wgpu-mm VS webgpu-blas

This is very exciting! (I had suspected it would slip to 114)

WebGPU implementations are still pretty immature, but certainly enough to get started with. I've been implementing a Rust + WebGPU ML runtime for the past few months and have enjoyed writing WGSL.

I recently got a 250M parameter LLM running in the browser without much optimisation and it performs pretty well! (https://twitter.com/fleetwood___/status/1638469392794091520)

That said, matmuls are still pretty handicapped in the browser (especially considering the bounds checking enforced in the browser). From my benchmarking I've struggled to hit 50% of theoretical FLOPS, which is cut down to 30% when the bounds checking comes in. (Benchmarks here: https://github.com/FL33TW00D/wgpu-mm)

I look forward to accessing shader cores as they mentioned in the post.

Looks like no -- there appears to be no tensor core or similar support and this SGEMM (fp32 matrix multiply) benchmark gets awful results (my laptop gets 330gflops on this when it should be capable of 13000 gflops).

https://github.com/milhidaka/webgpu-blas

I think you will be interested to read this article about the future of data programming in JavaScript (http://benschmidt.org/post/2020-01-15/2020-01-15-webgpu/).

I do think that this kind of thing will be able to be built on top of WebGPU (I saw this experimental POC that did so recently: https://github.com/milhidaka/webgpu-blas). The only issue is that since JavaScript doesn't support operator overloading, the code might be a little less readable.