My platform:
GPU: GeForce 8800GT
OS: Ubuntu 10.04 x64
CUDA Driver: 195.36.15
CUDA Toolkit: 3.0
CUDA SDK: 3.0
I had installed the CUDA SDK/Toolkit 3.0 and had try to build the samples from SDK, all done right, I can run all the samples now.
and then, I download a CUDA benchmark source code and try to build it ( of course, this benchmark is not worten by me, I am just interesting) :
http://www.eecg.toronto.edu/~moshovos/CUDA…robenchmark.zip
this zip file was extract to “/home/edison/NVIDIA_GPU_Computing_SDK/projects/microbenchmark”
I had modify the “ROOTDIR=” of the “Makefile” to “/home/edison/NVIDIA_GPU_Computing_SDK/projects” ;
and replace the “common.mk” with the one that I success build for SDK samples ;
and copy the cutil.h to “/home/edison/NVIDIA_GPU_Computing_SDK/projects/microbenchmark” ;
and copy the folders: “common”, “lib” to “/home/edison/NVIDIA_GPU_Computing_SDK/projects/microbenchmark”
and type “make”, the building is running but after a few minutes, i get such error info:
edison@edison-desktop://home/edison/NVIDIA_GPU_Computing_SDK/projects/microbenchmark$ make
chmod a+x build_cubin
/usr/bin/ld: cannot find -lcutil_x86_64
collect2: ld returned 1 exit status
make: *** [bin/linux/release/main] Error 1
so, how can I fix this “/usr/bin/ld: cannot find -lcutil_x86_64” problem ?
thanks !