projects
/
oota-llvm.git
/ search
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
shortlog
|
log
|
commit
|
commitdiff
|
tree
first ⋅ prev ⋅ next
More FMA folding opportunities.
2015-01-06
Filipe Cabecinhas
Don't loop endlessly for MachO files with 0 ncmds
commit
|
commitdiff
|
tree
2014-10-22
Filipe Cabecinhas
Silence gcc's -Wcomment
commit
|
commitdiff
|
tree
2014-10-13
Filipe Cabecinhas
Fix a broadcast related regression on the vector shuffle...
commit
|
commitdiff
|
tree
2014-07-24
Filipe Cabecinhas
Fixed PR20411 - bug in getINSERTPS()
commit
|
commitdiff
|
tree
2014-06-22
Filipe Cabecinhas
Fix PR20087 by using the source index when changing...
commit
|
commitdiff
|
tree
2014-06-06
Filipe Cabecinhas
Fixed a bug in lowering shuffle_vectors to insertps
commit
|
commitdiff
|
tree
2014-05-31
Filipe Cabecinhas
Make blend tests more specific
commit
|
commitdiff
|
tree
2014-05-30
Filipe Cabecinhas
Convert a vselect into a concat_vector if possible
commit
|
commitdiff
|
tree
2014-05-30
Filipe Cabecinhas
Separate the check for blend shuffle_vector masks
commit
|
commitdiff
|
tree
2014-05-29
Filipe Cabecinhas
Added tests for shufflevector lowering to blend instrs.
commit
|
commitdiff
|
tree
2014-05-27
Filipe Cabecinhas
Post-commit fixes for r209643
commit
|
commitdiff
|
tree
2014-05-27
Filipe Cabecinhas
Convert some X86 blendv* intrinsics into IR.
commit
|
commitdiff
|
tree
2014-05-23
Filipe Cabecinhas
llvm-ar: Output the file we errored on.
commit
|
commitdiff
|
tree
2014-05-19
Filipe Cabecinhas
Added more insertps optimizations
commit
|
commitdiff
|
tree
2014-05-19
Filipe Cabecinhas
Change the blend tests to AVX, not AVX2.
commit
|
commitdiff
|
tree
2014-05-16
Filipe Cabecinhas
Added tests for the cost of lowering VSELECT instructions.
commit
|
commitdiff
|
tree
2014-05-16
Filipe Cabecinhas
Implemented special cases for PerformVSELECTCombine.
commit
|
commitdiff
|
tree
2014-05-16
Filipe Cabecinhas
Lower vselects into X86ISD::BLENDI when appropriate.
commit
|
commitdiff
|
tree
2014-05-16
Filipe Cabecinhas
Implemented LowerVSELECT to custom lower some instructions.
commit
|
commitdiff
|
tree
2014-05-11
Filipe Cabecinhas
Fixed a bug when lowering build_vector (PR19694)
commit
|
commitdiff
|
tree
2014-05-08
Filipe Cabecinhas
Optimize shufflevector that copies an i64/f64 and zeros...
commit
|
commitdiff
|
tree
2014-05-08
Filipe Cabecinhas
Lower certain build_vectors to insertps instructions
commit
|
commitdiff
|
tree
2014-05-05
Filipe Cabecinhas
Revert "Optimize shufflevector that copies an i64/f64...
commit
|
commitdiff
|
tree
2014-05-05
Filipe Cabecinhas
Optimize shufflevector that copies an i64/f64 and zeros...
commit
|
commitdiff
|
tree
2014-04-26
Filipe Cabecinhas
Appease the almighty buildbots.
commit
|
commitdiff
|
tree
2014-04-25
Filipe Cabecinhas
Optimization for certain shufflevector by using insertps.
commit
|
commitdiff
|
tree
2014-04-24
Filipe Cabecinhas
Optimize some special cases for SSE4a insertqi
commit
|
commitdiff
|
tree
2014-04-21
Filipe Cabecinhas
Rename X86insrtps to the proper instruction name.
commit
|
commitdiff
|
tree
2014-04-09
Filipe Cabecinhas
Revert "YAMLIO: Encode ambiguous hex strings explicitly"
commit
|
commitdiff
|
tree
2013-04-25
Filipe Cabecinhas
Allow users to choose identity used to sign tools.
commit
|
commitdiff
|
tree
2012-05-09
Filipe Cabecinhas
Fixed a typo
commit
|
commitdiff
|
tree