- be760b6 Simplify quantized multiplier by Georgios Pinitas · 3 years, 4 months ago
- 287015c Update test tolerance ahead of merging PR #227 by bjacob · 3 years, 4 months ago
- 58e3051 Change the default MulParams multiplier values to multiply by 1, not 0. by bjacob · 3 years, 5 months ago
- d65bcd7 Fixes for builds in open source projects with cpuinfo and googletest deps. by Benoit Jacob · 3 years, 5 months ago
- 45a876f Revert "Revert "Add CMake support with a converter from Bazel"" by bjacob · 3 years, 5 months ago
- 25df4d3 Corrected macro for detecting ppc platform (#83) by Nishidha · 3 years, 5 months ago
- 20b5eb0 Add a tracing framework (really just logging). by Benoit Jacob · 3 years, 5 months ago
- 4ed6216 Revert "Add CMake support with a converter from Bazel (#233)" (#243) by bjacob · 3 years, 5 months ago
- b87d6d2 Add CMake support with a converter from Bazel (#233) by bjacob · 3 years, 5 months ago
- fb9174e Corrected macro for detecting ppc platform. (#83) by Nishidha · 3 years, 5 months ago
- 2cbb179 Drop unneeded dependency from :context. by Benoit Jacob · 3 years, 5 months ago
- 3f655fa Cosmetics: class-ify TrMulTask, in particular put the trailing _ where they belong. by Benoit Jacob · 3 years, 5 months ago
- c162e5d Relax test tolerance against Eigen, adapting to a recent Eigen change between Eigen commits by Benoit Jacob · 3 years, 6 months ago
- 3fc7ae2 fix gcc warnings by Benoit Jacob · 3 years, 6 months ago
- 177062d Move the example out of the ruy/ruy directory, and add an example returning raw by Benoit Jacob · 3 years, 6 months ago
- 4790797 Fixing warnings on MSVC (comparing a bool with >). by Ben Vanik · 3 years, 7 months ago
- 7a6a38e Enforce x86 bit exactness by T.J. Alumbaugh · 3 years, 7 months ago
- d79362c MSVC fixes: by Benoit Jacob · 3 years, 7 months ago
- 7e1d379 Zero point checking disabled for uint8 x uint8 GEMMs by T.J. Alumbaugh · 3 years, 7 months ago
- dd1102a Update AVX, AVX2, AVX512 Rescale operations with Rounding Right Shift by T.J. Alumbaugh · 3 years, 8 months ago
- a28320a move example.cc into one directory by Leslie-Fang · 3 years, 8 months ago
- 034c0e2 Use movi NEON instruction to zero out registers by Lukas Geiger · 3 years, 8 months ago
- e59c55d It's _MSC_VER not __MSC_VER. by Benoit Jacob · 3 years, 8 months ago
- 503dd78 Enable x86 SIMD code paths on MSVC 2019 and similarly-versioned Clang-CL. by Benoit Jacob · 3 years, 8 months ago
- 3c363dc Add a few PMU counters. by Benoit Jacob · 3 years, 8 months ago
- 14569d2 Additional optimizations for AVX 8bit quantized kernel. by T.J. Alumbaugh · 3 years, 9 months ago
- fad2140 Optimize AVX/AVX2 quantized path by T.J. Alumbaugh · 3 years, 9 months ago
- d13c696 Fix buffer overrun on asan for AVX512 float. by T.J. Alumbaugh · 3 years, 9 months ago
- 8f08903 Optimize AVX512 float path by T.J. Alumbaugh · 3 years, 9 months ago
- be065e4 Optimize AVX/AVX2+FMA float path by T.J. Alumbaugh · 3 years, 9 months ago
- d7b739e AVX 8bit row major/col major packing code by T.J. Alumbaugh · 3 years, 9 months ago
- 74bfa70 AVX Pack inherits from StandardCpp by T.J. Alumbaugh · 3 years, 9 months ago
- 9e63749 AVX 8bit kernel. Forked from AVX2+FMA version by T.J. Alumbaugh · 3 years, 9 months ago
- ce0e559 Changes are excluded via Copybara by Ruy Contributors · 3 years, 10 months ago
- 4b1972b Changes are excluded via Copybara by Ruy Contributors · 3 years, 10 months ago
- 59c2de8 Rename kOutOfOrder -> kGeneric, kInOrder -> kA55ish, by Benoit Jacob · 3 years, 10 months ago
- 4f6a37b Reimplement :tune on top of :cpuinfo. by Benoit Jacob · 3 years, 10 months ago
- f99b42b Add bzl_library rules for .bzl files without one. by Ruy Contributors · 3 years, 10 months ago
- 2b24016 Adds AVX float packing code. by T.J. Alumbaugh · 3 years, 10 months ago
- 70d32d6 Adds AVX path and AVX float kernel. by T.J. Alumbaugh · 3 years, 10 months ago
- d4822f4 Adds AVX path and AVX float kernel. by T.J. Alumbaugh · 3 years, 10 months ago
- 18e34fa Adds AVX path and AVX float kernel. by T.J. Alumbaugh · 3 years, 10 months ago
- d7bd2a1 Print extra information in case of disagreeing TestResults. by T.J. Alumbaugh · 3 years, 11 months ago
- 5bb02fb check_macros improvements: promote operands before comparisons (avoids -Wsign-compare errors with GCC in cases like RUY_CHECK_NE(unsigned_bitmask_expression, 0)) and move all of the implementation to an inline function instead of having half of it in the macro. by Benoit Jacob · 3 years, 11 months ago
- f876353 Add missing #include of <cstring>. by Benoit Jacob · 3 years, 11 months ago
- bfe6e0d Simplify bias-loading code now that bias buffers are always rounded up to multiple of kernel size. by Benoit Jacob · 3 years, 11 months ago
- b53312b Use lambdas to shorten source code like we did in the avx512 kernel. by Benoit Jacob · 3 years, 11 months ago
- f611892 Handle per-column multipliers in the avx512 kernel without transposing the 16x16 accumulator block. by Benoit Jacob · 3 years, 11 months ago
- 1efd970 Optimized packing code path for row-major 8bit inputs for the x86 paths. by Benoit Jacob · 3 years, 11 months ago
- 257a0fc Optimized packing code path for row-major 8bit inputs for the kNeon path. Written in intrinsics to handle 3 cases at once: by Benoit Jacob · 3 years, 11 months ago
- 550655f Use lambdas to shorten Kernel8bitAvx512's source code, and to split the resulting non-opt binary code into smaller functions. This makes no difference in opt builds, but for non-opt builds this reduces the stack frame of this function from 60k down to 24k. This avoids stack overflows in some toolchains. by Benoit Jacob · 3 years, 11 months ago
- ec99c70 Optimized packing code path for row-major float inputs. by Benoit Jacob · 3 years, 11 months ago
- bebf022 Optimized packing code path for row-major 8bit inputs for the kNeonDotprod path. by Benoit Jacob · 3 years, 11 months ago
- d492ac8 Fix the build on some toolchains - a missing #include<cstring> and some avx512 intrinsic synonyms. by Benoit Jacob · 3 years, 11 months ago
- 90f7274 Rename packing code implementation functions now that they are explicitly about one specific source matrix storage order. by Benoit Jacob · 3 years, 11 months ago
- cd375d3 Templatize packing code paths on the source order, so that we support any combination source order, with the worst case being a fall back to the standard c++ packing code, which readily supports any storage order. by Benoit Jacob · 3 years, 11 months ago
- 5210e3e Simplification of FallBackToStandardCpp now that we are past the incremental steps toward supporting any channel_dimension. by Benoit Jacob · 3 years, 11 months ago
- 6d218c3 Efficient support for any channel_dimension for quantized kernels on AVX-512, part 2: handling of per-channel multipliers. by Benoit Jacob · 3 years, 11 months ago
- c1d5b4f Efficient support for any channel_dimension for quantized kernels on AVX-512, part 1: non-per-channel-multiplier case, so we only have to deal with bias vectors for now. by Benoit Jacob · 3 years, 11 months ago
- bb9349c Efficient support for any channel_dimension for quantized kernels on AVX2. by Benoit Jacob · 3 years, 11 months ago
- bd21e0c Simplify x86 kernels by using the fact that there always is a per-channel buffer to read from, even in the non-perchannel case (in that case, its size is just the kernel's width and one must use 0 as offset). by Benoit Jacob · 3 years, 11 months ago
- 98c5213 Simplify x86 kernels thanks to the new fact that perchannel buffers are rounded to next multiple of kernel width. by Benoit Jacob · 3 years, 11 months ago
- a776b5d Fix runtime detection of support for our AVX2+FMA code path: we were only checking for AVX2, which happens to imply FMA on Intel CPUs. by Benoit Jacob · 4 years ago
- 7784e18 FMA is technically a separate ISA extension from AVX2. by Benoit Jacob · 4 years ago
- 27d16d0 Efficient support for any channel_dimension for float kernels on AVX-512. by Benoit Jacob · 4 years ago
- 592d30c Efficient support for any channel_dimension for float kernels on AVX2. by Benoit Jacob · 4 years ago
- f88e08e Allow the user to specify that they have allocated a slightly larger capacity for the per-channel buffers, so that ruy can then avoid reallocating and copying these buffers. by Benoit Jacob · 4 years ago
- 388ffd2 Fix ARM32 packing code reading past the end of the source matrix, and finishing enabling the use of SeparateMappingVector in StorageMatrix in the test code to guard against that (It had discovered this issue). by Benoit Jacob · 4 years ago
- 856f0fd Add comments and some minor simplications to packing code. by Benoit Jacob · 4 years ago
- e600a4d Avoid overrunning per-channel buffers, whose size is that of the corresponding user-facing matrix dimension, but which assembly kernels tend to address as if they had the same size as the corresponding packed matrix dimension. AddressSanitizer can't see what asm kernels do. by Benoit Jacob · 4 years ago
- f5b52f9 Minor optimization of in-order arm64 kernels, interleave the dup's used in the channels-are-columns case with other instructions. by Benoit Jacob · 4 years ago
- 62aa923 Minor simplification of arm32 assembly: the add instruction itself can be conditional. by Benoit Jacob · 4 years ago
- ec970ca Efficient support for any channel_dimension for quantized kernels on ARM32. by Benoit Jacob · 4 years ago
- 53c5454 Efficient support for any channel_dimension for float kernels on ARM32. by Benoit Jacob · 4 years ago
- 3cacc71 Efficient support for any channel_dimension for kNeonDotprod quantized kernels on ARM64. by Benoit Jacob · 4 years ago
- ffb0866 Efficient support for any channel_dimension for kNeon quantized kernels on ARM64. by Benoit Jacob · 4 years ago
- 1f9e146 Ensure that the 1Col kernels are not used with channel_dimension==kCol, so that we don't need to update them. by Benoit Jacob · 4 years ago
- caf57cc Efficient support for any channel_dimension for float kernels on ARM64. by Benoit Jacob · 4 years ago
- df335bc Groundwork to pass channel_dimension down to kernels and to incrementally enable fast kernels in the channel_dimension==kCol case. by Benoit Jacob · 4 years ago
- cd4f776 Revisiting RUY_OPT(AVOID_ALIASING). by Benoit Jacob · 4 years ago
- b3edb05 Fix benchmarking of caching. by Benoit Jacob · 4 years ago
- 2d09352 Allow benchmarking any combination of storage orders, and disable the randomization of the channel_dimension in the case of benchmarking, so that the actual storage order of the destination matrix being benchmarked internally matches what is specified (no internal transposition). Randomization of the channel_dimension is kept in non-benchmark tests. by Benoit Jacob · 4 years ago
- c03ab18 Allow disabling the reference path in the benchmark. by Benoit Jacob · 4 years ago
- 33fa58e Remove RUY_OPT(NATIVE_ROUNDING) or rather, the ability to disable it. by Benoit Jacob · 4 years ago
- 8525a43 Make the reference/standard-cpp code in ApplyMultiplier match the ARM code, by changing the RoundingDivideByPOT function, which was borrowed from gemmlowp, to a RoundingRightShift function that is more like just a standard rounding arithmetic shift instruction, breaking ties upwards instead of away-from-zero. by Benoit Jacob · 4 years ago
- 7fb015f Avoid relying on std::max being constexpr, which is c++14 behavior but is not implemented on TensorFlow continuous integration on Ubuntu 16. by Benoit Jacob · 4 years ago
- e7f175f Remove ExpectedOutcome support, it was used for death tests in test_special_mul_params, which has been removed already. by Benoit Jacob · 4 years ago
- 03bbc8f Store perchannel members in a union with their non-perchannel counterpart. by Benoit Jacob · 4 years ago
- 39df743 Split the storage of MulParams data members into 3 separate template specializations for the floating-point, raw integer and quantized cases. by Benoit Jacob · 4 years ago
- f5e0fac Remove cpuinfo from s390x build as there is no support yet by cdavoudian · 4 years ago
- 8678f55 Reduce to the case of column-major destination matrix by transposing the whole Mul in the row-major destination matrix case. by Benoit Jacob · 4 years ago
- 5b496e0 Some refactoring in create_trmul_params.* ahead of implementing the transposition technique to reduce to column-major destination. by Benoit Jacob · 4 years ago
- c17ae28 Implement the channels_dimension==kCol case. by Benoit Jacob · 4 years ago
- 375895e Change Transpose functions to returning the result by value. by Benoit Jacob · 4 years ago
- e273e15 Store the MulParams by value, in a char[] buffer, in TrMulParams. by Benoit Jacob · 4 years ago
- fd803fb Add a channel_dimension member to MulParams, bringing the last piece to make Ruy's API fully LHS<->RHS symmetric, allowing the implementation to transpose the whole Mul to reduce to column major destination matrices. by Benoit Jacob · 4 years ago
- d2509b7 Make FixedKernelLayout internal by Benoit Jacob · 4 years ago
- 66961ae Fix up templates specialization for change by Ruy Contributors · 4 years ago
- 19b09a4 Make FixedKernelLayout internal by Robert David · 4 years ago
- c9f5f9c Clean up #includes and deps among kernel* and pack*. by Benoit Jacob · 4 years ago