[vtkusers] compiling VTK against CUDA 4.2

Tue Jul 31 12:16:42 EDT 2012

I see.  OK, well, thank you both very much for the information!  :-)  Saves me a lot of time.  
-- Rich 

On Jul 31, 2012, at 6:26 AM, David E DeMarle wrote:

> I don't think that file in VTK is meant to be compiled against CUDA as
> it stands. It does appear to have code in it that came from something
> that could be compiled with nvcc but I am guessing those code paths
> are unused in VTK.
> 
> The only part of VTK that I am aware of that uses CUDA is the brand new
> Accelerators/Piston module. It is in current git master (post VTK
> 5.10). Look at it's CMakeLists.txt to see how code that uses CUDA can
> be brought into VTK.
> 
> And I think that Dženan is right. The message you are getting means
> that CMake isn't able to use nvcc directly as a replacement for
> gcc/icc/vs/what_have_you. Compiling CMake itself with nvcc isn't the
> right path to follow.
> 
> David E DeMarle
> Kitware, Inc.
> R&D Engineer
> 21 Corporate Drive
> Clifton Park, NY 12065-8662
> Phone: 518-881-4909
> 
> 
> On Tue, Jul 31, 2012 at 4:12 AM, Dženan Zukić <dzenanz at gmail.com> wrote:
>> Even if you did somehow succeed at building CMake using nvcc, it would make
>> no difference. What that error is saying is: nvcc cannot compile even a
>> simple C Hello World.
>> 
>> I know that efficient interfacing between a scenegraph library and CUDA is
>> not simple.
>> 
>> On Tue, Jul 31, 2012 at 3:18 AM, Cook, Rich <cook47 at llnl.gov> wrote:
>>> 
>>> Hello
>>> I'm new to the list and apologize if this is obvious or seen before.  I
>>> Googled around a bit for the answer but cannot find it.
>>> 
>>> My question seemed simple before I tried it.  How does one compile VTK in
>>> order to use CUDA? A user of mine requested this;  I'm not a VTK expert and
>>> am not sure specifically what she plans to do with the result.
>>> 
>>> I see CUDA stuff like this in the source, making me think this should be
>>> possible and the question at least makes sense to ask:
>>> 
>>> ../VTK/Graphics/vtkYoungsMaterialInterface.cxx:77204:#ifdef __CUDACC__
>>> 
>>> So, I tried a naive approach:
>>> 
>>> export CC=nvcc
>>> export CXX=nvcc
>>> export CPP=nvcc
>>> 
>>> mkdir vtk-build
>>> cd vtk-build
>>> cmake ../VTK
>>> 
>>> cmake fails as follows, making me think maybe I have to build CMake with
>>> nvcc as well?  That would be a bit ugly, but I just want to make sure I'm
>>> going down the right path.
>>> Any help is greatly appreciated.
>>> 
>>> rcook at edge84 (vtk-build): cmake ../VTK/
>>> -- The C compiler identification is GNU
>>> -- The CXX compiler identification is GNU
>>> -- Check for working C compiler: /opt/cudatoolkit-4.2/bin/nvcc
>>> -- Check for working C compiler: /opt/cudatoolkit-4.2/bin/nvcc -- broken
>>> CMake Error at
>>> /collab/usr/global/tools/Kitware/CMake/2.8.7/chaos_5_x86_64_ib/share/cmake-2.8/Modules/CMakeTestCCompiler.cmake:52
>>> (MESSAGE):
>>>  The C compiler "/opt/cudatoolkit-4.2/bin/nvcc" is not able to compile a
>>>  simple test program.
>>> 
>>>  It fails with the following output:
>>> 
>>>   Change Dir: /nfs/tmp2/rcook/VTK/5.10/vtk-build/CMakeFiles/CMakeTmp
>>> 
>>> 
>>> 
>>>  Run Build Command:/usr/bin/gmake "cmTryCompileExec/fast"
>>> 
>>>  /usr/bin/gmake -f CMakeFiles/cmTryCompileExec.dir/build.make
>>>  CMakeFiles/cmTryCompileExec.dir/build
>>> 
>>>  gmake[1]: Entering directory
>>>  `/nfs/tmp2/rcook/VTK/5.10/vtk-build/CMakeFiles/CMakeTmp'
>>> 
>>>  /collab/usr/global/tools/Kitware/CMake/2.8.7/chaos_5_x86_64_ib/bin/cmake
>>> -E
>>>  cmake_progress_report
>>>  /nfs/tmp2/rcook/VTK/5.10/vtk-build/CMakeFiles/CMakeTmp/CMakeFiles 1
>>> 
>>>  Building C object CMakeFiles/cmTryCompileExec.dir/testCCompiler.c.o
>>>  /opt/cudatoolkit-4.2/bin/nvcc -o
>>>  CMakeFiles/cmTryCompileExec.dir/testCCompiler.c.o -c
>>>  /nfs/tmp2/rcook/VTK/5.10/vtk-build/CMakeFiles/CMakeTmp/testCCompiler.c
>>> 
>>>  Linking C executable cmTryCompileExec
>>> 
>>>  /collab/usr/global/tools/Kitware/CMake/2.8.7/chaos_5_x86_64_ib/bin/cmake
>>> -E
>>>  cmake_link_script CMakeFiles/cmTryCompileExec.dir/link.txt --verbose=1
>>> 
>>>  /opt/cudatoolkit-4.2/bin/nvcc
>>>  CMakeFiles/cmTryCompileExec.dir/testCCompiler.c.o -o cmTryCompileExec
>>>  -rdynamic
>>> 
>>>  nvcc fatal : Unknown option 'rdynamic'
>>> 
>>>  gmake[1]: *** [cmTryCompileExec] Error 255
>>> 
>>>  gmake[1]: Leaving directory
>>>  `/nfs/tmp2/rcook/VTK/5.10/vtk-build/CMakeFiles/CMakeTmp'
>>> 
>>>  gmake: *** [cmTryCompileExec/fast] Error 2
>>> 
>>> 
>>> 
>>> 
>>> 
>>>  CMake will not be able to correctly generate this project.
>>> Call Stack (most recent call first):
>>>  CMakeLists.txt:15 (PROJECT)
>>> 
>>> 
>>> 
>>> --
>>> ✐Richard Cook
>>> ✇ Lawrence Livermore National Laboratory
>>> Bldg-453 Rm-4024, Mail Stop L-557
>>> 7000 East Avenue,  Livermore, CA, 94550, USA
>>> ☎ (office) (925) 423-9605
>>> ☎ (fax) (925) 423-6961
>>> ---
>>> Information Management & Graphics Grp., Services & Development Div.,
>>> Integrated Computing & Communications Dept.
>>> (opinions expressed herein are mine and not those of LLNL)
>>> 
>>> 
>>> 
>>> _______________________________________________
>>> Powered by www.kitware.com
>>> 
>>> Visit other Kitware open-source projects at
>>> http://www.kitware.com/opensource/opensource.html
>>> 
>>> Please keep messages on-topic and check the VTK FAQ at:
>>> http://www.vtk.org/Wiki/VTK_FAQ
>>> 
>>> Follow this link to subscribe/unsubscribe:
>>> http://www.vtk.org/mailman/listinfo/vtkusers
>> 
>> 
>> 
>> _______________________________________________
>> Powered by www.kitware.com
>> 
>> Visit other Kitware open-source projects at
>> http://www.kitware.com/opensource/opensource.html
>> 
>> Please keep messages on-topic and check the VTK FAQ at:
>> http://www.vtk.org/Wiki/VTK_FAQ
>> 
>> Follow this link to subscribe/unsubscribe:
>> http://www.vtk.org/mailman/listinfo/vtkusers
>> 

-- 
✐Richard Cook   
✇ Lawrence Livermore National Laboratory
Bldg-453 Rm-4024, Mail Stop L-557        
7000 East Avenue,  Livermore, CA, 94550, USA
☎ (office) (925) 423-9605    
☎ (fax) (925) 423-6961
---
Information Management & Graphics Grp., Services & Development Div., Integrated Computing & Communications Dept.
(opinions expressed herein are mine and not those of LLNL)