================
@@ -0,0 +1,1056 @@
+// RUN: %clang_cc1 %s -flax-vector-conversions=none -ffreestanding 
-triple=x86_64 -target-feature +avx10-v2-aux \
+// RUN: -emit-llvm -o - -Wno-invalid-feature-combination -Wall -Werror | 
FileCheck %s
+// RUN: %clang_cc1 %s -flax-vector-conversions=none -ffreestanding 
-triple=i386 -target-feature +avx10-v2-aux \
+// RUN: -emit-llvm -o - -Wno-invalid-feature-combination -Wall -Werror | 
FileCheck %s
+
+#include <immintrin.h>
+
+//
+// Group A: VCVTPS2BF8 / VCVTPS2BF8S / VCVTPS2HF8 / VCVTPS2HF8S /
+//          VCVTROPS2HF8 / VCVTROPS2HF8S
+//
+
+// VCVTPS2BF8 - 128-bit
+
+__m128i test_mm_cvtps_bf8(__m128 __A) {
+  // CHECK-LABEL: @test_mm_cvtps_bf8(
+  // CHECK: call <16 x i8> @llvm.x86.avx10.mask.vcvtps2bf8.128(
----------------
mahesh-attarde wrote:

* `avx10` intrinsic namespace sounds good. I dont have any motivation to add 
new namespace unless anyone objects. I preferred `avx10v2`. 

*  I think we should add intrinsic arguments.
```
call <16 x i8> @llvm.x86.avx10.mask.vcvtps2bf8.128(<4 x float>  [[A:%.*]], <16 
x i8> zeroinitializer, i8 -1)
```

https://github.com/llvm/llvm-project/pull/206888
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to