[PATCH] D80752: [AArch64]: BFloat MatMul Intrinsics&CodeGen

[Mikhail Maltsev via Phabricator via llvm-commits]

miyuki added inline comments.


================
Comment at: clang/test/CodeGen/aarch64-bf16-dotprod-intrinsics.c:2
+// RUN: %clang_cc1 -triple aarch64-arm-none-eabi \
+// RUN:   -O2 -target-feature +neon -target-feature +bf16 \
+// RUN:   -emit-llvm -o - %s | FileCheck %s
----------------
Is it possible to avoid running the whole -O2 pipeline and instead run, say,

```
%clang_cc1 -triple aarch64-arm-none-eabi -target-feature +neon -target-feature +bf16 \
-disable-O0-optnone -emit-llvm %s -o - | opt -S -mem2reg -instcombine | FileCheck %s
```

Also, I suggest auto-generating the checks using `llvm/utils/update_cc_test_checks.py`.  Sorry, I should have mentioned it in the previous review iteration.


================
Comment at: clang/test/CodeGen/aarch64-bf16-dotprod-intrinsics.c:10
+// CHECK-NEXT  %0 = bitcast <4 x bfloat> %a to <8 x i8>
+// CHECK-NEXT  %1 = bitcast <4 x bfloat> %b to <8 x i8>
+// CHECK-NEXT  %vbfdot1.i = tail call <2 x float> @llvm.aarch64.neon.bfdot.v2f32.v8i8(<2 x float> %r, <8 x i8> %0, <8 x i8> %1)
----------------
`CHECK-NEXT:`


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D80752/new/

https://reviews.llvm.org/D80752