cmd/compile: potential compiler regression in 1.16 with heavy arithmetic #43971
Labels
FrozenDueToAge
NeedsInvestigation
Someone must examine and confirm this is a valid issue and not a duplicate of an existing one.
Performance
I just noticed a performance regression in a small benchmark with 1.16. Here's the code in question:
I'm benching on a
Intel(R) Core(TM) i9-9880H CPU @ 2.30GHz
usinggo test -v -benchtime 5s -bench Benchmark*
.Go 1.15:
Go 1.16:
The bitsliced performance got worse and the lookup table performance got better.
I'm not overly concerned about this, but it is somewhat curious. CC @josharian @bradfitz
The text was updated successfully, but these errors were encountered: