Ecosyste.ms: Issues

An open API service for providing issue and pull request metadata for open source projects.

GitHub / bluss/matrixmultiply issues and pull requests

#86 - Add `clippy` to CI

Issue - State: closed - Opened by cospectrum 3 months ago - 1 comment

#85 - Fix alignment in s390x and cross test

Pull Request - State: closed - Opened by bluss 4 months ago

#84 - Update CI actions and tests with cargo careful's sanitizer

Pull Request - State: closed - Opened by bluss 9 months ago

#83 - Allow usage of the crate without alloc

Issue - State: open - Opened by stexa about 1 year ago

#82 - Transposed operations segfault

Issue - State: open - Opened by RylanYancey about 1 year ago - 1 comment

#81 - Give the user the ability to allocate all memory himself

Issue - State: open - Opened by cospectrum about 1 year ago - 4 comments

#80 - request 8-byte buffer alignment on macos

Pull Request - State: closed - Opened by bluss about 1 year ago - 1 comment

#79 - Bazel: link or target filename contains space

Issue - State: closed - Opened by QuentinPerez over 1 year ago - 2 comments

#78 - Remove space from file names

Pull Request - State: closed - Opened by xander-zitara over 1 year ago - 3 comments

#77 - Fix nostd build for complex

Pull Request - State: closed - Opened by bluss over 1 year ago

#76 - 0.3.5 fails to build with `--no-default-features --features cgemm`

Issue - State: closed - Opened by autarch over 1 year ago - 3 comments

#75 - Special packing for complex, specialize packing for avx2

Pull Request - State: closed - Opened by bluss over 1 year ago - 1 comment

#74 - Use slice in packing function for noalias optimization

Pull Request - State: closed - Opened by bluss over 1 year ago

#73 - Arm64/AArch64 Neon kernels

Pull Request - State: closed - Opened by bluss over 1 year ago

#72 - Request only 16-byte alignment on macos

Pull Request - State: closed - Opened by bluss over 1 year ago

#71 - Tiny matmul discrepancy between single- and multi- thread(?)

Issue - State: closed - Opened by kstavro over 1 year ago - 19 comments

#70 - Silence suspicious Send/Sync impls warning

Pull Request - State: closed - Opened by bluss over 2 years ago

#69 - Fix Send/Sync impls for future compat warning

Pull Request - State: closed - Opened by bluss over 2 years ago

#68 - Updated comment in function kernel_x86_avx

Pull Request - State: closed - Opened by Tastaturtaste over 2 years ago - 1 comment

#67 - Fix Miri error with -Zmiri-tag-raw-pointers

Pull Request - State: closed - Opened by jturner314 almost 3 years ago - 6 comments

#66 - Run CI on macos too

Pull Request - State: closed - Opened by bluss about 3 years ago

#65 - Run the benchmark loop script in ci

Pull Request - State: closed - Opened by bluss about 3 years ago

#64 - Fix usize parsing on 32-bit arch

Pull Request - State: closed - Opened by bluss about 3 years ago

#63 - Fix crates.io badge

Pull Request - State: closed - Opened by atouchet about 3 years ago - 1 comment

#62 - Run miri in the testuite

Pull Request - State: closed - Opened by bluss about 3 years ago

#61 - Links and other knowledge about applied matrix multiplication

Issue - State: open - Opened by bluss about 3 years ago

#60 - Run CI on macOS too

Pull Request - State: closed - Opened by bluss about 3 years ago - 1 comment

#59 - Allow tweaking size parameters at compile time

Pull Request - State: closed - Opened by bluss about 3 years ago - 1 comment

#58 - Add experimental support for complex: cgemm/zgemm

Pull Request - State: closed - Opened by bluss about 3 years ago

#57 - Apple M1 SIMD support

Issue - State: closed - Opened by KaedeTai about 3 years ago - 6 comments
Labels: enhancement, help wanted

#56 - Align mask buffer pointer manually

Pull Request - State: closed - Opened by bluss over 3 years ago - 1 comment

#55 - Test involving many 6x6 matrices fails randomly on Mac OS

Issue - State: closed - Opened by vihdzp over 3 years ago - 16 comments

#54 - Add benchmark runner as an "example" binary

Pull Request - State: closed - Opened by bluss almost 4 years ago

#53 - Add github actions to replace travis

Pull Request - State: closed - Opened by bluss almost 4 years ago - 1 comment

#52 - Add threading support

Pull Request - State: closed - Opened by bluss almost 4 years ago
Labels: breaking-change

#51 - no_std support

Pull Request - State: closed - Opened by vadixidav over 4 years ago - 17 comments

#50 - error: array lengths can't depend on generic parameters

Issue - State: closed - Opened by jonathanstrong almost 5 years ago - 5 comments

#49 - Fix test build for non-x86

Pull Request - State: closed - Opened by bluss about 5 years ago - 2 comments

#48 - Building tests fails on non-x86 architectures

Issue - State: closed - Opened by ignatenkobrain about 5 years ago - 8 comments

#47 - chore: Update itertools to 0.8

Pull Request - State: closed - Opened by ignatenkobrain about 5 years ago - 1 comment

#46 - (WIP) Add no_std support

Pull Request - State: closed - Opened by jturner314 about 5 years ago - 6 comments

#45 - Update rawpointer dependency to 0.2

Pull Request - State: closed - Opened by bluss about 5 years ago

#44 - Typo

Pull Request - State: closed - Opened by felixrabe over 5 years ago - 1 comment

#43 - Igemm experiment

Pull Request - State: open - Opened by SuperFluffy almost 6 years ago - 9 comments

#42 - Minor changes to kernel masking

Pull Request - State: closed - Opened by bluss almost 6 years ago

#41 - Only add in masked kernel loop

Pull Request - State: closed - Opened by SuperFluffy almost 6 years ago - 2 comments

#40 - Prepare 0.2.2

Pull Request - State: closed - Opened by bluss almost 6 years ago

#39 - Change and rename testing environment variable to `MMTEST_FEATURE=`

Pull Request - State: closed - Opened by bluss almost 6 years ago - 1 comment

#38 - Test fma instructions on travis

Issue - State: closed - Opened by SuperFluffy almost 6 years ago - 1 comment

#37 - More flexible kernel selection

Pull Request - State: closed - Opened by bluss almost 6 years ago - 3 comments

#36 - Implement sgemm and dgemm using fma

Pull Request - State: closed - Opened by SuperFluffy almost 6 years ago - 15 comments

#35 - Use fma, fused multiply add, for architectures supporting fma

Issue - State: closed - Opened by SuperFluffy almost 6 years ago - 4 comments

#34 - Use optimal kernel parameters (architectures, matrix layouts)

Issue - State: open - Opened by SuperFluffy almost 6 years ago - 7 comments

#33 - Implement DGEMM kernel using avx intrinsics

Pull Request - State: closed - Opened by SuperFluffy almost 6 years ago - 9 comments

#32 - Don't shadow c in sgemm_kernel::kernel_x86_avx

Issue - State: closed - Opened by SuperFluffy almost 6 years ago - 1 comment

#31 - Explore performance of _mm256_blend_ps vs _mm256_shuffle_ps

Issue - State: closed - Opened by SuperFluffy almost 6 years ago - 3 comments

#30 - Investigate if _mm256_broadcast_ss outperforms _mm256_set1_ps

Issue - State: closed - Opened by SuperFluffy almost 6 years ago - 3 comments

#29 - Panic when benchmarking with target-feature=sse

Issue - State: closed - Opened by SuperFluffy almost 6 years ago - 4 comments

#28 - WIP: i32 gemm experiment

Pull Request - State: open - Opened by bluss almost 6 years ago - 9 comments

#27 - In the sgemm avx kernel, transpose if we can match C's layout

Pull Request - State: closed - Opened by bluss almost 6 years ago

#26 - Speed up packing by using copy_nonoverlapping

Pull Request - State: closed - Opened by bluss almost 6 years ago - 1 comment

#25 - Allow operations on transposed matrices, i.e. Op(A) and Op(B), and DSYRK

Issue - State: open - Opened by SuperFluffy about 6 years ago - 8 comments

#24 - Integer matrices

Issue - State: open - Opened by SuperFluffy about 6 years ago - 9 comments

#23 - Use ifunc strategy or other ways to only check target feature existance once

Issue - State: open - Opened by bluss about 6 years ago - 1 comment

#22 - Use std::arch SIMD and runtime target feature detection

Pull Request - State: closed - Opened by bluss about 6 years ago - 4 comments

#21 - Fix handling of zero-size arrays

Pull Request - State: closed - Opened by jturner314 about 6 years ago - 3 comments

#20 - Relax debug assertion on strides of C matrix

Pull Request - State: closed - Opened by jturner314 about 6 years ago - 2 comments

#19 - Add .gitignore

Pull Request - State: closed - Opened by jturner314 about 6 years ago

#18 - ICE's on nightly rust: resolving bounds after type-checking

Issue - State: closed - Opened by bluss over 7 years ago - 1 comment

#17 - Use CARGO_CFG_TARGET_FEATURE to pick sgemm 8x8 if avx exists

Pull Request - State: closed - Opened by bluss almost 8 years ago

#16 - Use no local arrays

Pull Request - State: closed - Opened by bluss about 8 years ago

#15 - Improve unrolling in sgemm kernel

Pull Request - State: closed - Opened by bluss about 8 years ago - 2 comments

#14 - Revert the workaround for array zeroing

Issue - State: closed - Opened by bluss about 8 years ago - 2 comments

#13 - set up benchmarks to run on stable with cargo bench (Test)

Pull Request - State: closed - Opened by bluss over 8 years ago - 1 comment

#12 - Nozeroed (test)

Pull Request - State: closed - Opened by bluss over 8 years ago

#11 - Run benchmarks using travis

Pull Request - State: closed - Opened by bluss over 8 years ago

#10 - Use mem::zeroed to fill the gemm kernel's array for the vectors

Pull Request - State: closed - Opened by bluss over 8 years ago

#9 - Performance regression on nightly

Issue - State: closed - Opened by bluss over 8 years ago

#8 - SNB Performance

Issue - State: closed - Opened by millardjn over 8 years ago - 2 comments

#7 - Use one Vec for both packing buffers

Pull Request - State: closed - Opened by bluss over 8 years ago - 12 comments

#6 - ref_mat_mul not always the slower version

Issue - State: closed - Opened by MagaTailor over 8 years ago - 16 comments

#5 - Test build i686 with travis

Pull Request - State: closed - Opened by bluss over 8 years ago

#4 - Use a 4-by-8 microkernel for sgemm

Pull Request - State: closed - Opened by bluss over 8 years ago

#3 - Add sgemm and dgemm asm microkernels from BLIS

Pull Request - State: closed - Opened by bluss over 8 years ago - 4 comments

#2 - Non-square µ-kernels, aligned buffers and a 8-by-4 kernel for dgemm

Pull Request - State: closed - Opened by bluss over 8 years ago

#1 - A faster multiplication with more ymm1; Also make the mask kernel more generic

Pull Request - State: closed - Opened by bluss over 8 years ago - 5 comments