I’m absolutely rookie.
I installed the “Nvidia driver” and Cuda Toolkit in Ubuntu 14 with the runfile installer .
Everything seems fine, but when trying to verify the installation compiling an example, those included in the toolkit, I do something wrong.
I am unable to compile the examples, I get an error that I can not interpret.
After using cd to put myself in the appropriate directory ( I believe) I run the command make
clite @ pcubuntu: / usr / local / cuda / samples / 0_Simple / inlinePTX $ make
I copy the result (the same error if I try the compilation on cuda-7.0):
What file or directory or permissions are missing ?. Sorry for this question, probably, trivial.
Thanks in advance.
perhaps the next helps to identify the error:
nvcc -V gives me: tools Cuda compilation, release 7.0, V7.0.27
my PATH: /usr/local/cuda-7.0/bin:/home/clite/anaconda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games
my LD_LIBRARY_PATH: /usr/local/cuda-7.0/lib64
Thanks for this quick reply.
my first attempt fails. I copy entirely the terminal output and translate the Spanish error to English ((in double parentheses))
root@pcubuntu:/usr/local/cuda/samples# make
make[1]: se ingresa al directorio «/usr/local/cuda-7.0/samples/0_Simple/inlinePTX»
“/usr/local/cuda-7.0”/bin/nvcc -ccbin g++ -I…/…/common/inc -m64 -gencode arch=compute_20,code=sm_20 -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_52,code=compute_52 -o inlinePTX.o -c inlinePTX.cu
g++: No such file or directory
make[1]: *** [inlinePTX.o] Error 1
make[1]: se sale del directorio ((my translation: IT EXITS THE DIRECTORY)) «/usr/local/cuda-7.0/samples/0_Simple/inlinePTX»
make: *** [0_Simple/inlinePTX/Makefile.ph_build] Error 2
root@pcubuntu:/usr/local/cuda/samples#
Another try on one more deep dir
root@pcubuntu:/usr/local/cuda/samples/0_Simple/inlinePTX# make
“/usr/local/cuda-7.0”/bin/nvcc -ccbin g++ -I…/…/common/inc -m64 -gencode arch=compute_20,code=sm_20 -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_52,code=compute_52 -o inlinePTX.o -c inlinePTX.cu
g++: No such file or directory
make: *** [inlinePTX.o] Error 1
root@pcubuntu:/usr/local/cuda/samples/0_Simple/inlinePTX#
My interpretation of the message, was totally wrong. The message is simply saying that g ++ is not installed.!
What a shame!
After installing g ++, make runs without errors and finished building CUDA samples.