projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 760d1d5) | patch
[x86] Lower some trunc + shuffle patterns to vpmov[q|d][b|w]
authorMikhail Dvoretckii <mikhail.dvoretckii@intel.com>
Thu, 21 Jun 2018 14:16:45 +0000 (14:16 +0000)
committerMikhail Dvoretckii <mikhail.dvoretckii@intel.com>
Thu, 21 Jun 2018 14:16:45 +0000 (14:16 +0000)
commit22c82af5c85e33508fee35d1785e801bc8136acc
treec44766f104db700f5be55fee0c3288451690b922tree | snapshot
parent760d1d574183400344d4c82889d70fc4535fbbe5commit | diff
[x86] Lower some trunc + shuffle patterns to vpmov[q|d][b|w]

This should help in lowering the following four intrinsics:
 _mm256_cvtepi32_epi8
 _mm256_cvtepi64_epi16
 _mm256_cvtepi64_epi8
 _mm512_cvtepi64_epi8

Differential Revision: https://reviews.llvm.org/D46957

llvm-svn: 335238
llvm/lib/Target/X86/X86ISelLowering.cpp diff | blob | history
llvm/test/CodeGen/X86/shuffle-vs-trunc-256.ll diff | blob | history
llvm/test/CodeGen/X86/shuffle-vs-trunc-512.ll diff | blob | history
Domain: System / Toolchain;