Kelsey Gilbert

WGSL 2022-08-09 Minutes * (KG: This might be a larger issue, but I want to timebox it regardless this week) * MM: Idea is compiler team things it’s valuable to...

This is what I came up with for why this might be useful, even absent platform `trap` support: ```rs var arr: array fn max_in(from: u32, to: u32) { var vmax:...

I can give more context: max_in__clamp is the naive clamp-using translation of max_in. max_in__discard1 is the naive trap/discard-using translation. max_in__discard2 and max_in__discard3 are reasonable transforms for the translater/compiler to do,...

Is that a reasonable generalizable transform to expect from compilers?

Is this a reasonable transform to expect from **our** compilers?

Like, the other aspect here is that this is applicable to this toy case, but not for a case where the core alg is expected to run `to-from` times. You'd...

WGSL 2022-08-16 Minutes * KG: Don’t know where to start. Guess one thing to mention is that I didn't’ think this used to be a good idea but see why...

WGSL 2022-08-23 Minutes * DN: wanted to visit one thing from last week. Feel like flipping side saying why have to expand behaviours and why can’t JS compiler trick work...

WGSL meeting minutes 2022-03-08 * MM: Discussion with Robin. This is a proxy for new users vs. expert users. New users would probably want more characters, whereas expert uses would...

WGSL meeting minutes 2022-06-28 * Previously: [https://github.com/gpuweb/gpuweb/issues/2118#issuecomment-1117637250](https://github.com/gpuweb/gpuweb/issues/2118#issuecomment-1117637250) * AB: Something for discussion, internally forced us to think a bit more. * KG: Why 64k * AB: exactly. Thought that was...