SIMD subgroup meeting on 2021-06-11

[ngzhian]

This is a joint meeting with flexible-vectors and any SIMD related presentations are also welcome, e.g. new architectures or toolchains, longer term discussion that don't fit into flexible-vectors or relaxed-simd.

The meeting will be on Friday, June 11 at 9:00AM - 10:00AM PDT/ 5:00PM - 6:00PM CET.

If this meeting doesn't already appear on your calendar, or you are a new attendee, please fill out this form to attend.

[ngzhian]

@penzn fyi

[nemequ]

Regarding the last item we discussed (somewhere to house performance data about WASM SIMD), I spent a bit of time putting together something at https://github.com/nemequ/wasm-simd-data.

It's a bit more ambitious than what we discussed since it's more than just performance data, but I think it makes sense…

One thing that makes me hesitate a bit is that I'm relying on SIMDe pretty heavily, which feels a bit self-serving. That really wasn't my intent when I brought up the idea yesterday, but the more I thought about how to implement it the more abusing SIMDe made sense to me. The README in the repo should explain my position a bit, but once you see how everything is put together I think you'll understand. That said, I'm still open to other ideas.

If anyone is interested in discussing anything about this new project, please use the "Discussion" tab (or "Issues" if appropriate) on that repo. I don't want to make too much noise in this repo since I'm guessing a lot of people watching probably don't care about the details.

FWIW, I have no problem adding relaxed/flexible stuff to both that repo and SIMDe as it is being developed.

[ngzhian]

Meeting notes are at https://github.com/WebAssembly/meetings/blob/main/simd/2021/SIMD-06-11.md

[penzn]

I don't want to make too much noise in this repo since I'm guessing a lot of people watching probably don't care about the details.

@nemequ, if you run into something unexpected you should still post it here, in case not everybody is following your project. I think the data repo started on the right foot and we can add upcoming SIMD proposals to it eventually. I think we eventually need to compare the data coming from SIMDe with the output of the engines, just in case.

[nemequ]

The web interface is done now, too: https://nemequ.github.io/waspr/

[ngzhian]

The website looks awesome!