sp_matmul_rs
sp_matmul_rs provides a fast way to perform a sparse matrix multiplication (SpGEMM) followed by top-n multiplication result selection.
It is a Rust port of sparse_dot_topn, created and optimised for use in string_grouper — where the dominant workload is matching every row of a very large TF-IDF matrix against every other row and keeping only the top-n nearest neighbours.
sparse_dot_topn is based on Gustavson's algorithm
but retains only the best n values per output row. sp_matmul_rs retains the same algorithmic core
but is built around L1/L2 cache-blocking: the default driver column-chunks B so the dense per-row
accumulator and the streamed B fragments stay resident in L1 (chunk sizes are automatically calculated based on cache size), which is the dominant performance lever once the working set stops fitting in cache.
Parallelism is provided by rayon (enabled by default) and the Python distribution releases the GIL around the kernel.
On Apple M5 Pro over a 663k × 193k TF-IDF self-similarity workload sp_matmul_rs runs in as little as 41% of the time of sparse_dot_topn when retaining the top 10 values per row and utilising 10 cores.
See the benchmark directory for details.
Usage
sp_matmul_topn supports {CSR, CSC, COO} matrices with {32, 64}bit {int, float} data.
Note that COO and CSC inputs are converted to the CSR format and are therefore slower.
The Python surface (sp_matmul, sp_matmul_topn, zip_sp_matmul_topn) mirrors sparse_dot_topn, so migration from the C++ extension is typically a one-line import swap.
Note that sp_matmul_topn(A, B, top_n=B.shape[1]) is equal to sp_matmul(A, B) and A.dot(B).
import numpy as np
import scipy.sparse as sparse
from sp_matmul_rs import sp_matmul, sp_matmul_topn
A = sparse.random(1000, 100, density=0.1, format="csr", dtype=np.float64)
B = sparse.random(100, 2000, density=0.1, format="csr", dtype=np.float64)
# Compute C and retain the top 10 values per row
C = sp_matmul_topn(A, B, top_n=10)
# or parallelised matrix multiplication without top-n selection
C = sp_matmul(A, B, n_threads=2)
# or with top-n selection
C = sp_matmul_topn(A, B, top_n=10, n_threads=2)
# pass n_threads=-1 to use all but one physical core
C = sp_matmul_topn(A, B, top_n=10, n_threads=-1)
# If you are only interested in values above a certain threshold
C = sp_matmul_topn(A, B, top_n=10, threshold=0.8)
# If you set the threshold we cannot easily determine the number of non-zero
# entries beforehand. Therefore, we allocate memory for `ceil(top_n * A.shape[0] * density)`
# non-zero entries. You can set the expected density to reduce the amount pre-allocated
# entries. Note that if we allocate too little an expensive copy(ies) will need to happen.
C = sp_matmul_topn(A, B, top_n=10, threshold=0.8, density=0.1)
The package imports as sp_matmul_rs and does not share its namespace with sparse_dot_topn — both can be installed side-by-side in the same environment.
Installation
sp_matmul_rs provides wheels for CPython 3.9 to 3.13 for:
- Windows (64bit)
- Linux (x86_64 and aarch64, both manylinux and musllinux)
- macOS (x86_64 and ARM)
pip install sp_matmul_rs
sp_matmul_rs relies on a Rust extension for the computationally intensive multiplication routine.
Note that the wheels ship with rayon enabled, providing parallelisation out-of-the-box.
If you need to disable threading at runtime, omit the n_threads argument (or pass n_threads=1).
Installing from source requires a Rust toolchain (1.75+) and maturin:
pip install maturin
pip install sp_matmul_rs --no-binary sp_matmul_rs
Supported
| Values | Indices | Python | Platforms |
|---|---|---|---|
| f32, f64, i32, i64 | i32, i64 | 3.9 – 3.13 | manylinux x86_64/aarch64, musllinux x86_64, macOS x86_64/arm64, Windows x86_64 |
Rust crate
For Rust callers, the same kernels are available without the Python bindings:
[dependencies]
sp_matmul_rs = "0.1"
use sp_matmul_rs::{sp_matmul_topn, CsrView, SortMode, TopNOptions};
let c = sp_matmul_topn::<f64, i32>(
a_view,
b_view,
10,
TopNOptions { sort: SortMode::ByValue, ..Default::default() },
);
Features
| Feature | Default | What it enables |
|---|---|---|
rayon |
yes | Parallel column-chunked driver (parallel::*). |
python |
no | PyO3 + numpy bindings — used by the Python distribution. |
Module layout
| Module | Responsibility |
|---|---|
scalar |
Scalar trait + impls for f32, f64, i32, i64 |
index |
Index trait + impls for i32, i64 |
csr |
CsrView (borrowed) and CsrMatrix (owned) |
maxheap |
Bounded max-heap retaining top-n scores per row |
matmul |
Sequential sp_matmul (no top-n) |
matmul_topn |
Sequential sp_matmul_topn (single full-column path) |
chunked |
Column-chunked driver — the L1/L2 cache-blocking pillar |
zip |
zip_sp_matmul_topn for distributed/cluster results |
parallel |
Rayon-backed variants (feature rayon) |
python |
PyO3 + numpy bindings (feature python) |
Build from source
# Rust crate
cargo build --release # default features: rayon
cargo build --release --no-default-features
cargo test
cargo bench
# Python extension (requires maturin and CPython >=3.9)
pip install maturin
maturin develop --release --features python,rayon
python -c "import sp_matmul_rs; print(sp_matmul_rs.kernel_info())"
x86-64: FMA/AVX2 via runtime dispatch
Baseline x86-64 builds (including the PyPI wheels) cannot assume FMA/AVX2 at compile time, so
the chunked kernels ship AVX2+FMA clones selected by runtime CPU detection — worth ~17–18 %
on real string-matching workloads, and available on effectively every x86-64 CPU since 2013
(Haswell/Zen). No configuration needed; check which path is active with
sp_matmul_rs.kernel_info()['runtime_simd'] ("avx2+fma", "baseline", or "compile-time"),
and set SP_MATMUL_RS_FORCE_BASELINE=1 to disable the clones for A/B measurements.
Building from source with RUSTFLAGS="-C target-cpu=native" still applies the full native
feature set (including AVX-512 where present — the runtime dispatch tier is AVX2+FMA only) to
the whole crate and compiles the dispatch machinery away; on AVX2-class hardware it measures
the same as the runtime-dispatched wheel. See
bench/results_x86_i7-6500U.md for x86-64 numbers and the
chunk_cols / SP_MATMUL_RS_CHUNK_COLS tuning knob.
Benchmarks
bench/bench_rust_vs_cpp_scipy.py is a three-way comparison against the upstream C++ extension (sparse_dot_topn) and scipy.sparse on a word-level TF-IDF matrix over the EDGAR company-name corpus (Kaggle: dattapiy/sec-edgar-companies-list).
Numbers below were measured on an Apple M5 Pro (10 performance cores), 48 GB, macOS 26.4, with both libraries built with parallelism enabled (sparse_dot_topn against Homebrew libomp, sp_matmul_rs with rayon).
Times are per-call min across timed runs (one untimed warm-up first).
Full EDGAR corpus — A @ A.T (663 000 × 193 190)
The workload that string_grouper actually runs in production: every company name matched against every other.
Scipy and the no-top-n baseline are omitted because the dense result is too large to materialise on a 48 GB machine.
repeats = 3.
| Benchmark | top_n | n_threads | C++ (s) | Rust (s) | Rust vs C++ |
|---|---|---|---|---|---|
| sp_matmul_topn | 10 | 1 | 132.537 | 68.585 | 1.9x |
| sp_matmul_topn | 10 | 8 | 28.143 | 12.903 | 2.2x |
| sp_matmul_topn | 10 | 10 | 28.429 | 11.567 | 2.5x |
| sp_matmul_topn | 20 | 1 | 131.878 | 70.350 | 1.9x |
| sp_matmul_topn | 20 | 8 | 27.505 | 13.154 | 2.1x |
| sp_matmul_topn | 20 | 10 | 27.524 | 11.415 | 2.4x |
At workload-realistic scale, sp_matmul_rs consistently runs in 40–50 % of the time of the C++ extension and continues to scale past 8 threads where the OpenMP build plateaus. The L1/L2 cache-blocking driver does the heavy lifting — at 663 000 rows the per-row working set no longer fits in cache, and the chunked layout keeps the dense accumulator L1-resident.
Sampled shape — 20 000 × 193 190 × 20 000
The same shape as the upstream C++ bench README, included for direct comparability.
repeats = 5.
| Benchmark | top_n | n_threads | Scipy (s) | C++ (s) | Rust (s) | Rust vs Scipy | Rust vs C++ |
|---|---|---|---|---|---|---|---|
| sp_matmul_topn | 10 | 1 | 0.110 | 0.102 | 0.083 | 1.3x | 1.2x |
| sp_matmul_topn | 10 | 8 | 0.107 | 0.014 | 0.015 | 6.9x | -1.1x |
| sp_matmul_topn | 30 | 1 | 0.108 | 0.124 | 0.119 | -1.1x | 1.0x |
| sp_matmul_topn | 30 | 8 | 0.109 | 0.017 | 0.021 | 5.3x | -1.2x |
| sp_matmul_topn | 100 | 8 | 0.107 | 0.031 | 0.041 | 2.6x | -1.3x |
| sp_matmul_topn | 1000 | 8 | 0.107 | 0.112 | 0.180 | -1.7x | -1.6x |
At this smaller shape the per-row working set still fits in cache, so the C++ extension's flat layout is competitive (Rust trails by ~1.2–1.6× at large top_n).
The sampled bench is included for direct comparability with the upstream C++ project's published table — the full-corpus numbers above are the more relevant indicator for real string-matching workloads.
The full {top_n, n_threads} grid (including the plain sp_matmul baseline and Scipy timings) lives in bench/README.md.
License
Apache-2.0. See LICENSE.
1 maintainer
Metadata
1 maintainer
Release files for sp-matmul-rs 0.2.1
For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.
Source distribution (sdist)
| File | Size | Uploaded | Provenance | |
|---|---|---|---|---|
| sp_matmul_rs-0.2.1.tar.gz | 116.6 kB | With provenance | Details |
Built distributions (wheels)
Total release size: 15.6 MB
Release files / sp_matmul_rs-0.2.1.tar.gz
| Download URL | sp_matmul_rs-0.2.1.tar.gz |
|---|---|
| Size | 116.6 kB |
| Tags | Source |
|
SHA-256 checksum How to use checksums |
d3654a5abcadd33a6a30e5a953e8495abf45b87f27049a81458fc9fd21df2e8e
|
|
BLAKE2b-256 checksum How to use checksums |
d5035d89223cbdf0510e16ed3fe904e768609b16a566dd7ae370f2b436b0c6f8
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp313-cp313-win_amd64.whl
| Download URL | sp_matmul_rs-0.2.1-cp313-cp313-win_amd64.whl |
|---|---|
| Size | 467.9 kB |
| Tags | CPython 3.13 Windows x86-64 |
|
SHA-256 checksum How to use checksums |
0804d173494397c10970b80a2a4a2c9126c303bd672d571da9c6ee0d939ebbe5
|
|
BLAKE2b-256 checksum How to use checksums |
272abfd70d09016bfcbae68594f13d485f91b58d9add796e5acea06a3d32ad8f
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp313-cp313-musllinux_1_2_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp313-cp313-musllinux_1_2_x86_64.whl |
|---|---|
| Size | 836.6 kB |
| Tags | CPython 3.13 Linux musl 1.2+ x86-64 |
|
SHA-256 checksum How to use checksums |
7e5c9e64ddab1c97bbbe27c07917a048264f8e25c3f076a532df4fcf8deeb395
|
|
BLAKE2b-256 checksum How to use checksums |
8e9072246d74b750cab2afb9709590883b7b6e1cdf42fe1a51c0a5974c33f3e8
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp313-cp313-manylinux_2_17_x86_64.manylinux2014_x86_64.whl |
|---|---|
| Size | 626.4 kB |
| Tags | CPython 3.13 Linux glibc 2.17+ x86-64 |
|
SHA-256 checksum How to use checksums |
07dd62d68fcbe2238938cf13edd9bbf8220279fce67c1b429f525a6d21042685
|
|
BLAKE2b-256 checksum How to use checksums |
6f060df3b14a97b9f0f245f07fe2112994e0cdcba112c45da4bbed258283dbf7
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
| Download URL | sp_matmul_rs-0.2.1-cp313-cp313-manylinux_2_17_aarch64.manylinux2014_aarch64.whl |
|---|---|
| Size | 599.7 kB |
| Tags | CPython 3.13 Linux glibc 2.17+ ARM64 |
|
SHA-256 checksum How to use checksums |
60af7c166c4981ce9292c740f7b901b88b20e1ee56a7c9c12c402a47b3a50df4
|
|
BLAKE2b-256 checksum How to use checksums |
80b6a13125a7b3bc66038b7b2f097493b46148e74c24df0c74262d6ac6dbcd30
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp313-cp313-macosx_11_0_arm64.whl
| Download URL | sp_matmul_rs-0.2.1-cp313-cp313-macosx_11_0_arm64.whl |
|---|---|
| Size | 561.6 kB |
| Tags | CPython 3.13 macOS 11.0+ ARM64 |
|
SHA-256 checksum How to use checksums |
83d632c6cd30a9bd7468d584fa048abbf75a0858644600a5a17a6701ef0233b9
|
|
BLAKE2b-256 checksum How to use checksums |
2cac5ccb05da81ccfab6c87d239001fdff3d532f5b71dba3b3650df52344206a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp312-cp312-win_amd64.whl
| Download URL | sp_matmul_rs-0.2.1-cp312-cp312-win_amd64.whl |
|---|---|
| Size | 468.6 kB |
| Tags | CPython 3.12 Windows x86-64 |
|
SHA-256 checksum How to use checksums |
723043a2b6950670878afd6ac090d66094fed2dcb4783a1e119b3e7f111e2a02
|
|
BLAKE2b-256 checksum How to use checksums |
7044a221a3ebfdb6d370303df3bbda88ebff05107c4670df4ca7c7e56aa4ff4a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp312-cp312-musllinux_1_2_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp312-cp312-musllinux_1_2_x86_64.whl |
|---|---|
| Size | 836.7 kB |
| Tags | CPython 3.12 Linux musl 1.2+ x86-64 |
|
SHA-256 checksum How to use checksums |
4b362e98823588e38044e3df37b2c9d0b7c525825d7cdbf982474087ecea00cf
|
|
BLAKE2b-256 checksum How to use checksums |
3efeaf3fb56b63e02fa80f4f22efbb759d81e3e13b8ed4ce7fdeaf668c8958cc
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp312-cp312-manylinux_2_17_x86_64.manylinux2014_x86_64.whl |
|---|---|
| Size | 626.4 kB |
| Tags | CPython 3.12 Linux glibc 2.17+ x86-64 |
|
SHA-256 checksum How to use checksums |
aed6dd1633152e23319e2c0972c28e667a07aeef3c076c1d509d5984afceb8a5
|
|
BLAKE2b-256 checksum How to use checksums |
cb56ace7ec2be21681fe96b4d66ef37cb5bf820ec7d0a74c130ee5d50848db8a
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
| Download URL | sp_matmul_rs-0.2.1-cp312-cp312-manylinux_2_17_aarch64.manylinux2014_aarch64.whl |
|---|---|
| Size | 599.7 kB |
| Tags | CPython 3.12 Linux glibc 2.17+ ARM64 |
|
SHA-256 checksum How to use checksums |
54fd573a3309be30690c025b9ed38051a3f7f78b59760d041ce2921876702f49
|
|
BLAKE2b-256 checksum How to use checksums |
ad2b1dcc72e31d7cf7f1075f5ec2e0d94af7c67872b49d12caf120f5c921075e
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp312-cp312-macosx_11_0_arm64.whl
| Download URL | sp_matmul_rs-0.2.1-cp312-cp312-macosx_11_0_arm64.whl |
|---|---|
| Size | 562.2 kB |
| Tags | CPython 3.12 macOS 11.0+ ARM64 |
|
SHA-256 checksum How to use checksums |
d18652edb94525ffb2ca24968912db9a83b71acba3f99f0be3a40d4ca996cb14
|
|
BLAKE2b-256 checksum How to use checksums |
9abb601411dcffed2f1d7ce4bd9a55e6d66fc1b20c78c31c56b887d9966fa3fa
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp311-cp311-win_amd64.whl
| Download URL | sp_matmul_rs-0.2.1-cp311-cp311-win_amd64.whl |
|---|---|
| Size | 466.5 kB |
| Tags | CPython 3.11 Windows x86-64 |
|
SHA-256 checksum How to use checksums |
6bdf156a75a950947a60b990ff04f9dcdbe063362f8fe3b572154f70c67a0d26
|
|
BLAKE2b-256 checksum How to use checksums |
5ea697274a8f0c62e0ea3d10b6f509c67603bd953fbd51752f89f9441a9e0978
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp311-cp311-musllinux_1_2_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp311-cp311-musllinux_1_2_x86_64.whl |
|---|---|
| Size | 835.8 kB |
| Tags | CPython 3.11 Linux musl 1.2+ x86-64 |
|
SHA-256 checksum How to use checksums |
9dd0c558f4336672e7adfe38affd6d93b305b4fbdf47a5231e110ed7befc40a8
|
|
BLAKE2b-256 checksum How to use checksums |
bc66c87dd63e867da3fd6b32bef2c84876bbb442b11115e1dd31ebe6c7f3ef32
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp311-cp311-manylinux_2_17_x86_64.manylinux2014_x86_64.whl |
|---|---|
| Size | 625.6 kB |
| Tags | CPython 3.11 Linux glibc 2.17+ x86-64 |
|
SHA-256 checksum How to use checksums |
6051237a461d2618ae3553bd00450c42c2af0677098399a4b674a903c69a5b42
|
|
BLAKE2b-256 checksum How to use checksums |
51c8686aa07eb9858a361896a02b191f38f91d1db66f729d6345534f575f6f5c
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
| Download URL | sp_matmul_rs-0.2.1-cp311-cp311-manylinux_2_17_aarch64.manylinux2014_aarch64.whl |
|---|---|
| Size | 599.0 kB |
| Tags | CPython 3.11 Linux glibc 2.17+ ARM64 |
|
SHA-256 checksum How to use checksums |
18f98734832579edc92564416275af2f907546a9cb12a21ae352fa7d52e2744a
|
|
BLAKE2b-256 checksum How to use checksums |
46fdead7ed4f893d9eb8ebf44fca73c8476655415f5b5bc5644442c7e9824919
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp311-cp311-macosx_11_0_arm64.whl
| Download URL | sp_matmul_rs-0.2.1-cp311-cp311-macosx_11_0_arm64.whl |
|---|---|
| Size | 563.8 kB |
| Tags | CPython 3.11 macOS 11.0+ ARM64 |
|
SHA-256 checksum How to use checksums |
a750744377e9b0f7021498d52b3fde793ae97e90c857775aa607f2fed2ec0c5d
|
|
BLAKE2b-256 checksum How to use checksums |
387a25006a05b9ae78557798be28a2c6212e2cada5eb57a9fc245326ec7086fc
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp310-cp310-win_amd64.whl
| Download URL | sp_matmul_rs-0.2.1-cp310-cp310-win_amd64.whl |
|---|---|
| Size | 466.5 kB |
| Tags | CPython 3.10 Windows x86-64 |
|
SHA-256 checksum How to use checksums |
8e1b6d190a04a010cfb8817f3cddad4a9c3243e17f8f816075f34968232d3e6b
|
|
BLAKE2b-256 checksum How to use checksums |
a0ea1f5e27e7ba08d99df9e9fd610ccf16ae0c9206e154612cf0d512dbe47ba2
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp310-cp310-musllinux_1_2_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp310-cp310-musllinux_1_2_x86_64.whl |
|---|---|
| Size | 835.9 kB |
| Tags | CPython 3.10 Linux musl 1.2+ x86-64 |
|
SHA-256 checksum How to use checksums |
0296d07ee46e842ab12420f3625440977a613298094496f5ebd4bc0691f633bc
|
|
BLAKE2b-256 checksum How to use checksums |
720cfbe1111d46c8dd3e1860c66d8bb2d59b8809e642fad62e0dbc0671edc806
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl |
|---|---|
| Size | 625.4 kB |
| Tags | CPython 3.10 Linux glibc 2.17+ x86-64 |
|
SHA-256 checksum How to use checksums |
107df351fdbe5fc659053a80bbe1f04e9197d16ab235b8b00c2b01094fde0232
|
|
BLAKE2b-256 checksum How to use checksums |
a1e455a380e5a7fe76e1345df09b83655835d6dab78905009e228f3969e0eaea
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
| Download URL | sp_matmul_rs-0.2.1-cp310-cp310-manylinux_2_17_aarch64.manylinux2014_aarch64.whl |
|---|---|
| Size | 599.2 kB |
| Tags | CPython 3.10 Linux glibc 2.17+ ARM64 |
|
SHA-256 checksum How to use checksums |
50939db8dc5ebfa04ff2d2c6c7e4cee19e83b60d114e0e9a6a77e5ddefea1d6f
|
|
BLAKE2b-256 checksum How to use checksums |
660e19ba82a074a50fcb9e1e9a52b0565ad32b299045d5609f88a5b6d93ece87
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp310-cp310-macosx_11_0_arm64.whl
| Download URL | sp_matmul_rs-0.2.1-cp310-cp310-macosx_11_0_arm64.whl |
|---|---|
| Size | 563.9 kB |
| Tags | CPython 3.10 macOS 11.0+ ARM64 |
|
SHA-256 checksum How to use checksums |
3570632d2a8b03e176555b9fdd7f2e3db80015101ce07d27293d08cddbf2f7b0
|
|
BLAKE2b-256 checksum How to use checksums |
3bcce35d6523228430abf737edb0b9535e2a9294b0f5145455558e27e1710f03
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp39-cp39-win_amd64.whl
| Download URL | sp_matmul_rs-0.2.1-cp39-cp39-win_amd64.whl |
|---|---|
| Size | 468.8 kB |
| Tags | CPython 3.9 Windows x86-64 |
|
SHA-256 checksum How to use checksums |
cbff374a3e4070ca5a906f43e37e669cbfef07344874af8697f1f193b7df0279
|
|
BLAKE2b-256 checksum How to use checksums |
36ac5da46672bc252076d07375c46df5c3cecac03e4c1c3f3f8943b4c771a32b
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp39-cp39-musllinux_1_2_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp39-cp39-musllinux_1_2_x86_64.whl |
|---|---|
| Size | 838.2 kB |
| Tags | CPython 3.9 Linux musl 1.2+ x86-64 |
|
SHA-256 checksum How to use checksums |
aacbd33bc538e8cb8eaf764fd89d8eb4d169880dfe336c7c67e679cf81b71c6f
|
|
BLAKE2b-256 checksum How to use checksums |
785719fdf8da84c999160a692237b6ae2fc2395cc72ea8e14df0ff9882628768
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl
| Download URL | sp_matmul_rs-0.2.1-cp39-cp39-manylinux_2_17_x86_64.manylinux2014_x86_64.whl |
|---|---|
| Size | 628.3 kB |
| Tags | CPython 3.9 Linux glibc 2.17+ x86-64 |
|
SHA-256 checksum How to use checksums |
b05a114f6e04761f0aaaf2ae1e5a39da50fd5b081b9f61bc08b53773351730e0
|
|
BLAKE2b-256 checksum How to use checksums |
4c9666dc7592b95da6e10b263fac2c1f17c1643e9b3d727411afe1a73dcab736
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl
| Download URL | sp_matmul_rs-0.2.1-cp39-cp39-manylinux_2_17_aarch64.manylinux2014_aarch64.whl |
|---|---|
| Size | 601.8 kB |
| Tags | CPython 3.9 Linux glibc 2.17+ ARM64 |
|
SHA-256 checksum How to use checksums |
c27e2ef01d121156119f9216a07b59d6c877721b61b342bcd78561a0daaebff7
|
|
BLAKE2b-256 checksum How to use checksums |
0ac551f0c63e73aea3f0fc2927f5dd9bb774675cb29a2e2115ceda7c1ff0f803
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logRelease files / sp_matmul_rs-0.2.1-cp39-cp39-macosx_11_0_arm64.whl
| Download URL | sp_matmul_rs-0.2.1-cp39-cp39-macosx_11_0_arm64.whl |
|---|---|
| Size | 565.5 kB |
| Tags | CPython 3.9 macOS 11.0+ ARM64 |
|
SHA-256 checksum How to use checksums |
8ae18c777ed0107d039a4e4c2f275b3262d17d93703be35ac41022fae16a4507
|
|
BLAKE2b-256 checksum How to use checksums |
37ab7465191148d357da16ce8f93735d1f85b789a22b219186f6924a6794bceb
|
| Upload date | |
|
Uploaded using Trusted Publishing? What is trusted publishing? |
Yes |
| Uploaded via |
twine/6.1.0 CPython/3.13.12
|
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logSecurity for sp-matmul-rs 0.2.1
Provenance
Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.
Full provenance
All 26 files in this release have been uploaded with provenance metadata.
Attestation for sp_matmul_rs-0.2.1.tar.gz
PyPI Publish Attestation
PyPI verified that this artifact, at this checksum, originated from the publisher listed below.
Signed by GitHub Actions, verified by PyPI on Jul 15, 2026.
Transparency logA note on package security
PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.
To learn more about protecting yourself from malicious actors, visit our package security documentation.
