[llvm-branch-commits] [clang] clang/AMDGPU: Emit atomicrmw for __builtin_amdgcn_global_atomic_fadd_{f32|f64} (PR #96872)

[Matt Arsenault via llvm-branch-commits]

================
@@ -49,7 +49,7 @@ void test_s_wait_event_export_ready() {
 }
 
 // CHECK-LABEL: @test_global_add_f32
-// CHECK: {{.*}}call{{.*}} float @llvm.amdgcn.global.atomic.fadd.f32.p1.f32(ptr addrspace(1) %{{.*}}, float %{{.*}})
+// CHECK: = atomicrmw fadd ptr addrspace(1) %addr, float %x syncscope("agent") seq_cst, align 4, !amdgpu.no.fine.grained.memory !{{[0-9]+}}, !amdgpu.ignore.denormal.mode !{{[0-9]+$}}
----------------
arsenm wrote:

We have all the codegen tests in the backends. Clang tests checking ISA are discouraged. The set of atomic tests is hard enough to maintain as it is 

https://github.com/llvm/llvm-project/pull/96872