sgemm-bi 0.1.1

Deterministic, 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.
  • Feature flags
  • capi

sgemm-bi

There is very little structured metadata to build this page from currently. You should check the main library docs, readme, or Cargo.toml in case the author documented the features in them.

This version has 1 feature flags, 0 of them enabled by default.

capi

This feature flag does not enable additional features.