[flang-commits] [clang] [flang] [llvm] clang: Start using new amdgpu subarch triples (PR #206482)

[Matt Arsenault via flang-commits]

================
@@ -915,7 +915,9 @@ class ToolChain {
       return;
     }
 
-    if (TT.isAMDGPU()) {
+    if (TT.isAMDGCN()) {
----------------
arsenm wrote:

It's the same, amdgcn and amdgpu are aliases. This is just dropping r600 from this case


https://github.com/llvm/llvm-project/pull/206482