diff options
| author | Robert Maynard <robert.maynard@kitware.com> | 2018-03-27 18:59:34 (GMT) |
|---|---|---|
| committer | Brad King <brad.king@kitware.com> | 2018-03-28 13:38:43 (GMT) |
| commit | 41eab150a8ef42bbebff18ff84652e9da1ef4e75 (patch) | |
| tree | 80799f497ace9a2151b7d900e4f71fadf8353443 /Tests/CudaOnly/CMakeLists.txt | |
| parent | 88c7abb7409c235d411c4cc899377385b91075e2 (diff) | |
| download | CMake-41eab150a8ef42bbebff18ff84652e9da1ef4e75.zip CMake-41eab150a8ef42bbebff18ff84652e9da1ef4e75.tar.gz CMake-41eab150a8ef42bbebff18ff84652e9da1ef4e75.tar.bz2 | |
CUDA: Pass more link libraries to device linking
Previously we dropped non-target items from the device link line because
nvcc rejects paths to shared library files, and only with target items
do we know the kind of library. However, this also prevents projects
from linking to system-provided libraries like `cublas_device` that
contain device code.
Fix this by passing more link items to device linking. Items that are
not file paths, such as `-lfoo`, can simply be passed unconditionally.
Items that are targets known to be shared libraries can still be
skipped. Items that are paths to library files can be passed directly
if they end in `.a`. Otherwise, pass them using `-Xnvlink` to bypass
nvcc's front-end. The nvlink tool knows to ignore shared library files.
Issue: #16317
Diffstat (limited to 'Tests/CudaOnly/CMakeLists.txt')
| -rw-r--r-- | Tests/CudaOnly/CMakeLists.txt | 1 |
1 files changed, 1 insertions, 0 deletions
diff --git a/Tests/CudaOnly/CMakeLists.txt b/Tests/CudaOnly/CMakeLists.txt index 5ad6e6b..565baca 100644 --- a/Tests/CudaOnly/CMakeLists.txt +++ b/Tests/CudaOnly/CMakeLists.txt @@ -2,6 +2,7 @@ ADD_TEST_MACRO(CudaOnly.EnableStandard CudaOnlyEnableStandard) ADD_TEST_MACRO(CudaOnly.ExportPTX CudaOnlyExportPTX) ADD_TEST_MACRO(CudaOnly.GPUDebugFlag CudaOnlyGPUDebugFlag) +ADD_TEST_MACRO(CudaOnly.LinkSystemDeviceLibraries CudaOnlyLinkSystemDeviceLibraries) ADD_TEST_MACRO(CudaOnly.ResolveDeviceSymbols CudaOnlyResolveDeviceSymbols) ADD_TEST_MACRO(CudaOnly.SeparateCompilation CudaOnlySeparateCompilation) ADD_TEST_MACRO(CudaOnly.WithDefs CudaOnlyWithDefs) |