cmd/compile: 3.14% regression in AppendMsgResyncTargetsInfo-8 sec/op at 61bc17f

[mknyszek]

Discovered a regression in sec/op of 3.14% for benchmark AppendMsgResyncTargetsInfo-8 at 61bc17f.

CC @randall77

Note: This might not matter. This benchmark has a weird history.

[randall77]

I don't see any differences in assembly for any of the of the functions that appear in a cpu profile of this benchmark.
Closing, probably just code alignment issues.

[mknyszek]

I'm inclined to delete this benchmark, based on that.