[llvm-branch-commits] [clang] clang/AMDGPU: Remove artificial restriction on --gpu-max-threads-per-block (PR #204864)

[via llvm-branch-commits]

llvmorg-github-actions[bot] wrote:


<!--LLVM PR SUMMARY COMMENT-->

@llvm/pr-subscribers-backend-hexagon

Author: Matt Arsenault (arsenm)

<details>
<summary>Changes</summary>

Previously this flag was only handled for HIP, and would produce an unused
argument warning. Also use a simpler method for forwarding the flag to cc1.

---
Full diff: https://github.com/llvm/llvm-project/pull/204864.diff


2 Files Affected:

- (modified) clang/lib/Driver/ToolChains/AMDGPU.cpp (+2-8) 
- (added) clang/test/Driver/amdgpu-openmp-max-threads.c (+5) 


``````````diff
diff --git a/clang/lib/Driver/ToolChains/AMDGPU.cpp b/clang/lib/Driver/ToolChains/AMDGPU.cpp
index 6611b133057f5..495f724338250 100644
--- a/clang/lib/Driver/ToolChains/AMDGPU.cpp
+++ b/clang/lib/Driver/ToolChains/AMDGPU.cpp
@@ -878,16 +878,10 @@ void AMDGPUToolChain::addClangTargetOptions(
       if (DriverArgs.hasArgNoClaim(options::OPT_hipstdpar))
         CC1Args.append({"-mllvm", "-amdgpu-enable-hipstdpar"});
     }
-
-    // FIXME: This should not depend on HIP
-    StringRef MaxThreadsPerBlock =
-        DriverArgs.getLastArgValue(options::OPT_gpu_max_threads_per_block_EQ);
-    if (!MaxThreadsPerBlock.empty()) {
-      CC1Args.push_back(DriverArgs.MakeArgString(
-          Twine("--gpu-max-threads-per-block=") + MaxThreadsPerBlock));
-    }
   }
 
+  DriverArgs.AddLastArg(CC1Args, options::OPT_gpu_max_threads_per_block_EQ);
+
   // Default to "hidden" visibility, as object level linking will not be
   // supported for the foreseeable future.
   // TODO: remove the SPIR-V bypass once it can encode (hidden) visibility.
diff --git a/clang/test/Driver/amdgpu-openmp-max-threads.c b/clang/test/Driver/amdgpu-openmp-max-threads.c
new file mode 100644
index 0000000000000..52f95e0182419
--- /dev/null
+++ b/clang/test/Driver/amdgpu-openmp-max-threads.c
@@ -0,0 +1,5 @@
+// Test that --gpu-max-threads-per-block is not ignored by openmp.
+// RUN: %clang -### -fopenmp --offload-arch=gfx906  --gpu-max-threads-per-block=256 %s 2>&1 | FileCheck %s
+
+// CHECK: "-cc1" "-triple" "amdgcn-amd-amdhsa"
+// CHECK-SAME: "--gpu-max-threads-per-block=256"

``````````

</details>


https://github.com/llvm/llvm-project/pull/204864