[clang] [llvm] [DXIL][SPIRV] Lower WaveActiveCountBits intrinsic (PR #113382)

[Finn Plummer via llvm-commits]

================
@@ -0,0 +1,22 @@
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -triple \
+// RUN:   dxil-pc-shadermodel6.3-compute %s -emit-llvm -disable-llvm-passes -o - | \
+// RUN:   FileCheck %s --check-prefixes=CHECK,CHECK-DXIL
+// RUN: %clang_cc1 -std=hlsl2021 -finclude-default-header -triple \
+// RUN:   spirv-pc-vulkan-compute %s -emit-llvm -disable-llvm-passes -o - | \
+// RUN:   FileCheck %s --check-prefixes=CHECK,CHECK-SPIRV
+
+// Test basic lowering to runtime function call.
+
+// CHECK-LABEL: test_bool
+int test_bool(bool expr) {
+  // CHECK-SPIRV: %[[#entry_tok:]] = call token @llvm.experimental.convergence.entry()
+  // CHECK-SPIRV:  %[[RET:.*]] = call spir_func i32 @llvm.spv.wave.active.countbits(i1 %{{.*}}) [ "convergencectrl"(token %[[#entry_tok]]) ]
----------------
inbelic wrote:

I guess when the change to add the convergence token was first introduced it was important to check that the convergence token was correctly generated for the appropriate spirv operations. But now it is probably quite redundant to add that on each test, and checking for the convergent attribute is sufficient.

Will update to use `-DTARGET` as suggested.

https://github.com/llvm/llvm-project/pull/113382