Expand description
SIMD-accelerated sparse matrix-vector multiply.
Provides spmv_simd, which dispatches to an architecture-specific
implementation when the simd feature is enabled, and falls back to a
portable scalar loop otherwise.
Functions§
- spmv_
scalar - Scalar fallback implementation of SpMV.
- spmv_
scalar_ f64 - Scalar fallback for f64 SpMV.
- spmv_
simd - Sparse matrix-vector multiply with optional SIMD acceleration.
- spmv_
simd_ f64 - Sparse matrix-vector multiply with optional SIMD acceleration for f64.