================
@@ -25,4 +25,27 @@ define amdgpu_ps void @llvm_log2_bf16_s(ptr addrspace(1)
%out, bfloat inreg %src
ret void
}
+define amdgpu_ps void @llvm_exp2_bf16_v(ptr addrspace(1) %out, bfloat %src) {
+; GCN-LABEL: llvm_exp2_bf16_v:
+; GCN: ; %bb.0:
+; GCN-NEXT: v_exp_bf16_e32 v2, v2
+; GCN-NEXT: global_store_b16 v[0:1], v2, off
+; GCN-NEXT: s_endpgm
+ %exp = call bfloat @llvm.exp2.bf16(bfloat %src)
----------------
arsenm wrote:
There's no legalization if the instruction works correctly. There's no reason
to have the clang builtin, though the codegen for the target intrinsic should
still work with the new type if legal
https://github.com/llvm/llvm-project/pull/149229
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
