[clang] [llvm] [AArch64][clang][llvm] Add ACLE Armv9.7 matrix multiply-accumulate intrinsics (PR #193017)

[Kerry McLaughlin via llvm-commits]

================
@@ -0,0 +1,16 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 6
+; RUN: llc -mtriple=aarch64-none-linux-gnu -mattr=+sve,+sve-b16mm < %s | FileCheck %s
+
+define <vscale x 8 x bfloat> @bfmmla_bf16(<vscale x 8 x bfloat> %acc, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b) #0 {
+; CHECK-LABEL: bfmmla_bf16:
+; CHECK:       // %bb.0: // %entry
+; CHECK-NEXT:    bfmmla z0.h, z1.h, z2.h
+; CHECK-NEXT:    ret
+entry:
+  %out = tail call <vscale x 8 x bfloat> @llvm.aarch64.sve.bfmmla.bf16(<vscale x 8 x bfloat> %acc, <vscale x 8 x bfloat> %a, <vscale x 8 x bfloat> %b)
+  ret <vscale x 8 x bfloat> %out
+}
+
+declare <vscale x 8 x bfloat> @llvm.aarch64.sve.bfmmla.bf16(<vscale x 8 x bfloat>, <vscale x 8 x bfloat>, <vscale x 8 x bfloat>)
+
+attributes #0 = { "target-features"="+sve,+sve-b16mm" }
----------------
kmclaughlin-arm wrote:

These attributes have already been added by the `RUN` lines

https://github.com/llvm/llvm-project/pull/193017