Hi Robert Find attached a test case that reproduces the error.
Jamil
On Fri, 3 Nov 2017 at 21:00 Jamil Appa <jamil.a...@zenotech.com> wrote:
> Hi Robert
>
> I will put a minimal case together using one of the CUDA samples and
> sent it through for testing with the new cmake support.
>
> On your second point, CUDA_CUBLAS_LIBRARIES should be empty if no
> cublas libraries are present so that use case should be automatically
> handled unless I am missing something.
>
> I will create a fork and topic branch with the change and submit it for
> review.
>
> Thanks
>
> Jamil
>
>
> On Fri, 3 Nov 2017 at 18:08 Robert Maynard <robert.mayn...@kitware.com>
> wrote:
>
>> HI Jamil,
>>
>> If you are interested in submitting a fix I can review, but I would like
>> to also make sure that this issue doesn't exist inside the new cmake
>> support for CUDA as a first class language. Can you produce a minimal test
>> case that we can rework to use the new support and add to CMake's test
>> suite.
>>
>> You will also need to track down what version of CUDA first added
>> cublas_device
>> so that we properly guard the inclusion.
>>
>> On Fri, Nov 3, 2017 at 7:39 AM, Jamil Appa <jamil.a...@zenotech.com>
>> wrote:
>>
>>> Hi Robert
>>>
>>> I am using CUDA_ADD_CUBLAS_TO_TARGET but the error happens during the
>>> prelink phase when using separable compilation. (nvcc -dlink phase)
>>>
>>> My guess is that Nvidia have not included cublas_device.a in their
>>> dlink phase as you don't have to provide any other libraries. I have asked
>>> the Nvidia team if that is the case, but no reply yet.
>>>
>>> The general fix seems to be to modify lines 1759/1772 of
>>> FindCUDA.cmake (in gitlab master) to include CUBLAS_LIBRARIES
>>>
>>> COMMAND ${CUDA_NVCC_EXECUTABLE} ${nvcc_flags} -dlink ${object_files}
>>> ${CUDA_CUBLAS_LIBRARIES} -o ${output_file}
>>>
>>> This will have no impact if the cublas symbols are not used in the
>>> object files.
>>>
>>> If you are happy with this I can submit the change for review.
>>>
>>> Jamil
>>>
>>>
>>> On Thu, 2 Nov 2017 at 13:53 Robert Maynard <robert.mayn...@kitware.com>
>>> wrote:
>>>
>>>> Hi,
>>>>
>>>> Have you tried using the 'CUDA_ADD_CUBLAS_TO_TARGET' command on the
>>>> target that has FindCUDA separable compilation enabled?
>>>>
>>>> On Wed, Nov 1, 2017 at 11:16 AM, Jamil Appa <jamil.a...@zenotech.com>
>>>> wrote:
>>>>
>>>>> Hi
>>>>>
>>>>> When using FindCUDA with separable compilation combined with device
>>>>> code that uses cublas_device I get missing symbols when linking the
>>>>> intermediate file at line 1758
>>>>> <https://gitlab.kitware.com/cmake/cmake/blob/release/Modules/FindCUDA.cmake#L1758>
>>>>> FindCUDA.cmake
>>>>>
>>>>> If I change line 1758 to include the cublas_device library the
>>>>> symbols are resolved correctly.
>>>>>
>>>>> COMMAND ${CUDA_NVCC_EXECUTABLE} ${nvcc_flags} -dlink
>>>>> ${object_files} -o ${output_file} -lcublas_device
>>>>>
>>>>> Is there a way to pass through a library to
>>>>> the CUDA_LINK_SEPARABLE_COMPILATION_OBJECTS function that I am missing?
>>>>>
>>>>> Thanks
>>>>>
>>>>> Jamil
>>>>>
>>>>> *Jamil Appa* | Co-Founder and Director | Zenotech
>>>>> [image: Papercut]
>>>>> [image: Tel:] +44 (0)7747 606 788 <+44%207747%20606788> [image:
>>>>> Zenotech LTD - Simulation Unlimited] <http://www.zenotech.com/>
>>>>> [image: Email:] jamil.a...@zenotech.com
>>>>> [image: Web:] www.zenotech.com
>>>>> [image: Papercut]
>>>>> [image: linkedin:]
>>>>> <http://uk.linkedin.com/pub/jamil-appa/1/165/120>[image:
>>>>> Twitter:] <https://twitter.com/zenotech>[image: Location:]
>>>>> <https://www.google.co.uk/maps/place/Bristol+%26+Bath+Science+Park/@51.500921,-2.478567,17z/data=!3m1!4b1!4m2!3m1!1s0x48719ab86a5a9f7d:0xd17394f3400abb0a>
>>>>>
>>>>> Company Registration No : 07926926 | VAT No : 128198591
>>>>>
>>>>> Registered Office : 1 Larkfield Grove, Chepstow, Monmouthshire, NP16
>>>>> 5UF, UK
>>>>> <https://maps.google.com/?q=1+Larkfield+Grove,+Chepstow,+Monmouthshire,+NP16+5UF,+UK&entry=gmail&source=g>
>>>>>
>>>>> Address : Bristol & Bath Science Park, Dirac Cres, Emersons Green,
>>>>> Bristol BS16 7FR
>>>>>
>>>>>
>>>>> --
>>>>>
>>>>> Powered by www.kitware.com
>>>>>
>>>>> Please keep messages on-topic and check the CMake FAQ at:
>>>>> http://www.cmake.org/Wiki/CMake_FAQ
>>>>>
>>>>> Kitware offers various services to support the CMake community. For
>>>>> more information on each offering, please visit:
>>>>>
>>>>> CMake Support: http://cmake.org/cmake/help/support.html
>>>>> CMake Consulting: http://cmake.org/cmake/help/consulting.html
>>>>> CMake Training Courses: http://cmake.org/cmake/help/training.html
>>>>>
>>>>> Visit other Kitware open-source projects at
>>>>> http://www.kitware.com/opensource/opensource.html
>>>>>
>>>>> Follow this link to subscribe/unsubscribe:
>>>>> http://public.kitware.com/mailman/listinfo/cmake
>>>>>
>>>>
>>>>
>>
cuda-test.tar
Description: Unix tar archive
-- Powered by www.kitware.com Please keep messages on-topic and check the CMake FAQ at: http://www.cmake.org/Wiki/CMake_FAQ Kitware offers various services to support the CMake community. For more information on each offering, please visit: CMake Support: http://cmake.org/cmake/help/support.html CMake Consulting: http://cmake.org/cmake/help/consulting.html CMake Training Courses: http://cmake.org/cmake/help/training.html Visit other Kitware open-source projects at http://www.kitware.com/opensource/opensource.html Follow this link to subscribe/unsubscribe: http://public.kitware.com/mailman/listinfo/cmake
