[CMake][OpenMP] Customize default offloading arch

For the shuffle instructions in reductions we need at least sm_30
but the user may want to customize the default architecture.

Differential Revision: https://reviews.llvm.org/D38883

git-svn-id: https://llvm.org/svn/llvm-project/cfe/trunk@315996 91177308-0d34-0410-b5e6-96231b3b80d8

CMakeLists.txt

@@ -235,6 +235,17 @@ endif()
 set(CLANG_DEFAULT_OPENMP_RUNTIME "libomp" CACHE STRING
   "Default OpenMP runtime used by -fopenmp.")
 
+# OpenMP offloading requires at least sm_30 because we use shuffle instructions
+# to generate efficient code for reductions.
+set(CLANG_OPENMP_NVPTX_DEFAULT_ARCH "sm_30" CACHE STRING
+  "Default architecture for OpenMP offloading to Nvidia GPUs.")
+string(REGEX MATCH "^sm_([0-9]+)$" MATCHED_ARCH "${CLANG_OPENMP_NVPTX_DEFAULT_ARCH}")
+if (NOT DEFINED MATCHED_ARCH OR "${CMAKE_MATCH_1}" LESS 30)
+  message(WARNING "Resetting default architecture for OpenMP offloading to Nvidia GPUs to sm_30")
+  set(CLANG_OPENMP_NVPTX_DEFAULT_ARCH "sm_30" CACHE STRING
+    "Default architecture for OpenMP offloading to Nvidia GPUs." FORCE)
+endif()
+
 set(CLANG_VENDOR ${PACKAGE_VENDOR} CACHE STRING
   "Vendor-specific text for showing with version information.")
 

include/clang/Config/config.h.cmake

@@ -20,6 +20,9 @@
 /* Default OpenMP runtime used by -fopenmp. */
 #define CLANG_DEFAULT_OPENMP_RUNTIME "${CLANG_DEFAULT_OPENMP_RUNTIME}"
 
+/* Default architecture for OpenMP offloading to Nvidia GPUs. */
+#define CLANG_OPENMP_NVPTX_DEFAULT_ARCH "${CLANG_OPENMP_NVPTX_DEFAULT_ARCH}"
+
 /* Multilib suffix for libdir. */
 #define CLANG_LIBDIR_SUFFIX "${CLANG_LIBDIR_SUFFIX}"
 

lib/Driver/ToolChains/Cuda.cpp

@@ -542,9 +542,9 @@ CudaToolChain::TranslateArgs(const llvm::opt::DerivedArgList &Args,
   // flags are not duplicated.
   // Also append the compute capability.
   if (DeviceOffloadKind == Action::OFK_OpenMP) {
-    for (Arg *A : Args){
+    for (Arg *A : Args) {
       bool IsDuplicate = false;
-      for (Arg *DALArg : *DAL){
+      for (Arg *DALArg : *DAL) {
         if (A == DALArg) {
           IsDuplicate = true;
           break;
@@ -555,14 +555,9 @@ CudaToolChain::TranslateArgs(const llvm::opt::DerivedArgList &Args,
     }
 
     StringRef Arch = DAL->getLastArgValue(options::OPT_march_EQ);
-    if (Arch.empty()) {
-      // Default compute capability for CUDA toolchain is the
-      // lowest compute capability supported by the installed
-      // CUDA version.
-      DAL->AddJoinedArg(nullptr,
-          Opts.getOption(options::OPT_march_EQ),
-          CudaInstallation.getLowestExistingArch());
-    }
+    if (Arch.empty())
+      DAL->AddJoinedArg(nullptr, Opts.getOption(options::OPT_march_EQ),
+                        CLANG_OPENMP_NVPTX_DEFAULT_ARCH);
 
     return DAL;
   }

lib/Driver/ToolChains/Cuda.h

@@ -76,17 +76,6 @@ public:
   std::string getLibDeviceFile(StringRef Gpu) const {
     return LibDeviceMap.lookup(Gpu);
   }
-  /// \brief Get lowest available compute capability
-  /// for which a libdevice library exists.
-  std::string getLowestExistingArch() const {
-    std::string LibDeviceFile;
-    for (auto key : LibDeviceMap.keys()) {
-      LibDeviceFile = LibDeviceMap.lookup(key);
-      if (!LibDeviceFile.empty())
-        return key;
-    }
-    return "sm_20";
-  }
 };
 
 namespace tools {