Horace He
Horace He
Well, it is kinda hackish, but imo, if there's no difference for the user, then I would consider it a "proper fix".
@skawaii @CendaL that's the type of feedback I was looking for. The `zc` issue is a separate one: one that we didn't fix before because folds were practically unusable anyways...
@CendaL Yep that's currently an unavoidable consequence of this hack :( @skawaii @zwing99 Could one of you open up a new issue to track this new bug?
@skawaii @zwing99 I fixed the zc issue in the newest version.
I don't have anything concretely to add since my implementation is also slow :P But here's my implementation (I implemented the scratchpad + barrier approach outlined in the paper to...
I would also very much like this. I think the lack of the source-mapping also makes debugging IMAs with compute-sanitizer much harder as well.
I wrote one here that only uses CuteDSL (as well as numpy/matplotlib): https://gist.github.com/Chillee/e2b07157caeade8c6b0bdf463d10f833 I could clean it up and land it if folks have opinions about what API it should...
@ccecka where are the dependencies coming from? (e.g. congruent, coprofile, etc.) and yeah, in my testing I found that matplotlib ends up being a bottleneck at larger sizes. It's about...
@fengxie haha probably not. Just making the point that I think matplotlib actually scales "ok-ish" even to fairly large sizes.