projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: dccf83a) | patch
[X86][AVX] Cleanup gather_v8i32_v8i32 special test case
authorSimon Pilgrim <llvm-dev@redking.me.uk>
Wed, 24 Mar 2021 10:57:50 +0000 (10:57 +0000)
committerSimon Pilgrim <llvm-dev@redking.me.uk>
Wed, 24 Mar 2021 11:14:38 +0000 (11:14 +0000)
commit69b71d245f27ff2538043710bb4207d2d1748099
tree16d93bf7c79a4121bcbed4ecd793dd67b89c549ctree | snapshot
parentdccf83acf95530da0425ab6293addb410fb062fdcommit | diff
[X86][AVX] Cleanup gather_v8i32_v8i32 special test case

Cleanup the gather_v8i32_v8i32 IR to more closely match how the middle-end will optimise the vector geps (exposing more splats).

This helps the gather scalarization case a lot, but shows a missed opportunity for AVX512 gathers to recognise uniform-constant indices.

And none of the cases realise that some of the gathers are really blended broadcasts....
llvm/test/CodeGen/X86/masked_gather.ll diff | blob | history
Domain: System / Toolchain;