Fix missing optimization for amd-fftw & upgrade to 4.2 #84
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Per #82, amd-fftw uses
CFLAGS
as a replacement rather than an append, so all options need to be supplied by us. This pull also upgrades to 4.2, because we can.A vanilla build applies
-O3 -fomit-frame-pointer -mtune=native -malign-double -fstrict-aliasing -fno-schedule-insns
. On a modern GCC build, all of the other options are actually implied by-O3
, except for two:-mtune=native
. We omit this because a build on a virtual machine might leave us with a badly tuned library.-fno-schedule-insns
. This turns off instruction scheduling. A benchmark on my end showed no improvements from this option, I think it's better to trust the compiler here. Like the other options this might be a holdover from times where GCC's optimization was worse.This also patches
build.sh
so it can process configure commands with quoted options. The other spots in the file already useeval
, just this one didn't. I retested the build to ensure this 1) doesn't break anything and 2) the exportedCFLAGS
don't end up on other builds.