[PATCH] D55059: [ARM] FP16: constant initialised v4f16 and v8f16 vectors
Sjoerd Meijer via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Fri Nov 30 01:43:24 PST 2018
SjoerdMeijer added a comment.
Thanks for taking a look!
I am actually going to split this up in 3 patches:
- the changes in ARMISelDAGToDAG.cpp, the are related to (post-increment) VLD1, and should also be tested explicitly
- the changes in ARMInstrNEON.td, the bitcasts,
- and then this one to support BUILD_VECTOR.
> Please clean up the testcase so it only contains the relevant operations; the whole loop isn't necessary.
I think for testing this BUILD_VECTOR change, a loop is necessary, as this code is generated from:
%vec.phi = phi <8 x half> [ zeroinitializer, %vector.ph ], [ %2, %vector.body ]
but I will have a look again.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D55059/new/
https://reviews.llvm.org/D55059
More information about the llvm-commits
mailing list