Greatly speed up nativeCodeGen/seqBlocks
authorJoachim Breitner <mail@joachim-breitner.de>
Sat, 16 May 2015 15:47:31 +0000 (17:47 +0200)
committerJoachim Breitner <mail@joachim-breitner.de>
Sat, 16 May 2015 19:29:18 +0000 (21:29 +0200)
commit8e4dc8fb63b8d3bfee485c1c830776f3ed704f4d
tree80194342c14dc6ca1ea6de9ea046ff69bc78ad1a
parentc256357242ee2dd282fd0516260edccbb7617244
Greatly speed up nativeCodeGen/seqBlocks

When working on #10397, I noticed that "reorder" in
nativeCodeGen/seqBlocks took more than 60% of the time. With this
refactoring, it does not even show up in the profile any more. This
fixes #10422.

Differential Revision: https://phabricator.haskell.org/D893
compiler/nativeGen/AsmCodeGen.hs
testsuite/tests/perf/compiler/all.T