summaryrefslogtreecommitdiff
path: root/test/lit/passes/inlining-optimizing_optimize-level=3.wast
diff options
context:
space:
mode:
authorAlon Zakai <azakai@google.com>2021-08-18 16:05:19 -0700
committerGitHub <noreply@github.com>2021-08-18 16:05:19 -0700
commit29a51335bfb81fb0f2916039ead3fb5385ef92a4 (patch)
tree572dc5206b5d4f3ce9e4dfaef001b696ecdc6834 /test/lit/passes/inlining-optimizing_optimize-level=3.wast
parent3d6ddaf16232c42ab9a82f5114d562c2d8807870 (diff)
downloadbinaryen-29a51335bfb81fb0f2916039ead3fb5385ef92a4.tar.gz
binaryen-29a51335bfb81fb0f2916039ead3fb5385ef92a4.tar.bz2
binaryen-29a51335bfb81fb0f2916039ead3fb5385ef92a4.zip
Optimize LocalCSE hash computations using a stack. NFC (#4091)
Before, we'd compute the hash of a child, then store that in a map, then the parent would find the child's hash in the map using the pointer to the child. But as we do a simple postorder walk, we can use a stack, and avoid hashing the child pointers. This makes it 10% faster or so.
Diffstat (limited to 'test/lit/passes/inlining-optimizing_optimize-level=3.wast')
0 files changed, 0 insertions, 0 deletions