Commit Graph

102 Commits

Author SHA1 Message Date
James Bigler 6a19150b54 FindCUDA: Add support for separable compilation
This adds a new variable, CUDA_SEPARABLE_COMPILATION, and two new
functions, CUDA_COMPUTE_SEPARABLE_COMPILATION_OBJECT_FILE_NAME and
CUDA_LINK_SEPARABLE_COMPILATION_OBJECTS.

When CUDA_SEPARABLE_COMPILATION is specified then CUDA runtime objects
will be compiled with the separable compilation flag.  These object
files are collected in a target named variable that can be used in
CUDA_COMPUTE_SEPARABLE_COMPILATION_OBJECT_FILE_NAME and
CUDA_LINK_SEPARABLE_COMPILATION_OBJECTS.
2013-03-12 13:38:09 -04:00
James Bigler 3497bc2cee FindCUDA: Remove linkage against CUDA driver library (#13084)
The CUDA runtime library (libcudart) doesn't depend on the driver API
library (libcuda) and shouldn't be used to link CUDA run time codes.
2013-03-12 13:37:44 -04:00
David Cole 3dea63d04d Merge topic 'topics/FindCUDA/Add-CUDA_HOST_COMPILER'
db207e4 FindCUDA: Added CUDA_HOST_COMPILER variable.
2012-09-18 16:41:20 -04:00
David Cole 286f7c9644 Merge topic 'topics/FindCUDA/Add-CUDA_SOURCE_PROPERTY_FORMAT'
787287c Added CUDA_SOURCE_PROPERTY_FORMAT. Allows setting per file format (OBJ or PTX)
2012-08-24 14:24:11 -04:00
David Cole 6dab13c56e Merge topic 'topics/FindCUDA/Fix-g3-again'
a20c41d Replace -g3 with -g for CUDA 4.1 and 4.2 in addition to CUDA < 3.0.
2012-08-24 14:23:21 -04:00
James Bigler 787287cc46 Added CUDA_SOURCE_PROPERTY_FORMAT. Allows setting per file format (OBJ or PTX)
Added CUDA_SOURCE_PROPERTY_FORMAT that allows users to set the format (OBJ or PTX) on a
per file basis.
2012-08-20 17:33:35 -06:00
James Bigler db207e4065 FindCUDA: Added CUDA_HOST_COMPILER variable.
Added a new CUDA variable for specifying the CUDA_HOST_COMPILER.  This will allow users to
be able to specify which host compiler to use for invoking NVCC with.  By default it will
use the compiler used for host compilation.  This is convenient for when you want to
specify a different compiler than the default compiler.  You end up using the same
compiler for both the NVCC compilation and the host compilation instead of using the
default compiler in the path for NVCC.
2012-08-20 15:14:37 -06:00
Rolf Eike Beer b04650e163 use the find_* functions ENV parameter
Instead of directly passing $ENV{SOMEVAR} to a find_* call pass in ENV SOMEVAR.
This will make sure the paths will get correctly handled through different
platforms, especially on Windows.

Also fixes one place where paths with windows delimiters (\) were hardcoded to
use forward slashes.
2012-08-14 21:02:58 +02:00
Kitware Robot 9db3116226 Remove CMake-language block-end command arguments
Ancient versions of CMake required else(), endif(), and similar block
termination commands to have arguments matching the command starting the
block.  This is no longer the preferred style.

Run the following shell code:

for c in else endif endforeach endfunction endmacro endwhile; do
    echo 's/\b'"$c"'\(\s*\)(.\+)/'"$c"'\1()/'
done >convert.sed &&
git ls-files -z -- bootstrap '*.cmake' '*.cmake.in' '*CMakeLists.txt' |
egrep -z -v '^(Utilities/cm|Source/kwsys/)' |
egrep -z -v 'Tests/CMakeTests/While-Endwhile-' |
xargs -0 sed -i -f convert.sed &&
rm convert.sed
2012-08-13 14:19:16 -04:00
Kitware Robot 77543bde41 Convert CMake-language commands to lower case
Ancient CMake versions required upper-case commands.  Later command
names became case-insensitive.  Now the preferred style is lower-case.

Run the following shell code:

cmake --help-command-list |
grep -v "cmake version" |
while read c; do
    echo 's/\b'"$(echo $c | tr '[:lower:]' '[:upper:]')"'\(\s*\)(/'"$c"'\1(/g'
done >convert.sed &&
git ls-files -z -- bootstrap '*.cmake' '*.cmake.in' '*CMakeLists.txt' |
egrep -z -v '^(Utilities/cm|Source/kwsys/)' |
xargs -0 sed -i -f convert.sed &&
rm convert.sed
2012-08-13 14:19:16 -04:00
Rolf Eike Beer 188b96b100 fix some typos 2012-06-19 21:05:14 +02:00
James Bigler a20c41d032 Replace -g3 with -g for CUDA 4.1 and 4.2 in addition to CUDA < 3.0.
CUDA 4.1 reintroduced the bug where -g3 breaks compilation.  This
wasn't fixed until CUDA 5.0, so both CUDA 4.1 and 4.2 have this bug.
2012-06-19 10:04:51 -06:00
Rolf Eike Beer 7ec2ebdbcf fix the same typos as found by Debian in other places, too 2012-02-18 16:13:34 +01:00
James Bigler 154f53738f Added support for curand, cusparse, npp, nvcuenc and nvcuvid libraries.
This addresses Bug 11882 which provided a sample implementation for adding
support for cusparse.  I went ahead and added all the libraries I thought
appropriate.
2011-12-20 14:20:04 -07:00
David Cole 0ea95b99ce Merge topic 'topics/FindCUDA/Misc-fixes'
aa36082 Miscellaneous fixes.
2011-12-07 16:45:35 -05:00
David Cole c26e28754c Merge topic 'topics/FindCUDA/Multi-dir-clash'
80e279d Make CUDA working directory unique for each target.
2011-12-07 16:45:00 -05:00
David Cole 9f18f64c7c Merge topic 'topics/FindCUDA/Quote-fixes'
c3c7a0c Fixes for handling quotes in args and other places (Fix Bug 11726 and 12099).
2011-12-06 15:08:12 -05:00
David Cole 35017df53c Merge topic 'topics/FindCUDA/linux-double-build'
aa59544 Reset dependency file list when a dependency disappeared.
2011-12-06 15:07:48 -05:00
James Bigler aa36082a2b Miscellaneous fixes. 2011-12-05 19:04:48 -07:00
James Bigler 80e279d37c Make CUDA working directory unique for each target.
This allows you to have more than source file with the same name but different
directories.  The intermediate and configuration files are now in this same directory.
2011-12-05 19:00:00 -07:00
James Bigler c3c7a0cfb8 Fixes for handling quotes in args and other places (Fix Bug 11726 and 12099). 2011-12-05 17:16:05 -07:00
James Bigler aa59544078 Reset dependency file list when a dependency disappeared.
Fix a long outstanding bug when a file in the dependency list wasn't found.  This bug
wouldn't reset the dependencies, so the makefile would still want the missing file when
building.  The work around was to configure twice, but this is no longer necessary.
2011-12-05 16:29:28 -07:00
James Bigler 8930938351 Added support for CUDA_PATH which is present in the CUDA toolkit 3.2 onward.
This required changing how the paths were used.  I now use the PATH_SUFFIXES parameter
instead of putting the whole path in the command.
2011-12-05 16:04:00 -07:00
Brad King c4275592a8 Modules: Include builtin FindPackageHandleStandardArgs directly
The FindPackageHandleStandardArgs module was originally created outside
of CMake.  It was added for CMake 2.6.0 by commit e118a627 (add a macro
FIND_PACKAGE_HANDLE_STANDARD_ARGS..., 2007-07-18).  However, it also
proliferated into a number of other projects that at the time required
only CMake 2.4 and thus could not depend on CMake to provide the module.
CMake's own find modules started using the module in commit b5f656e0
(use the new FIND_PACKAGE_HANDLE_STANDARD_ARGS in some of the FindXXX
modules..., 2007-07-18).

Then commit d358cf5c (add 2nd, more powerful mode to
find_package_handle_standard_args, 2010-07-29) added a new feature to
the interface of the module that was fully optional and backward
compatible with all existing users of the module.  Later commit 5f183caa
(FindZLIB: use the FPHSA version mode, 2010-08-04) and others shortly
thereafter started using the new interface in CMake's own find modules.
This change was also backward compatible because it was only an
implementation detail within each module.

Unforutnately these changes introduced a problem for projects that still
have an old copy of FindPackageHandleStandardArgs in CMAKE_MODULE_PATH.
When any such project uses one of CMake's builtin find modules the line

  include(FindPackageHandleStandardArgs)

loads the copy from the project which does not have the new interface!
Then the including find module tries to use the new interface with the
old module and fails.

Whether this breakage can be considered a backward incompatible change
in CMake is debatable.  The situation is analagous to copying a standard
library header from one version of a compiler into a project and then
observing problems when the next version of the compiler reports errors
in its other headers that depend on its new version of the original
header.  Nevertheless it is a change to CMake that causes problems for
projects that worked with previous versions.

This problem was discovered during the 2.8.3 release candidate cycle.
It is an instance of a more general problem with projects that provide
their own versions of CMake modules when other CMake modules depend on
them.  At the time we resolved this instance of the problem with commit
b0118402 (Use absolute path to FindPackageHandleStandardArgs.cmake
everywhere, 2010-09-28) for the 2.8.3 release.

In order to address the more general problem we introduced policy
CMP0017 in commit db44848f (Prefer files from CMAKE_ROOT when including
from CMAKE_ROOT, 2010-11-17).  That change was followed by commit
ce28737c (Remove usage of CMAKE_CURRENT_LIST_DIR now that we have
CMP0017, 2010-12-20) which reverted the original workaround in favor of
using the policy.  However, existing project releases do not set the
policy behavior to NEW and therefore still exhibit the problem.

We introduced in commit a364daf1 (Allow users to specify defaults for
unset policies, 2011-01-03) an option for users to build existing
projects by adding -DCMAKE_POLICY_DEFAULT_CMP0017=NEW to the command
line.  Unfortunately this solution still does not allow such projects to
build out of the box, and there is no good way to suggest the use of the
new option.

The only remaining solution to keep existing projects that exhibit this
problem building is to restore the change originally made in commit
b0118402 (Use absolute path to FindPackageHandleStandardArgs.cmake
everywhere, 2010-09-28).  This also avoids policy CMP0017 warnings for
this particular instance of the problem the policy addresses.
2011-01-20 10:56:49 -05:00
Alex Neundorf ce28737c93 Remove usage of CMAKE_CURRENT_LIST_DIR now that we have CMP0017
This puts the new search behaviour for included files in action, i.e.
now when a file from Modules/ include()s another file, it also gets the
one from Modules/ included, i.e. the one it expects.

Alex
2011-01-04 08:20:08 -05:00
David Cole b0fb2ad3e1 Merge topic 'AddCMAKE_CURRENT_LIST_DIR'
b011840 Use absolute path to FindPackageHandleStandardArgs.cmake everywhere
41e4f1a Add automatic variable CMAKE_CURRENT_LIST_DIR(dir of CMAKE_CURRENT_LIST_FILE)
f9fc79c Remove trailing whitespace
2010-10-19 15:53:16 -04:00
Brad King ba3064b584 Merge topic 'FindCUDA-allow-g3'
6916f8d Allow -g3 for CUDA v3.0+.
2010-10-05 15:33:12 -04:00
James Bigler 6916f8dba7 Allow -g3 for CUDA v3.0+.
In versions of the CUDA toolkit previous to version 3.0 the use of -g3 would cause
compilation errors.  This was fixed in version 3.0.
2010-10-04 16:42:21 -06:00
James Bigler 1df8516c0e Fix for bug 0011263.
Added CUSDKCOMPUTE_ROOT to the list of paths when looking for CUDA_SDK_ROOT_DIR.
2010-10-04 16:11:08 -06:00
Alex Neundorf b01184022b Use absolute path to FindPackageHandleStandardArgs.cmake everywhere
This is to avoid getting an (older) copy of FPHSA.cmake which is
e.g. installed with KDE 4.5.0 and 4.5.1.

Alex
2010-09-28 22:30:31 +02:00
James Bigler 4f0636e979 Added CUDA 3.2 directory changes. Disable emulation mode for CUDA 3.1+.
CUDA 3.2 on Windows systems changed the layout of the library paths.  This adds the extra
directories needed to locate the new files.

I also explicitly disable emulation mode for CUDA 3.1+.  This feature was deprecated in
3.0 and ultimately removed in 3.1.  The script errors out if CUDA_BUILD_EMULATION is
turned on.  I didn't want to ignore emulation mode (even with a warning - which most
people may not even see) and have users confused as to why it wasn't working.
2010-09-10 11:48:38 -06:00
Alex Neundorf 656cd2fad2 Improved version checking for FindCUDA using the new mode of FPHSA
Alex
2010-08-07 23:09:14 +02:00
Kai Wasserbäch 9203e9187e Fix spelling errors reported by Lintian.
During a Lintian run on the binary packages of CMake in Debian I was
notified of many spelling mistakes.
2010-07-13 09:41:37 -04:00
Brad King 4ac531487b Merge branch 'CudaRTEmuLibraryForCUDA30' 2010-06-24 10:43:05 -04:00
James Bigler bb6acb8667 Add support for the emulation version of the cudart library.
In version 3.0 of the CUDA toolkit when building code for emulation, you need to link
against a new version of the cuda run time library called cudartemu.  This CL adds a check
for the new library and uses it when present and in emulation mode.  Note that this
library is not present in previous or subsequent versions of the CUDA toolkit.
2010-06-21 17:11:57 -06:00
James Bigler 1d5554201f CUDA_VERSION variable passed to REGEX needs quotes to work when not defined. 2010-06-15 10:03:40 -06:00
James Bigler 0d30e3fe91 Fixed: CUDA_VERSION_MAJOR/MINOR now computed after first run.
CUDA_VERSION_MAJOR and CUDA_VERSION_MINOR were only computed when CUDA_VERSION was first
computed.  Subsequent runs of FindCUDA would not have CUDA_VERSION_MAJOR/MINOR set.  We
now extract the major and minor versions from the CUDA_VERSION cache variable every run.
2010-05-12 16:38:51 -06:00
James Bigler aa495ad021 Add -rpath for cuda libraries on Apple. They use @rpath in the library link names. 2010-01-28 13:50:38 -05:00
James Bigler f04f912694 Updated the documentation on what kinds of flags you can use when setting target properties. 2010-01-28 13:08:56 -05:00
James Bigler 6faa4ae15e Changed warning string to use the WARNING flag. 2010-01-08 12:57:43 -05:00
James Bigler d3034f9b56 Disable the --host-compilation flag for CUDA >= 3.0 since it is derecated. 2010-01-08 12:12:54 -05:00
James Bigler bbfcb198f8 Fix #9970: Use execute_process instead of deprecated exec_program. 2009-12-01 16:28:26 -05:00
James Bigler b1817590c1 Fix -fPIC from being used on executable object files.
BUILD_SHARED_LIBS is now only recognized when calling CUDA_ADD_LIBRARY.  If you want the CMAKE_SHARED_LIBRARY_C/CXX_FLAGS to be used, pass SHARED as an argument.  This prevents -fPIC from being used on objects destined for executables by default.
2009-11-04 01:15:31 -05:00
James Bigler 9605af8c2e Look for nvcc in the 32 bit bin directory before the 64 bin directory. 2009-10-27 12:00:28 -04:00
James Bigler 94fff7af35 Fix Xcode build.
Move the make_directory command to the main target, so that CMAKE_CFG_INTDIR
will get expanded by the build tool.
2009-10-06 22:52:52 -04:00
James Bigler 7112227c18 Updated copyright notice to conform to NVIDIA guidelines. 2009-10-05 17:14:33 -04:00
James Bigler c66fb80171 Removed support for cutil library and header file.
Cutil was never intented to be used outside of the SDK.  The removal of this
code is in support of this.  The CUDA_SDK_ROOT_DIR will continue to be
supported, in case users wish to use this to find files in the SDK.  There are
also two examples of how to use CUDA_SDK_ROOT_DIR to find header files and
libraries if users so wish.
2009-09-28 23:41:40 -04:00
James Bigler a26dac2202 Fixed CUDA_PROPAGATE_HOST_FLAGS, added path for Mac SDK.
The CUDA_PROPAGATE_HOST_FLAGS was incorrect in that it prevented the CUDA_NVCC_FLAGS_CONFIG variable from getting filled.

Also, added a search path for the CUDA SDK install on Macs.
2009-09-28 23:17:13 -04:00
Brad King 1a527daa3b Convert newlines from CRLF to LF
These files were committed to the repository with Windows newlines.
This converts them to Unix newlines so they will show up natively.
2009-09-25 15:42:53 -04:00
James Bigler d468145fa2 Documentation fixes, new CUDA_PROPAGATE_HOST_FLAGS, changed output directory.
- Finished updating and formatting documentation.
- Added CUDA_PROPAGATE_HOST_FLAGS (Default ON) that can disable the C flag
propagation to the host compiler.
_ Changed the output directory for support files from
${CMAKE_CURRENT_BINARY_DIR} to ${CMAKE_CURRENT_BINARY_DIR}/CMakeFiles.  This
will hopefully reduce the clutter in the binary directory.
2009-09-25 00:28:12 -04:00
James Bigler b7142e9214 Updated formatting of documentation plus a little reorganization. 2009-09-23 01:01:51 -04:00
James Bigler eaaf71d7b3 Initial version of FindCUDA script. Still needs documentation formatting. 2009-09-15 02:38:20 -04:00