T0b1/wasmtime
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Optimizations to egraph framework (#5391)

Browse Source 
* Optimizations to egraph framework:

- Save elaborated results by canonical value, not latest value (union
  value). Previously we were artificially skipping and re-elaborating
  some values we already had because we were not finding them in the
  map.

- Make some changes to handling of icmp results: when icmp became
  I8-typed (when bools went away), many uses became `(uextend $I32 (icmp
  $I8 ...))`, and so patterns in lowering backends were no longer
  matching.

  This PR includes an x64-specific change to match `(brz (uextend (icmp
  ...)))` and similarly for `brnz`, but it also takes advantage of the
  ability to write rules easily in the egraph mid-end to rewrite selects
  with icmp inputs appropriately.

- Extend constprop to understand selects in the egraph mid-end.

With these changes, bz2.wasm sees a ~1% speedup, and spidermonkey.wasm
with a fib.js input sees a 16.8% speedup:

```
$ time taskset 1 target/release/wasmtime run --allow-precompiled --dir=. ./spidermonkey.base.cwasm ./fib.js
1346269
taskset 1 target/release/wasmtime run --allow-precompiled --dir=.  ./fib.js  2.14s user 0.01s system 99% cpu 2.148 total
$ time taskset 1 target/release/wasmtime run --allow-precompiled --dir=. ./spidermonkey.egraphs.cwasm ./fib.js
1346269
taskset 1 target/release/wasmtime run --allow-precompiled --dir=.  ./fib.js  1.78s user 0.01s system 99% cpu 1.788 total
```

* Review feedback.
This commit is contained in:
Chris Fallin
2022-12-07 13:23:13 -08:00
committed by GitHub
parent c5379051c4
commit 8c55b81300
8 changed files with 90 additions and 41 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
cranelift/codegen/src/isa/x64/lower.isle
View File

@@ -2916,10 +2916,19 @@
(let ((cmp IcmpCondResult (invert_icmp_cond_result (emit_cmp cc a b))))
(side_effect (jmp_cond_icmp cmp taken not_taken))))
(rule 2 (lower_branch (brz (uextend (icmp cc a b)) _ _) (two_targets taken not_taken))
(let ((cmp IcmpCondResult (invert_icmp_cond_result (emit_cmp cc a b))))
(side_effect (jmp_cond_icmp cmp taken not_taken))))
(rule 2 (lower_branch (brz (fcmp cc a b) _ _) (two_targets taken not_taken))
(let ((cmp FcmpCondResult (emit_fcmp (floatcc_inverse cc) a b)))
(side_effect (jmp_cond_fcmp cmp taken not_taken))))
(rule 2 (lower_branch (brz (uextend (fcmp cc a b)) _ _) (two_targets taken not_taken))
(let ((cmp FcmpCondResult (emit_fcmp (floatcc_inverse cc) a b)))
(side_effect (jmp_cond_fcmp cmp taken not_taken))))
(rule 1 (lower_branch (brz val @ (value_type $I128) _ _) (two_targets taken not_taken))
(side_effect (jmp_cond_icmp (cmp_zero_i128 (CC.NZ) val) taken not_taken)))
@@ -2936,6 +2945,13 @@
(let ((cmp FcmpCondResult (emit_fcmp cc a b)))
(side_effect (jmp_cond_fcmp cmp taken not_taken))))
(rule 2 (lower_branch (brnz (uextend (icmp cc a b)) _ _) (two_targets taken not_taken))
(side_effect (jmp_cond_icmp (emit_cmp cc a b) taken not_taken)))
(rule 2 (lower_branch (brnz (uextend (fcmp cc a b)) _ _) (two_targets taken not_taken))
(let ((cmp FcmpCondResult (emit_fcmp cc a b)))
(side_effect (jmp_cond_fcmp cmp taken not_taken))))
(rule 1 (lower_branch (brnz val @ (value_type $I128) _ _) (two_targets taken not_taken))
(side_effect (jmp_cond_icmp (cmp_zero_i128 (CC.Z) val) taken not_taken)))