- User Since
- May 5 2014, 7:26 AM (198 w, 3 d)
Don't inline matchUnaryPredicate/matchBinaryPredicate
Commit it by itself, no need to wait for another review, as long as its a separate commit its fine to tag it with 'Differential Revision: https://reviews.llvm.org/D41278' (like you should have done for rL325217 as well).
Tue, Feb 20
@mgrang I noticed you resolved PR35135 - what are you intending to do with this patch?
LGTM with a couple of minors.
LGTM - still don't understand why we don't have a single overidable function instead of the getShiftAmountTy/getScalartShiftAmountTy duplication.
Mon, Feb 19
@craig.topper mentioned that IMUL support is missing from SandyBridge as well (PR36084)
LGTM - as you've said in the comment I'd still like us to look into improving target specific cost values instead of relying on these defaults.
Replaced with D43449
Sun, Feb 18
Does anyone know why we split getShiftAmountTy / getScalarShiftAmountTy like we do? Couldn't we just get rid of getScalarShiftAmountTy and perform it all in getShiftAmountTy?
I noticed MSP430, did you see anything else? Everything else seemed to default to i32 which should be OK.
Sat, Feb 17
I still think it'd be better if you treated splat vectors as a vector instead of a scalar - your change to matchUnaryPredicate means that we're accepting UNDEF elements where we weren't before, which for DIV/REM opcodes is supposed to be a big no-no.
Rebased, and dropped remnants of UMIN/UMAX support.
@fhahn Do you have any ARM/AARCH64 concerns? Otherwise, I'm happy with this change.
OK to commit then?
Fri, Feb 16
LGTM, thanks. TBH I don't think its worth adding a non-const version of getRawData()
Thu, Feb 15
LGTM - don't suppose we have any load fold tests for BT do we?
Wed, Feb 14
Finally gotten back to a non-MSVC build machine, I'll get those signed/unsigned warning fixed up properly in the next diff.
Tue, Feb 13
What is happening with this patch? It's been in development for over a year now and still seems to be having problems. Getting PR30787 fixed would be VERY useful and I'm thinking we should be looking at alternatives if this patch is going to carry on stalling/reverting.
LGTM - I'm curious whether it's be useful for those combines to support vectors at some point, X86 combineANDXORWithAllOnesIntoANDNP might match more cases then.
Mon, Feb 12
@sdardis Are you wanting to fix the mips buildvector undef handling before this patch can land?