[llvm-branch-commits] [clang] [llvm] Builtins & Codegen support for v_cvt_scalef32_pk_{fp|bf}8_{f|bf}16 for gfx950 (PR #117742)

[Shilei Tian via llvm-branch-commits]

================
@@ -15,9 +15,10 @@ typedef unsigned int uint2 __attribute__((ext_vector_type(2)));
 typedef half __attribute__((ext_vector_type(2))) half2;
 typedef short __attribute__((ext_vector_type(2))) short2;
 typedef float __attribute__((ext_vector_type(2))) float2;
+typedef __bf16 __attribute__((ext_vector_type(2))) bfloat2;
 
 void test(global uint* out, global uint2* out_v2u32, uint a, uint b, global half2* out_v2f16, global float* out_f32, float scale, global short2* out_v2i16, float src0, float src1,
-          global float2* out_v2f32) {
+          global float2* out_v2f32, half2 src0_v2f16, bfloat2 src0_v2bf16) {
----------------
shiltian wrote:

and then all the following patches were built on top of it

https://github.com/llvm/llvm-project/pull/117742