Blame README.md at 666db680aa5e3e817ac67a38e6ae60344bdb16c5

3Compares matrix-matrix multiply (GEMM) performance across implementations

4running in the browser, plus a native reference measured outside it.

6**Live:** https://concept-collection.github.io/matmul-bench/

8Methods:

10- **JavaScript/TypeScript** — plain triple loop over `Float64Array` (f64), no

11 special-casing. The "no special effort" baseline.

12- **WebGPU** — a tiled compute shader (f32; WGSL has no double precision).

13- **WASM from custom C** — naive, blocked+SIMD (`f64x2`, `-msimd128`), and a

14 threaded (WASM pthreads) version of the blocked+SIMD kernel, built with

15 emscripten (f64). The threaded one needs cross-origin isolation.

16- **libFLAME/BLIS in WASM** — a real BLAS (`dgemm` via libFLAME + BLIS)

17 compiled to WebAssembly, single-threaded and multi-threaded (WASM pthreads),

18 from [libflame2wasm](https://github.com/magland/libflame2wasm) (f64). The

19 threaded build needs cross-origin isolation (SharedArrayBuffer), supplied on

20 GitHub Pages by a vendored service worker.

21- **Native LAPACK/OpenBLAS** — a fixed reference table from `dgemm` run

22 outside the browser (`native/`).

24## Develop

26```bash

27npm install

28npm run dev # local dev server (sends COOP/COEP for threaded methods)

29wasm/build-wasm.sh # rebuild wasm/dist/matmul.* + public/matmul/matmul_mt.*

30 # (needs emsdk)

31blis/build-blis-wasm.sh # rebuild public/matmul/matmul_blis_{st,mt}.* (needs emsdk

32 # + the prebuilt .a files from a sibling libflame2wasm)

33npm run build # tsc -b && vite build

34```

36The libFLAME/BLIS WASM artifacts under `public/matmul/` are committed (they link

37~30 MB of prebuilt static libraries that live in a separate

38[libflame2wasm](https://github.com/magland/libflame2wasm) checkout), so the

39GitHub Pages build doesn't rebuild them. The threaded custom module

40(`public/matmul/matmul_mt.*`) is built in CI from `wasm/matmul.c`.

42## Native reference

44```bash

45native/build.sh

46OPENBLAS_NUM_THREADS=1 native/bench_native 128 256 512 1024 2048

47OPENBLAS_NUM_THREADS=12 native/bench_native 128 256 512 1024 2048

48```

50Hand-copy the results into `src/data/nativeReference.ts`.