emankov added inline comments.

================
Comment at: clang/lib/Driver/ToolChains/Cuda.cpp:158
     if (!VersionFile) {
-      // CUDA 7.0 doesn't have a version.txt, so guess that's our version if
-      // version.txt isn't present.
-      Version = CudaVersion::CUDA_70;
+      // CUDA 7.0 and CUDA 11.1+ do not have version.txt file.
+      // Use libdevice file to distinguish 7.0 from the new versions.
----------------
CUDA 11.0+, actually


================
Comment at: clang/lib/Driver/ToolChains/Cuda.cpp:161
+      if (FS.exists(LibDevicePath + "/libdevice.10.bc")) {
+        Version = CudaVersion::LATEST;
+        DetectedVersionIsNotSupported = Version > 
CudaVersion::LATEST_SUPPORTED;
----------------
Do we have any other mechanism besides version.txt for determining an exact 
CUDA version? Setting the latest version in case of absence of version.txt 
doesn't suit all the needs: sometimes the exact version is taken into account, 
for instance in [[ https://github.com/ROCm-Developer-Tools/HIPIFY#clang | 
hipify-clang ]].


================
Comment at: clang/test/Driver/cuda-version-check.cu:13
 // RUN:    FileCheck %s --check-prefix=UNKNOWN_VERSION
+// CUDA-11.1 does not carry version.txt file. Make sure we still detect it as a
+// new version and handle it the same as we handle other new CUDA versions.
----------------
CUDA-11.0 Update 1 doesn't carry version.txt as well.


================
Comment at: clang/test/Driver/cuda-version-check.cu:72
 
-// UNKNOWN_VERSION: Unknown CUDA version 999.999. Assuming the latest 
supported version
+// UNKNOWN_VERSION: Unknown CUDA version{{.*}}. Assuming the latest supported 
version
 // UNKNOWN_VERSION_CXX-NOT: Unknown CUDA version
----------------
missing space before {{


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D89752/new/

https://reviews.llvm.org/D89752

_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits

Reply via email to