[PATCH] D47823: AMDGPU: Try a lot harder to emit scalar loads

Matt Arsenault via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Wed Jun 6 23:42:42 PDT 2018


arsenm added inline comments.


================
Comment at: lib/Target/AMDGPU/SIISelLowering.cpp:5364
+                                MVT::i32, SL, Ld->getChain(), Ptr,
+                                DAG.getUNDEF(Ptr.getValueType()),
+                                Ld->getPointerInfo(), MVT::i32,
----------------
rampitec wrote:
> arsenm wrote:
> > rampitec wrote:
> > > arsenm wrote:
> > > > rampitec wrote:
> > > > > Why do not you want to get pointer, pointer info, AA and the rest from original load?
> > > > Question wording is broken.
> > > > 
> > > > Everything remains the same, except the memory and possibly result type. I don't think the pointer info access size needs to increase, although I'm not 100% sure
> > > If later there will be some more info on an original load it will be lost here and it will need patching. IMO it is better to copy what is possible to copy.
> > This is already copying everything?
> I would copy a pointer instead if creating a new undef. Ponter info, tbaa and range are probably cannot be just copied.
Range needs to be adjusted or dropped, but tbaa and the rest of the pointer info should be fine. This isn't changing anything from an aliasing perspective and we don't actually use the high bits, so tbaa and the rest of the pointer info should be fine


https://reviews.llvm.org/D47823





More information about the llvm-commits mailing list