[PATCH] D39720: [X86][AVX512] lowering kunpack intrinsic - llvm part

[jina via Phabricator via llvm-commits]

jina.nahias added inline comments.


================
Comment at: test/CodeGen/X86/avx512-intrinsics-fast-isel.ll:6
 ; NOTE: This should use IR equivalent to what is generated by clang/test/CodeGen/avx512f-builtins.c
-
 define <8 x i64> @test_mm512_mask_set1_epi32(<8 x i64> %__O, i16 zeroext %__M, i32 %__A) {
----------------
jina.nahias wrote:
> craig.topper wrote:
> > jina.nahias wrote:
> > > jina.nahias wrote:
> > > > there are no tests here cause of a bug not related to this patch. will be added after his fixing. 
> > > >>! In D39720#918931, @craig.topper wrote:
> > > > I don't see any tests that produce kunpckbw after this change.
> > > 
> > > it should be here. @craig.topper 
> > I only see a test case from avx512-intrinsics.ll moved to avx512-intrinsics-upgrade.ll, but its not producing a kunpckbw instruction. I don't see a test case that produces a kunpck from shift and or in this diff. Unless there's one already in trunk?
> there are no tests producing a kunpckbw instruction cause of a bug not related to this patch. will be added after his fixing. i wont push this commit before his fixing.
> @guyblank @craig.topper 
Done @craig.topper 


https://reviews.llvm.org/D39720