diff options
| author | Chandler Carruth <chandlerc@gmail.com> | 2015-05-30 04:05:11 +0000 |
|---|---|---|
| committer | Chandler Carruth <chandlerc@gmail.com> | 2015-05-30 04:05:11 +0000 |
| commit | 2599da3cfd519ed59a574d446cd7b25bf543be74 (patch) | |
| tree | 7b023a8615792cbd45647e4e5eacd802c9c96278 /llvm/lib/CodeGen | |
| parent | 6ba9730a4ef3515653d1813fb716988398ca2c5d (diff) | |
| download | bcm5719-llvm-2599da3cfd519ed59a574d446cd7b25bf543be74.tar.gz bcm5719-llvm-2599da3cfd519ed59a574d446cd7b25bf543be74.zip | |
[x86] Restore the bitcasts I removed when refactoring this to avoid
shifting vectors of bytes as x86 doesn't have direct support for that.
This removes a bunch of redundant masking in the generated code for SSE2
and SSE3.
In order to avoid the really significant code size growth this would
have triggered, I also factored the completely repeatative logic for
shifting and masking into two lambdas which in turn makes all of this
much easier to read IMO.
llvm-svn: 238637
Diffstat (limited to 'llvm/lib/CodeGen')
0 files changed, 0 insertions, 0 deletions
