Changes in README.md

5e751f86 - Cranelift: add note to opts/README.md describing new mid-end rule guidance after #7999. (#8015)

Wed, 28 Feb 2024 18:58:02 +0000

Cranelift: add note to opts/README.md describing new mid-end rule guidance after #7999. (#8015)While debugging #7999, we learned of a possible, extremely subtle,interaction between the design of our ISLE mid-end prelude and matchingrules with a certain structure. Because a `Value` represents an entireeclass, separate left-hand sides (as in helper rules or if-let clauses)that match on the value may match against different enodes returned bythe multi-match extractor's iterator. We then may infer some property ofone enode, another property of another enode, and perform a rewrite thatis only valid if both of those matches are on the same enode.The precise distinction is whether it is a property of the *value* -- e.g.,nonzero, or even, or within a range -- or a property of the *operation*and matched subpieces. The former is fine, the latter runs intotrouble. We found that #7719 added a helper that determined whether avalue "was a certain operator" -- actually, had any enode of a certainoperator -- and separately, matched "the operator" and extracted itsopcode and parameters (actually, *any* binary operator). The first halfcan match an opcode we support simplifying, and the second half can getthe arguments and `op` and blindly use them in the rewrite.This PR adds new guidance to avoid complex helpers and be aware ofmulti-matching behavior, preferring to write patterns directly (as thefix in #8005 does) instead. Longer-term, we also have other ideas, e.g.@jameysharp's suggestion to disallow at-patterns on multi-extractors inleft hand sides to reduce the chance of hitting this footgun. List of files: /wasmtime-44.0.1/cranelift/codegen/src/opts/README.md

7464bbcb - Add missing subsume uses in egraph rules (#7879)

Tue, 06 Feb 2024 20:16:03 +0000

Add missing subsume uses in egraph rules (#7879)* Fix a few egraph rules that needed `subsume`There were a few rules that dropped value references from the LHSwithout using subsume. I think they were probably benign as theyproduced constant results, but this change is in the spirit of ourrevised guidelines for egraph rules.* Augment egraph rule guideline 2 to talk about constants List of files: /wasmtime-44.0.1/cranelift/codegen/src/opts/README.md

5b2ae836 - Cranelift: Use a fixpoint loop to compute the best value for each eclass (#7859)

Mon, 05 Feb 2024 22:42:32 +0000

Cranelift: Use a fixpoint loop to compute the best value for each eclass (#7859)* Cranelift: Use a fixpoint loop to compute the best value for each eclassFixes #7857* Remove fixpoint loop early-continue optimization* Add document describing optimization rule invariants* Make select optimizations use subsume* Remove invalid debug assert* Remove now-unused methods* Add commutative adds to cost tests List of files: /wasmtime-44.0.1/cranelift/codegen/src/opts/README.md

b9a58148 - ISLE: split algebraic.isle into several files (#6140)

Tue, 11 Apr 2023 21:39:18 +0000

ISLE: split algebraic.isle into several files (#6140)* ISLE: split algebraic.isle into several files* delete `algebraic.clif`* Add `README.md`* Remove old `algebraic.clif` tests---------Co-authored-by: Jamey Sharp List of files: /wasmtime-44.0.1/cranelift/codegen/src/opts/README.md