigned ThreadCount = - std::min(Options.LinkOpts.Threads, DebugMapPtrsOrErr->size()); - ThreadPool Threads(ThreadCount); + unsigned ThreadCount = Options.LinkOpts.Threads; + if (!ThreadCount) + ThreadCount = DebugMapPtrsOrErr->size(); + ThreadPool Threads(hardware_concurrency(ThreadCount)); // If there is more than one link to execute, we need to generate // temporary files. diff --git a/llvm/tools/gold/gold-plugin.cpp b/llvm/tools/gold/gold-plugin.cpp index f68f718..2a9398f 100644 --- a/llvm/tools/gold/gold-plugin.cpp +++ b/llvm/tools/gold/gold-plugin.cpp @@ -134,8 +134,8 @@ namespace options { static unsigned OptLevel = 2; // Default parallelism of 0 used to indicate that user did not specify. // Actual parallelism default value depends on implementation. - // Currently only affects ThinLTO, where the default is - // llvm::heavyweight_hardware_concurrency. + // Currently only affects ThinLTO, where the default is the max cores in the + // system. static unsigned Parallelism = 0; // Default regular LTO codegen parallelism (number of partitions). static unsigned ParallelCodeGenParallelismLevel = 1; diff --git a/llvm/tools/llvm-cov/CodeCoverage.cpp b/llvm/tools/llvm-cov/CodeCoverage.cpp index 52e9958..625e234 100644 --- a/llvm/tools/llvm-cov/CodeCoverage.cpp +++ b/llvm/tools/llvm-cov/CodeCoverage.cpp @@ -947,9 +947,7 @@ int CodeCoverageTool::doShow(int argc, const char **argv, // If NumThreads is not specified, auto-detect a good default. if (NumThreads == 0) - NumThreads = - std::max(1U, std::min(llvm::heavyweight_hardware_concurrency(), - unsigned(SourceFiles.size()))); + NumThreads = SourceFiles.size(); if (!ViewOpts.hasOutputDirectory() || NumThreads == 1) { for (const std::string &SourceFile : SourceFiles) @@ -957,7 +955,7 @@ int CodeCoverageTool::doShow(int argc, const char **argv, ShowFilenames); } else { // In -output-dir mode, it's safe to use multiple threads to print files. - ThreadPool Pool(NumThreads); + ThreadPool Pool(heavyweight_hardware_concurrency(NumThreads)); for (const std::string &SourceFile : SourceFiles) Pool.async(&CodeCoverageTool::writeSourceFileView, this, SourceFile, Coverage.get(), Printer.get(), ShowFilenames); diff --git a/llvm/tools/llvm-cov/CoverageExporterJson.cpp b/llvm/tools/llvm-cov/CoverageExporterJson.cpp index 216b5e3..ba8ff5c 100644 --- a/llvm/tools/llvm-cov/CoverageExporterJson.cpp +++ b/llvm/tools/llvm-cov/CoverageExporterJson.cpp @@ -163,11 +163,9 @@ json::Array renderFiles(const coverage::CoverageMapping &Coverage, ArrayRef FileReports, const CoverageViewOptions &Options) { auto NumThreads = Options.NumThreads; - if (NumThreads == 0) { - NumThreads = std::max(1U, std::min(llvm::heavyweight_hardware_concurrency(), - unsigned(SourceFiles.size()))); - } - ThreadPool Pool(NumThreads); + if (NumThreads == 0) + NumThreads = SourceFiles.size(); + ThreadPool Pool(heavyweight_hardware_concurrency(NumThreads)); json::Array FileArray; std::mutex FileArrayMutex; diff --git a/llvm/tools/llvm-cov/CoverageReport.cpp b/llvm/tools/llvm-cov/CoverageReport.cpp index 8225954..187e2dc 100644 --- a/llvm/tools/llvm-cov/CoverageReport.cpp +++ b/llvm/tools/llvm-cov/CoverageReport.cpp @@ -356,11 +356,8 @@ std::vector CoverageReport::prepareFileReports( // If NumThreads is not specified, auto-detect a good default. if (NumThreads == 0) - NumThreads = - std::max(1U, std::min(llvm::heavyweight_hardware_concurrency(), - unsigned(Files.size()))); - - ThreadPool Pool(NumThreads); + NumThreads = Files.size(); + ThreadPool Pool(heavyweight_hardware_concurrency(NumThreads)); std::vector FileReports; FileReports.reserve(Files.size()); diff --git a/llvm/tools/llvm-lto2/llvm-lto2.cpp b/llvm/tools/llvm-lto2/llvm-lto2.cpp index 8deedd4..fc86fd9 100644 --- a/llvm/tools/llvm-lto2/llvm-lto2.cpp +++ b/llvm/tools/llvm-lto2/llvm-lto2.cpp @@ -65,8 +65,8 @@ static cl::opt "import files for the " "distributed backend case")); -static cl::opt Threads("thinlto-threads", - cl::init(llvm::heavyweight_hardware_concurrency())); +// Default to using all hardware cores in the system. +static cl::opt Threads("thinlto-threads", cl::init(0)); static cl::list SymbolResolutions( "r", diff --git a/llvm/tools/llvm-profdata/llvm-profdata.cpp b/llvm/tools/llvm-profdata/llvm-profdata.cpp index 424edf4..f05c7e6 100644 --- a/llvm/tools/llvm-profdata/llvm-profdata.cpp +++ b/llvm/tools/llvm-profdata/llvm-profdata.cpp @@ -307,8 +307,11 @@ static void mergeInstrProfile(const WeightedFileVector &Inputs, // If NumThreads is not specified, auto-detect a good default. if (NumThreads == 0) - NumThreads = - std::min(hardware_concurrency(), unsigned((Inputs.size() + 1) / 2)); + NumThreads = std::min(hardware_concurrency().compute_thread_count(), + unsigned((Inputs.size() + 1) / 2)); + // FIXME: There's a bug here, where setting NumThreads = Inputs.size() fails + // the merge_empty_profile.test because the InstrProfWriter.ProfileKind isn't + // merged, thus the emitted file ends up with a PF_Unknown kind. // Initialize the writer contexts. SmallVector, 4> Contexts; @@ -320,7 +323,7 @@ static void mergeInstrProfile(const WeightedFileVector &Inputs, for (const auto &Input : Inputs) loadInput(Input, Remapper, Contexts[0].get()); } else { - ThreadPool Pool(NumThreads); + ThreadPool Pool(hardware_concurrency(NumThreads)); // Load the inputs in parallel (N/NumThreads serial steps). unsigned Ctx = 0; diff --git a/llvm/unittests/Support/Host.cpp b/llvm/unittests/Support/Host.cpp index 2c17a50..6225234 100644 --- a/llvm/unittests/Support/Host.cpp +++ b/llvm/unittests/Support/Host.cpp @@ -37,7 +37,8 @@ protected: // Initially this is only testing detection of the number of // physical cores, which is currently only supported/tested for // x86_64 Linux and Darwin. - return (Host.getArch() == Triple::x86_64 && + return Host.isOSWindows() || + (Host.getArch() == Triple::x86_64 && (Host.isOSDarwin() || Host.getOS() == Triple::Linux)); } diff --git a/llvm/unittests/Support/TaskQueueTest.cpp b/llvm/unittests/Support/TaskQueueTest.cpp index 0a8aeca..4d8c3e4 100644 --- a/llvm/unittests/Support/TaskQueueTest.cpp +++ b/llvm/unittests/Support/TaskQueueTest.cpp @@ -22,7 +22,7 @@ protected: }; TEST_F(TaskQueueTest, OrderedFutures) { - ThreadPool TP(1); + ThreadPool TP(hardware_concurrency(1)); TaskQueue TQ(TP); std::atomic X{ 0 }; std::atomic Y{ 0 }; @@ -66,7 +66,7 @@ TEST_F(TaskQueueTest, OrderedFutures) { } TEST_F(TaskQueueTest, UnOrderedFutures) { - ThreadPool TP(1); + ThreadPool TP(hardware_concurrency(1)); TaskQueue TQ(TP); std::atomic X{ 0 }; std::atomic Y{ 0 }; @@ -96,7 +96,7 @@ TEST_F(TaskQueueTest, UnOrderedFutures) { } TEST_F(TaskQueueTest, FutureWithReturnValue) { - ThreadPool TP(1); + ThreadPool TP(hardware_concurrency(1)); TaskQueue TQ(TP); std::future F1 = TQ.async([&] { return std::string("Hello"); }); std::future F2 = TQ.async([&] { return 42; }); diff --git a/llvm/unittests/Support/ThreadPool.cpp b/llvm/unittests/Support/ThreadPool.cpp index a16adbb..237be87 100644 --- a/llvm/unittests/Support/ThreadPool.cpp +++ b/llvm/unittests/Support/ThreadPool.cpp @@ -8,11 +8,13 @@ #include "llvm/Support/ThreadPool.h" +#include "llvm/ADT/DenseSet.h" #include "llvm/ADT/STLExtras.h" #include "llvm/ADT/SmallVector.h" #include "llvm/ADT/Triple.h" #include "llvm/Support/Host.h" #include "llvm/Support/TargetSelect.h" +#include "llvm/Support/Threading.h" #include "gtest/gtest.h" @@ -69,6 +71,8 @@ protected: void SetUp() override { MainThreadReady = false; } + void TestAllThreads(ThreadPoolStrategy S); + std::condition_variable WaitMainThread; std::mutex WaitMainThreadMutex; bool MainThreadReady = false; @@ -131,7 +135,7 @@ TEST_F(ThreadPoolTest, Async) { TEST_F(ThreadPoolTest, GetFuture) { CHECK_UNSUPPORTED(); - ThreadPool Pool{2}; + ThreadPool Pool(hardware_concurrency(2)); std::atomic_int i{0}; Pool.async([this, &i] { waitForMainThread(); @@ -162,3 +166,45 @@ TEST_F(ThreadPoolTest, PoolDestruction) { } ASSERT_EQ(5, checked_in); } + +#if LLVM_ENABLE_THREADS == 1 + +void ThreadPoolTest::TestAllThreads(ThreadPoolStrategy S) { + // FIXME: Skip these tests on non-Windows because multi-socket system were not + // tested on Unix yet, and llvm::get_thread_affinity_mask() isn't implemented + // for Unix. + Triple Host(Triple::normalize(sys::getProcessTriple())); + if (!Host.isOSWindows()) + return; + + llvm::DenseSet ThreadsUsed; + std::mutex Lock; + unsigned Threads = 0; + { + ThreadPool Pool(S); + Threads = Pool.getThreadCount(); + for (size_t I = 0; I < 10000; ++I) { + Pool.async([&] { + waitForMainThread(); + std::lock_guard Guard(Lock); + auto Mask = llvm::get_thread_affinity_mask(); + ThreadsUsed.insert(Mask); + }); + } + ASSERT_EQ(true, ThreadsUsed.empty()); + setMainThreadReady(); + } + ASSERT_EQ(llvm::get_cpus(), ThreadsUsed.size()); +} + +TEST_F(ThreadPoolTest, AllThreads_UseAllRessources) { + CHECK_UNSUPPORTED(); + TestAllThreads({}); +} + +TEST_F(ThreadPoolTest, AllThreads_OneThreadPerCore) { + CHECK_UNSUPPORTED(); + TestAllThreads(llvm::heavyweight_hardware_concurrency()); +} + +#endif diff --git a/llvm/unittests/Support/Threading.cpp b/llvm/unittests/Support/Threading.cpp index 183c9aa..c76e6e4 100644 --- a/llvm/unittests/Support/Threading.cpp +++ b/llvm/unittests/Support/Threading.cpp @@ -21,7 +21,8 @@ TEST(Threading, PhysicalConcurrency) { auto Num = heavyweight_hardware_concurrency(); // Since Num is unsigned this will also catch us trying to // return -1. - ASSERT_LE(Num, thread::hardware_concurrency()); + ASSERT_LE(Num.compute_thread_count(), + hardware_concurrency().compute_thread_count()); } #if LLVM_ENABLE_THREADS diff --git a/mlir/lib/Pass/Pass.cpp b/mlir/lib/Pass/Pass.cpp index fc40ff1..41adb62 100644 --- a/mlir/lib/Pass/Pass.cpp +++ b/mlir/lib/Pass/Pass.cpp @@ -411,7 +411,8 @@ void OpToOpPassAdaptorParallel::runOnOperation() { // Create the async executors if they haven't been created, or if the main // pipeline has changed. if (asyncExecutors.empty() || hasSizeMismatch(asyncExecutors.front(), mgrs)) - asyncExecutors.assign(llvm::hardware_concurrency(), mgrs); + asyncExecutors.assign(llvm::hardware_concurrency().compute_thread_count(), + mgrs); // Run a prepass over the module to collect the operations to execute over. // This ensures that an analysis manager exists for each operation, as well as -- 2.7.4