Hi,
Im trying to build the nvgraph_SSSP example on MacOSX 10.11 with the latest RC toolkit but I get the error below. I also don’t see any libnvgraph or similar anywhere on the cuda paths…
/D/N/C/samples ❯❯❯ make -C 7_CUDALibraries/nvgraph_SSSP
/Developer/NVIDIA/CUDA-8.0/bin/nvcc -ccbin clang++ -m64 -Xcompiler -arch -Xcompiler x86_64 -Xlinker -rpath -Xlinker /Developer/NVIDIA/CUDA-8.0/lib -gencode arch=compute_30,code=sm_30 -gencode arch=compute_35,code=sm_35 -gencode arch=compute_37,code=sm_37 -gencode arch=compute_50,code=sm_50 -gencode arch=compute_52,code=sm_52 -gencode arch=compute_60,code=sm_60 -gencode arch=compute_60,code=compute_60 -o nvgraph_SSSP nvgraph_SSSP.o -lnvgraph
ld: library not found for -lnvgraph
clang: error: linker command failed with exit code 1 (use -v to see invocation)
make: *** [nvgraph_SSSP] Error 1
Am I missing something?
/D/N/C/samples ❯❯❯ kextstat | grep -i cuda ⏎
177 0 0xffffff7f833be000 0x2000 0x2000 com.nvidia.CUDA (1.1.0) DD792765-CA28-395A-8593-D6837F05C4FF <4 1>
/D/N/C/samples ❯❯❯ /usr/bin/cc --version
Apple LLVM version 7.3.0 (clang-703.0.31)
Target: x86_64-apple-darwin15.5.0
Thread model: posix
InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin
/D/N/C/samples ❯❯❯ nvcc -V
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2016 NVIDIA Corporation
Built on Tue_May_17_12:08:22_CDT_2016
Cuda compilation tools, release 8.0, V8.0.28
/D/N/C/samples ❯❯❯
I haven’t tested on MacOS but on linux, the following are installed in the usual place:
$ ls /usr/local/cuda/lib64
...
libnvgraph.so
libnvgraph.so.8.0
libnvgraph.so.8.0.27
...
I didn’t have any trouble building the nvgraph samples on linux:
$ ls /usr/local/cuda/samples/bin/x86_64/linux/release
...
nvgraph_Pagerank
nvgraph_SemiRingSpmv
nvgraph_SSSP
...
It may have been an omission/oversight in the Mac RC build.
pif… nothing on Mac
/D/N/CUDA-8.0 ❯❯❯ ls /usr/local/cuda/lib*
/usr/local/cuda/lib:
libcublas.8.0.dylib libnppicc_static.a
libcublas.dylib libnppicom.8.0.dylib
libcublas_device.a libnppicom.dylib
libcublas_static.a libnppicom_static.a
libcuda.dylib libnppidei.8.0.dylib
libcudadevrt.a libnppidei.dylib
libcudart.8.0.dylib libnppidei_static.a
libcudart.dylib libnppif.8.0.dylib
libcudart_static.a libnppif.dylib
libcufft.8.0.dylib libnppif_static.a
libcufft.dylib libnppig.8.0.dylib
libcufft_static.a libnppig.dylib
libcufftw.8.0.dylib libnppig_static.a
libcufftw.dylib libnppim.8.0.dylib
libcufftw_static.a libnppim.dylib
libcuinj.8.0.dylib libnppim_static.a
libcuinj.dylib libnppist.8.0.dylib
libculibos.a libnppist.dylib
libcurand.8.0.dylib libnppist_static.a
libcurand.dylib libnppisu.8.0.dylib
libcurand_static.a libnppisu.dylib
libcusolver.8.0.dylib libnppisu_static.a
libcusolver.dylib libnppitc.8.0.dylib
libcusolver_static.a libnppitc.dylib
libcusparse.8.0.dylib libnppitc_static.a
libcusparse.dylib libnpps.8.0.dylib
libcusparse_static.a libnpps.dylib
libnppc.8.0.dylib libnpps_static.a
libnppc.dylib libnvToolsExt.1.dylib
libnppc_static.a libnvToolsExt.dylib
libnppi.8.0.dylib libnvblas.8.0.dylib
libnppi.dylib libnvblas.dylib
libnppi_static.a libnvrtc-builtins.8.0.dylib
libnppial.8.0.dylib libnvrtc-builtins.dylib
libnppial.dylib libnvrtc.8.0.dylib
libnppial_static.a libnvrtc.dylib
libnppicc.8.0.dylib stubs
libnppicc.dylib
/usr/local/cuda/libnsight:
artifacts.xml features nsight plugins
configuration index nsight.app readme
epl-v10.html notice.html p2
/usr/local/cuda/libnvvp:
artifacts.xml epl-v10.html notice.html nvvp.app plugins
configuration features nvvp p2 readme
/D/N/CUDA-8.0 ❯❯❯
I have confirmed that this is a known issue with CUDA 8RC and should be fixed in CUDA 8 production release.