projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 9ac2e11) | patch
[X86] Reverse the operand order of the autoupgrade of the kunpack builtins.
authorCraig Topper <craig.topper@intel.com>
Mon, 12 Feb 2018 22:38:34 +0000 (22:38 +0000)
committerCraig Topper <craig.topper@intel.com>
Mon, 12 Feb 2018 22:38:34 +0000 (22:38 +0000)
commit8d19c6fba2fa1d7d0ee3ae999e9f5c3a93f8c067
treefdbb733f1c7b069a4b78edc8fa34fa578cf1396etree | snapshot
parent9ac2e1138519c78f6b4850a367b0801b5988c74ecommit | diff
[X86] Reverse the operand order of the autoupgrade of the kunpack builtins.

The second operand needs to be in the lower bits of the concatenation. This matches llvm 5.0, gcc, and icc behavior.

Fixes PR36360.

llvm-svn: 324953
llvm/lib/IR/AutoUpgrade.cpp diff | blob | history
llvm/test/CodeGen/X86/avx512-intrinsics-upgrade.ll diff | blob | history
llvm/test/CodeGen/X86/avx512bw-intrinsics-upgrade.ll diff | blob | history
Domain: System / Toolchain;