[clang] [HIP][SPIR-V] Enable optimizations for amdgcnspirv target (PR #189383)

[Marcos Maronas via cfe-commits]

================
@@ -56,6 +56,6 @@
 // CHECK-SPIRV-BACKEND-BINARY: "{{.*clang(\.exe)?}}" "-cc1" "-triple" "spirv64-amd-amdhsa" {{.*}} "-emit-obj"
 // CHECK-SPIRV-BACKEND-BC: "{{.*clang(\.exe)?}}" "-cc1" "-triple" "spirv64-amd-amdhsa" {{.*}} "-emit-llvm-bc"
 // CHECK-SPIRV-BACKEND-LL: "{{.*clang(\.exe)?}}" "-cc1" "-triple" "spirv64-amd-amdhsa" {{.*}} "-emit-llvm"
-// CHECK-SPIRV-BACKEND-BINARY-EQ-TRIPLE: "{{.*clang(\.exe)?}}" "-cc1" {{.*}}"-triple=spirv64-amd-amdhsa" {{.*}}"-emit-obj"
+// CHECK-SPIRV-BACKEND-BINARY-EQ-TRIPLE: "{{.*clang(\.exe)?}}" "-cc1" "-triple=spirv64-amd-amdhsa" "-emit-obj" "-mllvm" "-vectorize-loops=false"
----------------
maarquitos14 wrote:

What about `"-mllvm", "-vectorize-slp=false"`, `"-fno-unroll-loops"` and `"-mllvm", "-interleave-loops=false"`?

https://github.com/llvm/llvm-project/pull/189383