matrixmultiply 0.3.11

General matrix multiplication for f32 and f64 matrices. Operates on matrices with general layout (they can use arbitrary row and column stride). Detects and uses SIMD features on x86/x86-64 and AArch64 transparently for higher performance. Uses a microkernel strategy, so that the implementation is easy to parallelize and optimize. Supports multithreading.
Documentation

## Guidelines

As a current guiding principle, the entrance functions (i.e sgemm, dgemm etc)
are non-generic by design, so that the compile time cost of the library is
limited.

Threading is supported using thread-tree for its lower dispatch overhead but
it's welcome to replace it by rayon - if the low dispatch overhead can be
preserved.

## Test tricks

Use MMTEST_FEATURE=avx,fma,avx2 and so on to restrict target feature detection
to the exact listed features: comma-separated list of feature names.

## Benchmarks

To run benchmarks, use `./benches/benchloop.py`

## Wasm

To test and benchmark wasm, add the wasm32-wasip1 target using rustup and
install wasmtime-cli.