<div dir="ltr"><div>Peter,</div><div><br></div>I can't reproduce the problem with recent clang. I've tried compiling the same smaple code with cuda-7.0 and 7.5.<div>clang version 3.9.0 (trunk 268962) (llvm/trunk 268980)<br></div><div><br></div><div>Could you tell me what was clang version you used? If you still see the problem, please file a clang bug on <a href="http://llvm.org/bugs">llvm.org/bugs</a>.</div><div><br></div><div>--Artem</div><div><div><div><br></div><div><br></div><div><div>% clang++ -I../../common/inc --cuda-gpu-arch=sm_35 simpleCUFFT.cu -L/usr/local/cuda-7.5/lib64 -lcufft -lcudart -o simpleCUFFT-clang</div><div>% LD_LIBRARY_PATH=/usr/local/cuda-7.5/lib64 ./simpleCUFFT-clang<br></div><div>[simpleCUFFT] is starting...</div><div>GPU Device 0: "Tesla K40c" with compute capability 3.5</div><div><br></div><div>Transforming signal cufftExecC2C</div><div>Launching ComplexPointwiseMulAndScale<<< >>></div><div>Transforming signal back cufftExecC2C</div><div><br></div></div></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Apr 5, 2016 at 2:12 PM, Jingyue Wu <span dir="ltr"><<a href="mailto:jingyue@google.com" target="_blank">jingyue@google.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">Would you mind upload your simpleCUFFT.cu code? It looks related to device code generation because building was successful. </div><div class="HOEnZb"><div class="h5"><div class="gmail_extra"><br><div class="gmail_quote">On Tue, Apr 5, 2016 at 1:56 PM, Peter Steinbach <span dir="ltr"><<a href="mailto:steinbac@mpi-cbg.de" target="_blank">steinbac@mpi-cbg.de</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi guys,<br>
<br>
first of all, please accept my apologies for contacting you by mail. I was a bit lost, which mailing list to choose from as pointed to by<br>
<a href="http://llvm.org/docs/CompileCudaWithLLVM.html" rel="noreferrer" target="_blank">http://llvm.org/docs/CompileCudaWithLLVM.html</a><br>
and the subsequent<br>
<a href="http://llvm.org/docs/#mailing-lists" rel="noreferrer" target="_blank">http://llvm.org/docs/#mailing-lists</a><br>
Feel free to deflect this request to the relevant mailing list or bug tracker.<br>
<br>
In any case, I am very interested in using GPUCC in favor of NVCC for a multitude of reasons (C++1X, compilation speed, ...). I started to "port" my favorite samples from the nvidia SDK.<br>
With clang 3.8, samples-7.0.28/7_CUDALibraries/simpleCUFFT as compiled with clang produces an error at runtime! Here is what I see with a K20c:<br>
<br>
$ clang++ --cuda-path=/sw/apps/cuda/7.0.28 -I../../common/inc -m64 --cuda-gpu-arch=sm_35 --cuda-gpu-arch=sm_35 -o simpleCUFFT.o -c simpleCUFFT.cu<br>
$ clang++ --cuda-path=/sw/apps/cuda/7.0.28 -L/sw/apps/cuda/7.0.28/lib64 -lcudart -ldl -lrt -pthread -m64 -o simpleCUFFT.llvm simpleCUFFT.o -lcufft<br>
$ ./simpleCUFFT.llvm<br>
[simpleCUFFT] is starting...<br>
GPU Device 0: "Tesla K20c" with compute capability 3.5<br>
<br>
Transforming signal cufftExecC2C<br>
Launching ComplexPointwiseMulAndScale<<< >>><br>
simpleCUFFT.cu(132) : getLastCudaError() CUDA error : Kernel execution failed [ ComplexPointwiseMulAndScale ] : (8) invalid device function.<br>
<br>
The same source code works just fine with nvcc 7.0.<br>
Any help would be appreciated.<br>
<br>
Best,<br>
Peter<br>
<br>
PS. From random comments, I had the feeling that you are looking at the SHOC benchmarks with gpucc. If so, please comment on:<br>
<a href="https://github.com/vetter/shoc/issues/48" rel="noreferrer" target="_blank">https://github.com/vetter/shoc/issues/48</a><br>
I don't wanna do work that is either pointless (support for textures) or was already done. ;)<br>
--<br>
Peter Steinbach, Dr. rer. nat.<br>
HPC Developer, Scientific Computing Facility<br>
<br>
Max Planck Institute of Molecular Cell Biology and Genetics<br>
Pfotenhauerstr. 108<br>
01307 Dresden<br>
Germany<br>
<br>
<br>
phone <a href="tel:%2B49%20351%20210%202882" value="+493512102882" target="_blank">+49 351 210 2882</a><br>
fax <a href="tel:%2B49%20351%20210%201689" value="+493512101689" target="_blank">+49 351 210 1689</a><br>
<a href="http://www.mpi-cbg.de" rel="noreferrer" target="_blank">www.mpi-cbg.de</a><br>
</blockquote></div><br></div>
</div></div></blockquote></div><br><br>
</div></div>