[CUDA] Detect and link with CUDA's libdevice bitcode library.

- added detection of libdevice bitcode file and API to find one appropriate for the GPU we're compiling for.
- pass additional cc1 options for linking with detected libdevice bitcode
- added -nocudalib to prevent automatic linking with libdevice
- added test cases to verify new functionality

Differential Revision: http://reviews.llvm.org/D14556

llvm-svn: 253387

1 files changed, 41 insertions, 0 deletions

diff --git a/clang/lib/Driver/ToolChains.cpp b/clang/lib/Driver/ToolChains.cpp
index 34504732a12..531e7900b11 100644
--- a/clang/lib/Driver/ToolChains.cpp
+++ b/clang/lib/Driver/ToolChains.cpp
@@ -1647,6 +1647,31 @@ void Generic_GCC::CudaInstallationDetector::init(
           D.getVFS().exists(CudaLibDevicePath)))
       continue;
 
+    std::error_code EC;
+    for (llvm::sys::fs::directory_iterator LI(CudaLibDevicePath, EC), LE;
+         !EC && LI != LE; LI = LI.increment(EC)) {
+      StringRef FilePath = LI->path();
+      StringRef FileName = llvm::sys::path::filename(FilePath);
+      // Process all bitcode filenames that look like libdevice.compute_XX.YY.bc
+      const StringRef LibDeviceName = "libdevice.";
+      if (!(FileName.startswith(LibDeviceName) && FileName.endswith(".bc")))
+        continue;
+      StringRef GpuArch = FileName.slice(
+          LibDeviceName.size(), FileName.find('.', LibDeviceName.size()));
+      CudaLibDeviceMap[GpuArch] = FilePath.str();
+      // Insert map entries for specifc devices with this compute capability.
+      if (GpuArch == "compute_20") {
+        CudaLibDeviceMap["sm_20"] = FilePath;
+        CudaLibDeviceMap["sm_21"] = FilePath;
+      } else if (GpuArch == "compute_30") {
+        CudaLibDeviceMap["sm_30"] = FilePath;
+        CudaLibDeviceMap["sm_32"] = FilePath;
+      } else if (GpuArch == "compute_35") {
+        CudaLibDeviceMap["sm_35"] = FilePath;
+        CudaLibDeviceMap["sm_37"] = FilePath;
+      }
+    }
+
     IsValid = true;
     break;
   }
@@ -4195,6 +4220,22 @@ CudaToolChain::addClangTargetOptions(const llvm::opt::ArgList &DriverArgs,
                                      llvm::opt::ArgStringList &CC1Args) const {
   Linux::addClangTargetOptions(DriverArgs, CC1Args);
   CC1Args.push_back("-fcuda-is-device");
+
+  if (DriverArgs.hasArg(options::OPT_nocudalib))
+    return;
+
+  std::string LibDeviceFile = CudaInstallation.getLibDeviceFile(
+      DriverArgs.getLastArgValue(options::OPT_march_EQ));
+  if (!LibDeviceFile.empty()) {
+    CC1Args.push_back("-mlink-cuda-bitcode");
+    CC1Args.push_back(DriverArgs.MakeArgString(LibDeviceFile));
+
+    // Libdevice in CUDA-7.0 requires PTX version that's more recent
+    // than LLVM defaults to. Use PTX4.2 which is the PTX version that
+    // came with CUDA-7.0.
+    CC1Args.push_back("-target-feature");
+    CC1Args.push_back("+ptx42");
+  }
 }
 
 llvm::opt::DerivedArgList *