User Details
- User Since
- Feb 16 2018, 3:21 AM (266 w, 1 d)
Nov 30 2022
Nov 22 2022
Fixed indentation
Nov 20 2022
Jan 19 2020
Jan 18 2020
Dec 12 2019
Sep 2 2019
Aug 26 2019
Feb 27 2019
Feb 25 2019
Feb 8 2019
Feb 7 2019
Jan 26 2019
The two corresponding test cases are not yet commited,
https://reviews.llvm.org/D57297
Jan 16 2019
Oct 25 2018
Oct 22 2018
Oct 19 2018
Back to having only a CLI option, no constructor argument.
Oct 12 2018
Adding constructor argument as well -- making it more similar to the ExpensiveCombines option.
Oct 1 2018
Jul 7 2018
Jul 6 2018
Well, apparently the test fails with the typo fix.
There is no declare dllimport void @_ZSt9terminatev() line that could be matched for CHECK-DYNAMIC-IA-DAG.
Jul 5 2018
ping @spatel
As suggested, I added test cases with all predicates (in r336346).
Jul 4 2018
I wasn't 100% sure about these being the intended checks.
Jun 28 2018
Jun 22 2018
Jun 20 2018
I was overzealous with the intrinsics, I lower really only the packed comparison now.
Jun 19 2018
Added __builtin_ia32_cmpsd_mask & __builtin_ia32_cmpss_mask.
Added __builtin_ia32_cmpsd_mask & __builtin_ia32_cmpss_mask.
Jun 18 2018
The question still left is, should we remove, auto upgrade the LLVM intrinsics not used anymore, or keep them around for when the signal behaviour is going to matter?
Jun 14 2018
Jun 13 2018
Ignoring signaling behviour - and rounding mode with it.
Also lowering __builtin_ia32_cmpsd and __builtin_ia32_cmpss.
Jun 12 2018
Jun 11 2018
Ping @efriedma
I altered the code, to ignore the the signaling behaviour, as suggested.
Also, it handles more vector cmp builtins now.