- 45c7fd8 Avoid realloc for packed weights in ResizeBilinear operator by Marat Dukhan · 4 years, 6 months ago
- 787978d Eliminate __builtin_prefetch calls by Marat Dukhan · 4 years, 6 months ago
- 9c6c09b Use normal (unaligned) allocation for subconvolution buffer by Marat Dukhan · 4 years, 6 months ago
- f214916 Add a dummy variable in xnn_f32_default_params by Marat Dukhan · 4 years, 6 months ago
- 5ce30d9 Work around non-standard std::uniform_int_distribution<uint8_t> by Marat Dukhan · 4 years, 6 months ago
- 1b4b69d Specify "C" linkage for xnn_params structure by Marat Dukhan · 4 years, 6 months ago
- 02072e6 Avoid two more cases of arithmetics on void* by Marat Dukhan · 4 years, 6 months ago
- 57dccd8 NEON and SSE2 implementations of X32 UNPOOL micro-kernel by Marat Dukhan · 4 years, 6 months ago
- 68eef3f Update XNNPACK and cpuinfo build rules for armhf by Terry Heo · 4 years, 6 months ago
- d0cf9bd Replace VLAs with direct alloca calls by Marat Dukhan · 4 years, 6 months ago
- 0183625 Increase error tolerance in IBilinearMicrokernelTester by Marat Dukhan · 4 years, 6 months ago
- 1f4e461 F16 1x8 GEMM ld64 microkernel by Frank Barchard · 4 years, 6 months ago
- bcbae1e Minor fix in CMake configuration by Marat Dukhan · 4 years, 6 months ago
- 9f240d1 Fix typo in f32-vscaleextexp-test target in CMakeLists by Marat Dukhan · 4 years, 6 months ago
- 3fd4b29 Work around "too many sections" error in f32-igemm-minmax-test with MinGW by Marat Dukhan · 4 years, 6 months ago
- c2cfb97 Port AlignedAllocator to Windows by Marat Dukhan · 4 years, 6 months ago
- ef25e75 Use generator expression for Release/Debug flags by Marat Dukhan · 4 years, 6 months ago
- 6ae741a Specify Windows 7+ API target in CMake config by Marat Dukhan · 4 years, 6 months ago
- 666c271 Specify -msse2 for PSIMD micro-kernels by Marat Dukhan · 4 years, 6 months ago
- 0952f41 Update FP16 dependency by Marat Dukhan · 4 years, 6 months ago
- 8fc8776 Suppress type narrowing warning in X8-ZIP SSE2 micro-kernel by Marat Dukhan · 4 years, 6 months ago
- c5ee9ff Include missing <numeric> header in BinaryElementwiseOperatorTester by Marat Dukhan · 4 years, 6 months ago
- bdc8099 Avoid arithmetics on void* in indirection buffer setup by Marat Dukhan · 4 years, 6 months ago
- ef3e7dc Replace __builtin_lrintf with lrintf by Marat Dukhan · 4 years, 6 months ago
- 504f594 Port xnn_aligned_deallocate to Windows by Marat Dukhan · 4 years, 6 months ago
- 57133c0 Port xnn_initialize to Windows by Marat Dukhan · 4 years, 6 months ago
- 2acf108 Avoid arithetics on void* in NCHW Convolution by Marat Dukhan · 4 years, 6 months ago
- 8ac2b3a Include immintrin.h in sources using _mm_undefined_ps by Marat Dukhan · 4 years, 6 months ago
- b9d07cf Fix CMake build by Marat Dukhan · 4 years, 6 months ago
- 9fe932e Partial port to Windows and MSVC by Marat Dukhan · 4 years, 6 months ago
- 79eda12 Update pthreadpool and FXdiv dependencies by Marat Dukhan · 4 years, 6 months ago
- 9993660 Add MINMAX suffix to remaining micro-kernels by Marat Dukhan · 4 years, 6 months ago
- 36b76b6 1x16 LD32 F16 GEMM by Frank Barchard · 4 years, 6 months ago
- 3cb54f9 1x8 LD64 F32 GEMM by Frank Barchard · 4 years, 6 months ago
- 683f559 FP16 4x16 and 6x16 GEMM ld32 microkernels by Frank Barchard · 4 years, 6 months ago
- 91cd2b7 Rename binary operation micro-kernels by Marat Dukhan · 4 years, 6 months ago
- 355ab43 Rename SpMM micro-kernels by Marat Dukhan · 4 years, 6 months ago
- 869c62d Auto-switch to LINEAR GEMM/IGEMM/DWCONV micro-kernels by Marat Dukhan · 4 years, 6 months ago
- aefaef3 Prepare xnn_params for variations in fused activations by Marat Dukhan · 4 years, 6 months ago
- 163a7e6 Scalar & WAsm GEMM/IGEMM/DWCONV micro-kernels without activation by Marat Dukhan · 4 years, 6 months ago
- de06f49 Add MINMAX suffix to GEMM/IGEMM/DWCONV/PPMM micro-kernel names by Marat Dukhan · 4 years, 6 months ago
- 8452ff5 Refactor AVGPOOL micro-kernel parameters by Marat Dukhan · 4 years, 6 months ago
- 1c58711 Add MINMAX suffix to filenames of GEMM/IGEMM/PPMM/DWCONV micro-kernels by Marat Dukhan · 4 years, 6 months ago
- eb09a6b Rename F32/U8 output params to minmax params by Marat Dukhan · 4 years, 6 months ago
- a51cf48 Unify layout of min/max parameters by Marat Dukhan · 4 years, 6 months ago
- ca92c46 Update pthreadpool dependency by Marat Dukhan · 4 years, 6 months ago
- 1841b1a Avoid signbit in scalar PReLU micro-kernel by Marat Dukhan · 4 years, 6 months ago
- 0a7818b Merge pull request #443 from annxingyuan:master by XNNPACK Team · 4 years, 6 months ago
- 85da99c update build by Ann Yuan · 4 years, 6 months ago
- bf182e8 Remove redundant cpuinfo dependency in :operator_run by Marat Dukhan · 4 years, 6 months ago
- 71ff592 Update Android benchmarks by Marat Dukhan · 4 years, 6 months ago
- 0dc845c Update links to ML frameworks in README by Marat Dukhan · 4 years, 6 months ago
- 05702cf Dynamically choose micro-kernel depending on active core by Marat Dukhan · 4 years, 6 months ago
- bd18b51 Update cpuinfo dependency by Marat Dukhan · 4 years, 6 months ago
- bb0c2f8 Fix typo tf_http_archive -> http_archive by Marat Dukhan · 4 years, 6 months ago
- b038fdc Adapt XNNPACK to the move of ruy to its own GitHub repository. by Benoit Jacob · 4 years, 6 months ago
- d21fdcb Fix FP16 SpMM incompatibility with gcc by Marat Dukhan · 4 years, 6 months ago
- 7d3a8c3 Make e2e DWCONV benchmark compatible with older gcc by Marat Dukhan · 4 years, 6 months ago
- 91eb942 Support Bazel build for Linux/AArch64 by Marat Dukhan · 4 years, 6 months ago
- 0d1052c iOS 6x8 microkernel based on Cortex-A75 but with X18 avoided. by Frank Barchard · 4 years, 6 months ago
- 1b35463 Outline variables in BUILD files by Marat Dukhan · 4 years, 6 months ago
- 4195af5 Avoid push d8, d9 for xnn_f32_dwconv_ukernel_up4x9__aarch64_neonfma_cortex_a55. by Frank Barchard · 4 years, 6 months ago
- 6f8c966 Use x13 instead of x18. by Frank Barchard · 4 years, 6 months ago
- 479e78c Inline variables in BUILD files. by XNNPACK Team · 4 years, 7 months ago
- 1c8bc0c Fix bug in Average Pooling operator by Marat Dukhan · 4 years, 7 months ago
- c58bd34 Fix bug in dilated max pooling with padding by Marat Dukhan · 4 years, 7 months ago
- 5868d80 Automatically switch to GAVGPOOL micro-kernels in Average Pooling operator by Marat Dukhan · 4 years, 7 months ago
- b187835 Polyfill _cvtu32_mask16 for Apple Clang 10.x by Marat Dukhan · 4 years, 7 months ago
- 057e796 For IOS disable assembly, falling back on intrinsics. by Frank Barchard · 4 years, 7 months ago
- 8fb9055 4x8 GEMM and IGEMM microkernels for Cortex A55. 7.8% faster for e2e mobile net v2. by Frank Barchard · 4 years, 7 months ago
- 99103dc Improve compatibility with older gcc versions by Marat Dukhan · 4 years, 7 months ago
- 4313088 Fix CMake build by Marat Dukhan · 4 years, 7 months ago
- b3801eb Micro-kernel selection for newly detected microarchitectures by Marat Dukhan · 4 years, 7 months ago
- 36053aa 4x8 AARCH32 GEMM/IGEMM avoid r2 push/pop. by Frank Barchard · 4 years, 7 months ago
- 95f2032 Update cpuinfo.patch by Marat Dukhan · 4 years, 7 months ago
- b7dd29e 4x8 GEMM and IGEMM microkernels for AARCH32 Cortex A55. 11.5% faster end to end: by Frank Barchard · 4 years, 7 months ago
- d6a5463 Refactor Average Pooling setup by Marat Dukhan · 4 years, 7 months ago
- 7a16d8b Update Average Pooling operator benchmark by Marat Dukhan · 4 years, 7 months ago
- d9e92eb Fix AVX and AVX512F PReLU microkernels by Marat Dukhan · 4 years, 7 months ago
- 90eca0a AVX and AVX512F versions of PReLU micro-kernel by Marat Dukhan · 4 years, 7 months ago
- f32ae34 Unify the value of $ABC variable across all templates by Marat Dukhan · 4 years, 7 months ago
- 5c5fa96 Auto-generate CLAMP micro-kernels by Marat Dukhan · 4 years, 7 months ago
- a63a6fc Refactor GAVGPOOL micro-kernels by Marat Dukhan · 4 years, 7 months ago
- 660fd19 Rename BILINEAR microkernels into IBILINEAR by Marat Dukhan · 4 years, 7 months ago
- 20c3b92 Size test for Subgraph API by Marat Dukhan · 4 years, 7 months ago
- 64debce Cortex-A55 cpu detect and select 6x8 GEMM and IGEMM microkernels. 9% faster end to end on mobile net v2. 16.5% faster on mobile net v1. 6.4% faster on mobile net v3. by Frank Barchard · 4 years, 7 months ago
- fe7acb6 Targets for requantization tests and benchmarks by Marat Dukhan · 4 years, 7 months ago
- 91e1999 6x8 GEMM and IGEMM microkernels for Cortex A55. 9% faster end to end: by Frank Barchard · 4 years, 7 months ago
- 303ad00 Remove redundant dependency in NEON micro-kernels by Marat Dukhan · 4 years, 7 months ago
- 4245f43 Update size_test with newly added operators by Marat Dukhan · 4 years, 7 months ago
- 9fd7e25 Use cpuinfo_get_max_cache_size in bench-utils by Marat Dukhan · 4 years, 7 months ago
- b6862d1 Cleanup unused SoftMax benchmark presets by Marat Dukhan · 4 years, 7 months ago
- f42facc Completely remove NaCl and PNaCl support by Marat Dukhan · 4 years, 7 months ago
- 6206b4a Fix XNNPACK build on Linux/Android with Bazel by Marat Dukhan · 4 years, 7 months ago
- 8d3c693 Optional SoftMax benchmarks vs DNNL by Marat Dukhan · 4 years, 7 months ago
- cc51633 Update cpuinfo dependency by Marat Dukhan · 4 years, 7 months ago
- 49a5969 More descriptive entry point names for operators with groups by Marat Dukhan · 4 years, 7 months ago
- 892687b Fix typo in comment by Marat Dukhan · 4 years, 7 months ago
- f092a4a f32-maxpool microkernel for ARM Neon. by Frank Barchard · 4 years, 7 months ago
- 488c334 Update status of PyTorch integration in the readme by Marat Dukhan · 4 years, 7 months ago