diff options
| author | Chandler Carruth <chandlerc@gmail.com> | 2014-09-25 10:21:15 +0000 |
|---|---|---|
| committer | Chandler Carruth <chandlerc@gmail.com> | 2014-09-25 10:21:15 +0000 |
| commit | e91d68c47547d714ec13bb0b2f3b3bd0c11c3a5e (patch) | |
| tree | c11da6e11eead27bec9eab73df1a51f808524e90 /llvm/lib/CodeGen/MachineSink.cpp | |
| parent | 3256b26ef24106935a803addb212bbb0bccb50ea (diff) | |
| download | llvm-e91d68c47547d714ec13bb0b2f3b3bd0c11c3a5e.zip llvm-e91d68c47547d714ec13bb0b2f3b3bd0c11c3a5e.tar.gz llvm-e91d68c47547d714ec13bb0b2f3b3bd0c11c3a5e.tar.bz2 | |
[x86] Teach the new vector shuffle lowering a fancier way to lower
256-bit vectors with lane-crossing.
Rather than immediately decomposing to 128-bit vectors, try flipping the
256-bit vector lanes, shuffling them and blending them together. This
reduces our worst case shuffle by a pretty significant margin across the
board.
llvm-svn: 218446
Diffstat (limited to 'llvm/lib/CodeGen/MachineSink.cpp')
0 files changed, 0 insertions, 0 deletions