- 545ebe9 Implement 6D parallelization with 1D and no tiling by Marat Dukhan · 3 years, 6 months ago
- 13d90b4 Use __STDC_NO_ATOMICS__ to detect C11 compilers without stdatomic.h by Marat Dukhan · 3 years, 6 months ago
- 6c76abe Support pre-C11 GCC intrinsics for atomics by Marat Dukhan · 3 years, 6 months ago
- fa75e65 Fix MSVC build (#10) by peterjc123 · 3 years, 8 months ago
- 18a7156 Use cpuinfo_get_current_uarch_index_with_default for parallelization with uarch by Marat Dukhan · 4 years ago
- bfe07ff 3D/4D/5D parallelization functions with 1D or no tiling by Marat Dukhan · 4 years ago
- afb880d Guard against generating ARM yield instruction for unsupporting processors by Marat Dukhan · 4 years, 1 month ago
- 5d088b1 Reorder C11 atomics before MSVC x64 atomics by Marat Dukhan · 4 years, 1 month ago
- 6ae95d3 Use platform-specific yield/pause instructions by Marat Dukhan · 4 years, 1 month ago
- 5690b5c MSVC-compatible FPU state functions by Marat Dukhan · 4 years, 1 month ago
- ada4eed Thumb-1 compatible assembly for disable_fpu_denormals by Marat Dukhan · 4 years, 1 month ago
- 6525d8b Avoid including stdatomic.h in any WAsm builds by Marat Dukhan · 4 years, 1 month ago
- 0672a7e Fast path using atomic decrement instead of atomic compare-and-swap by Marat Dukhan · 4 years, 1 month ago
- 6acde9c Reorder C11 atomics before MSVC atomics by Marat Dukhan · 4 years, 1 month ago
- e918b20 Recognize Cygwin as Windows by Marat Dukhan · 4 years, 2 months ago
- a61ed1a Use load-acquire + store-release on synchronization variables by Marat Dukhan · 4 years, 2 months ago
- 4ac9270 Support Windows on ARM/ARM64 by Marat Dukhan · 4 years, 2 months ago
- ddd479b Replace atomic fetch_sub with decrement_fetch primitive by Marat Dukhan · 4 years, 2 months ago
- fa72613 Add compiler barriers to MSVC atomics implementation by Marat Dukhan · 4 years, 2 months ago
- 52a2f57 Fix race condition in Windows implementation by Marat Dukhan · 4 years, 2 months ago
- 8f8dd87 Rewrite work spreading between threads by Marat Dukhan · 4 years, 2 months ago
- cb89899 Direct implementation pthreadpool_try_decrement_relaxed_size_t by Marat Dukhan · 4 years, 2 months ago
- 7e8364c Return static thread pool pointer in shim implementation by Marat Dukhan · 4 years, 2 months ago
- 02b50e6 Minor fixes in Windows implementation by Marat Dukhan · 4 years, 2 months ago
- be1bd8e Windows implementation using Events by Marat Dukhan · 4 years, 2 months ago
- fa67ff5 Fix erroneous narrowing in pthreadpool_fetch_sub_relaxed_size_t by Marat Dukhan · 4 years, 2 months ago
- 4694c12 Optimized pthreadpool_parallelize_* functions by Marat Dukhan · 4 years, 2 months ago
- be9c893 Implementation using Grand Central Dispatch by Marat Dukhan · 4 years, 2 months ago
- fc793bc Refactor pthreadpool implementation by Marat Dukhan · 4 years, 2 months ago
- 5b41aa6 Remove unused per-thread wakeup_condvar by Marat Dukhan · 4 years, 2 months ago
- 7604215 Microarchitecture-aware parallelization functions by Marat Dukhan · 4 years, 2 months ago
- 6469659 Refactor multi-threaded case of parallelization functions by Marat Dukhan · 4 years, 2 months ago
- 832403f Implement atomic_decrement with LL-SC on ARM/ARM64 by Marat Dukhan · 4 years, 2 months ago
- 97c181b Minor refactoring in pthreadpool_destroy by Marat Dukhan · 4 years, 2 months ago
- ba8b08e Fix race conditions in non-futex implementation by Marat Dukhan · 4 years, 2 months ago
- 15f39bf Futex-based WebAssembly+Threads implementation by Marat Dukhan · 4 years, 2 months ago
- cb207d8 Support WebAssembly+Threads build by Marat Dukhan · 4 years, 2 months ago
- 0f57821 Remove redundant barriers by Marat Dukhan · 4 years, 2 months ago
- 75294ce Simplify parallel task initialization by Marat Dukhan · 4 years, 2 months ago
- 2bde094 Avoid spinning thread-pool when task has the only item by Marat Dukhan · 4 years, 2 months ago
- 39e0461 Remove Native Client support by Marat Dukhan · 4 years, 3 months ago
- ef23a4a PTHREADPOOL_FLAG_YIELD_WORKERS flag to bypass spin-wait by Marat Dukhan · 4 years, 3 months ago
- 00108cf Minor cleanup by Marat Dukhan · 4 years, 3 months ago
- ebd50d0 Build on Windows/mingw64 (#6) by mattn · 4 years, 3 months ago
- 71aacd3 Switch to C11 atomics to synchronization by Marat Dukhan · 4 years, 8 months ago
- 158098a Make inline assembly compatible with old toolchain by Marat Dukhan · 4 years, 8 months ago
- 9e2825c Fix typo in comment by Marat Dukhan · 4 years, 8 months ago
- 714c953 Enable spin-wait in the main thread by Marat Dukhan · 4 years, 8 months ago
- 29f0e2c New pthreadpool_parallelize_* API by Marat Dukhan · 4 years, 8 months ago
- afb23f9 Enable spin-wait in worker threads by Marat Dukhan · 4 years, 8 months ago
- e807f76 Refactor modulo increment in work-stealing loop by Marat Dukhan · 4 years, 8 months ago
- c06f293 Use caller thread as one of workers in the thread pool by Marat Dukhan · 4 years, 8 months ago
- e6be7fc Minor refactorings by Marat Dukhan · 4 years, 8 months ago
- 13da0b4 Support 3D and 4D tiled computation by Marat Dukhan · 6 years ago
- 7d77d72 pthreadpool_get_threads_count: support NULL threadpool by Marat Dukhan · 6 years ago
- 6615a1c Compatibility with old Android platforms in futex syscall by Marat Dukhan · 7 years ago
- 2da784d Use futex on Native Client by Marat Dukhan · 7 years ago
- d08554a Use futex on Linux by Marat Dukhan · 7 years ago
- d1d42da Shared command variable among all threads by Marat Dukhan · 7 years ago
- 38b24fe Remove redundant include by Marat Dukhan · 7 years ago
- d54565d Support Emscripten by Marat Dukhan · 7 years ago
- a69b9eb Check that there are non-finished threads before work-stealing by Marat Dukhan · 7 years ago
- 2cff4bb Minor refactoring by Marat Dukhan · 7 years ago
- eef99d4 Allow NULL threadpool in pthreadpool_destroy by Marat Dukhan · 7 years ago
- 630dfb6 Fix race condition by Marat Dukhan · 7 years ago
- 17747d7 Android: make the code work with pre-Android 4.2 platform by Marat Dukhan · 8 years ago
- f3c8d73 Faster implementation for NULL thread pool by Marat Dukhan · 8 years ago
- 1325d6e Use FXdiv for 2D loops by Marat Dukhan · 8 years ago
- fa98b4b Fix race condition in the work processing by Marat Dukhan · 9 years ago
- e76282f Tiled 1D loops by Marat Dukhan · 9 years ago
- ad0ca6a Implement 2D and tiled 2D compute functions by Marat Dukhan · 9 years ago
- a04943a Use __has_extension and __has_feature is a way compatible with old gcc by Marat Dukhan · 9 years ago
- af6468b Make PTHREADPOOL_STATIC_ASSERT compatible with old compilers by Marat Dukhan · 9 years ago
- 7b1f6e5 Return size_t in pthreadpool_get_threads_count by Marat Dukhan · 9 years ago
- 3a45d9a Use posix_memalign instead of memalign of OS X compatibility by Marat Dukhan · 9 years ago
- c058bd3 Fixed typo: use __uint128_t instead of uint128_t by Marat Dukhan · 9 years ago
- 0a31219 Initial thread pool implementation by Marat Dukhan · 9 years ago