Hacker News Reader

Comment by blurbleblurble | Hacker News Reader

This is really cool and good to know about. I hadn't heard of this technique before, but have caught some interest in SPIRV lately while learning about WebGPU. Well I just noticed that Polly is used in the LLVM SPIRV backend... Makes sense! https://github.com/KhronosGroup/LLVM-SPIRV-Backend/blob/mast...

I'm very excited about the potential for writing GPU code from whatever high level language is convenient. (Which will not be WSL, by the way).

View on HN

It would be news to me that the PolyhedralInfo interface is actually used anywhere. As the file comment says, this was "work in progress" a few years ago and it shares the same drawbacks as Polly (unfortunately).

Polyhedral development in LLVM stagnated a while ago, maybe we find the people and time to actually land the PolyhedralValueInfo (see for example this talk: https://www.youtube.com/watch?v=xSA0XLYJ-G0).

Albert Cohen from Google gave a really good lecture on polyhedral compilation in the real world at PLISS [1] this year: slides [2], lecture videos [3, 4].

A core problem of the polyhedral approach, is that the thing that makes it so appealing (it being expressive enough to enable the many interesting program transformations), is costly for large programs making, the overall optimisation process and hence the compiler hard to scale. Naturally, this difficulty makes for interesting research problems.

[1] https://pliss.org/

[2] https://pliss2019.github.io/albert_cohen_slides.pdf

[3] https://www.youtube.com/watch?v=mt6pIpt5Wk0

[4] https://www.youtube.com/watch?v=3TNT5rFVTUY

Reply on news.ycombinator.com