[clang] e8fd998 - [HIP] support --offload-arch=native

[Yaxun Liu via cfe-commits]

Author: Yaxun (Sam) Liu
Date: 2022-12-13T10:09:33-05:00
New Revision: e8fd998e6194d8749b83b30813e3cb74e5261770

URL: https://github.com/llvm/llvm-project/commit/e8fd998e6194d8749b83b30813e3cb74e5261770
DIFF: https://github.com/llvm/llvm-project/commit/e8fd998e6194d8749b83b30813e3cb74e5261770.diff

LOG: [HIP] support --offload-arch=native

This patch detects system GPU and use them
in --offload-arch if 'native' is specified. If system GPU
cannot be detected clang will fall back to the default GPU arch.

Reviewed by: Artem Belevich

Differential Revision: https://reviews.llvm.org/D139045

Added: 
    

Modified: 
    clang/lib/Driver/Driver.cpp
    clang/lib/Driver/ToolChains/AMDGPU.h

Removed: 
    


################################################################################
diff  --git a/clang/lib/Driver/Driver.cpp b/clang/lib/Driver/Driver.cpp
index 969d38560018b..c7efe60b23359 100644
--- a/clang/lib/Driver/Driver.cpp
+++ b/clang/lib/Driver/Driver.cpp
@@ -3067,6 +3067,17 @@ class OffloadingActionBuilder final {
           if (A->getOption().matches(options::OPT_no_offload_arch_EQ) &&
               ArchStr == "all") {
             GpuArchs.clear();
+          } else if (ArchStr == "native" &&
+                     ToolChains.front()->getTriple().isAMDGPU()) {
+            auto *TC = static_cast<const toolchains::HIPAMDToolChain *>(
+                ToolChains.front());
+            SmallVector<std::string, 1> GPUs;
+            auto Err = TC->detectSystemGPUs(Args, GPUs);
+            if (!Err) {
+              for (auto GPU : GPUs)
+                GpuArchs.insert(Args.MakeArgString(GPU));
+            } else
+              llvm::consumeError(std::move(Err));
           } else {
             ArchStr = getCanonicalOffloadArch(ArchStr);
             if (ArchStr.empty()) {

diff  --git a/clang/lib/Driver/ToolChains/AMDGPU.h b/clang/lib/Driver/ToolChains/AMDGPU.h
index 0ac13ecaa6345..3f5461a1c5696 100644
--- a/clang/lib/Driver/ToolChains/AMDGPU.h
+++ b/clang/lib/Driver/ToolChains/AMDGPU.h
@@ -107,6 +107,9 @@ class LLVM_LIBRARY_VISIBILITY AMDGPUToolChain : public Generic_ELF {
   llvm::Error getSystemGPUArch(const llvm::opt::ArgList &Args,
                                std::string &GPUArch) const;
 
+  llvm::Error detectSystemGPUs(const llvm::opt::ArgList &Args,
+                               SmallVector<std::string, 1> &GPUArchs) const;
+
 protected:
   /// Check and diagnose invalid target ID specified by -mcpu.
   virtual void checkTargetID(const llvm::opt::ArgList &DriverArgs) const;
@@ -126,8 +129,6 @@ class LLVM_LIBRARY_VISIBILITY AMDGPUToolChain : public Generic_ELF {
   /// Get GPU arch from -mcpu without checking.
   StringRef getGPUArch(const llvm::opt::ArgList &DriverArgs) const;
 
-  llvm::Error detectSystemGPUs(const llvm::opt::ArgList &Args,
-                               SmallVector<std::string, 1> &GPUArchs) const;
 };
 
 class LLVM_LIBRARY_VISIBILITY ROCMToolChain : public AMDGPUToolChain {