| commit | 699414a49357812bd428e9ccd37af398ca299c90 | |
| author | Nemanja Ivanovic <nemanja.i.ibm@gmail.com> | |
| Fri, 26 Oct 2018 03:19:13 +0000 (26 03:19 +0000) | ||
| committer | Nemanja Ivanovic <nemanja.i.ibm@gmail.com> | |
| Fri, 26 Oct 2018 03:19:13 +0000 (26 03:19 +0000) | ||
| tree | e7d25d9040f1b922f7a5cffc55099ed7c9062731 | treesnapshot (tar.gz zip) |
| parent | 5fca60507547fc445be04e1b29c4ba2ea533eb07 | commitdiff |
[PowerPC] Keep vector int to fp conversions in vector domain
At present a v2i16 -> v2f64 convert is implemented by extracts to scalar,
scalar converts, and merge back into a vector. Use vector converts instead,
with the int data permuted into the proper position and extended if necessary.
Patch by RolandF.
Differential revision: https://reviews.llvm.org/D53346
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@345361 91177308-0d34-0410-b5e6-96231b3b80d8
At present a v2i16 -> v2f64 convert is implemented by extracts to scalar,
scalar converts, and merge back into a vector. Use vector converts instead,
with the int data permuted into the proper position and extended if necessary.
Patch by RolandF.
Differential revision: https://reviews.llvm.org/D53346
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@345361 91177308-0d34-0410-b5e6-96231b3b80d8
| lib/Target/PowerPC/PPCISelLowering.cpp | diffblobblamehistory | |
| lib/Target/PowerPC/PPCISelLowering.h | diffblobblamehistory | |
| test/CodeGen/PowerPC/vec-itofp.ll | [new file with mode: 0644] | blob |