Stream: git-wasmtime

Topic: wasmtime / PR #8005 cranelift: Fix ireduce rules

view this post on Zulip Wasmtime GitHub notifications bot (Feb 28 2024 at 01:27):

jameysharp edited PR #8005.

view this post on Zulip Wasmtime GitHub notifications bot (Feb 28 2024 at 01:38):

elliottt submitted PR review.

view this post on Zulip Wasmtime GitHub notifications bot (Feb 28 2024 at 02:28):

jameysharp merged PR #8005.

view this post on Zulip Wasmtime GitHub notifications bot (Mar 28 2024 at 06:42):

scottmcm commented on PR #8005:

Oops! Sorry for the mess here.

However, we've decided that checking whether the rule is actually an improvement is not desirable. In general, that should be the job of the cost function. Blindly adding equivalent expressions gives us more opportunities for other rules to fire, and we have global recursion and growth limits to keep the process from going too wild.

Good to see the new guidance for this, since it had been uncertain back when I first started doing these rules in https://github.com/bytecodealliance/wasmtime/pull/7693#discussion_r1430497487

view this post on Zulip Wasmtime GitHub notifications bot (Mar 28 2024 at 15:30):

jameysharp commented on PR #8005:

Yeah, we're all still figuring out how best to use this egraph optimizer. The best we can do is keep trying things out, and I appreciate your efforts to do that!

Last updated: Dec 23 2024 at 12:05 UTC