llvm-project/llvm/lib/Transforms
Simon Pilgrim 216b1bf5ed [InstCombine] SSE4A constant folding and conversion to shuffles.
This patch improves support for combining the SSE4A EXTRQ(I) and INSERTQ(I) intrinsics:

1 - Converts INSERTQ/EXTRQ calls to INSERTQI/EXTRQI if the 'bit index' and 'length' operands are constant
2 - Converts INSERTQI/EXTRQI calls to shufflevector if the bit index/length are both byte aligned (we can already lower shuffles to INSERTQI/EXTRQI if its useful)
3 - Constant folding support
4 - Add zeroinitializer handling

Differential Revision: http://reviews.llvm.org/D13348

llvm-svn: 250609
2015-10-17 11:40:05 +00:00
..
Hello Revert r240137 (Fixed/added namespace ending comments using clang-tidy. NFC) 2015-06-23 09:49:53 +00:00
IPO Sample Profiles - Adjust integer types. Mostly NFC. 2015-10-15 16:36:21 +00:00
InstCombine [InstCombine] SSE4A constant folding and conversion to shuffles. 2015-10-17 11:40:05 +00:00
Instrumentation Revert "[safestack] Fast access to the unsafe stack pointer on AArch64/Android." 2015-10-15 21:26:49 +00:00
ObjCARC [PM/AA] Rebuild LLVM's alias analysis infrastructure in a way compatible 2015-09-09 17:55:00 +00:00
Scalar [RS4GC] Dont' propagate call attrs related to patchable statepoints 2015-10-16 02:41:23 +00:00
Utils Revert 250343 and 250344 2015-10-15 16:51:00 +00:00
Vectorize [LoopVectorize] Shrink integer operations into the smallest type possible 2015-10-12 12:34:45 +00:00
CMakeLists.txt
LLVMBuild.txt
Makefile