49
sgemm-bi
v0.1.1 ExperimentalDeterministic, batch-invariant CUDA GEMM engine with a full training triad (forward, dW, dX) in f32 / bf16 / f16, plus an opt-in tensor-core tier that is faster than cuBLAS PEDANTIC. Bit-identical results across runs; fixed reduction order; no atomics; no cuBLAS dependency.
MIT OR Apache-2.0 Edition 2024 MSRV 1.94
Quick Verdict
- ✓Actively maintained (updated 31d ago)
- !Pre-1.0: API may have breaking changes
- ✓Tiny footprint (101KB, 2 deps)
- ✓Permissive license (MIT OR Apache-2.0)
Security
Checking security advisories...
Downloads
34
Dependents
0
Releases
2
Size
101KB
Deep Insights
📉
Download decline
9 downloads in the last 30 days, down 64% from the previous period. May indicate migration to alternatives.
🪶
Minimal dependencies
Only 2 direct dependencies. Lean dependency tree means faster builds and lower supply chain risk.
Health Breakdown
Maintenance 12/25
Recency, release consistency, active ratio
Quality 14/25
Yanked ratio, deps, size, maturity, features
Community 6/20
Reverse deps, ownership, ecosystem
Popularity 2/15
Downloads, momentum, growth trend
Documentation 15/15
Docs, repo, license, metadata
Download Trend
Daily downloads · last 90 days
0/day avg
Version Adoption
v0.1.1
56%
v0.1.0
44%
Release Timeline
2 releasessince 2026
J
F
M
A
M
J
J
A
S
O
N
D
2026
2
LessMore
Feature Flags
capi
README
Loading README...
Maintainers
Dependencies
2
direct dependencies
Dependents
0
crates depend on sgemm-bi