After installing my GTX260 card I am looking forward to running CUDA software.
To start, I would like to compile the sample code. Using this code, and I get a customed to the syntax of the CUDA routines and start writing code myself.
However, when running make, I encounter the problem:
make[1]: Entering directory /home/lukas/NVIDIA_GPU_Computing_SDK/C/common' make[1]: Leaving directory /home/lukas/NVIDIA_GPU_Computing_SDK/C/common’
make[1]: Entering directory /home/lukas/NVIDIA_GPU_Computing_SDK/C/common' make[1]: Leaving directory /home/lukas/NVIDIA_GPU_Computing_SDK/C/common’
make[1]: Entering directory /home/lukas/NVIDIA_GPU_Computing_SDK/C/common' make[1]: Leaving directory /home/lukas/NVIDIA_GPU_Computing_SDK/C/common’
make -C src/simpleStreams/
make[1]: Entering directory /home/lukas/NVIDIA_GPU_Computing_SDK/C/src/simpleStreams' nvcc fatal : A single input file is required for a non-link phase when an outputfile is specified make[1]: *** [obj/release/simpleStreams.cu.o] Error 255 make[1]: Leaving directory /home/lukas/NVIDIA_GPU_Computing_SDK/C/src/simpleStreams’
make: *** [src/simpleStreams/Makefile.ph_build] Error 2
after which the compiling ends. Google reveals remarkable little about this problem.
A possible cause is me running Ubuntu 9.10, which is officially not yet supported.
To fix the problem with the gcc 4.4 compatibility I have changed common.mk:
CXX := g+±4.3
CC := gcc-4.3
LINK := g+±4.3 -fPIC
But the same error message occurs.
There is a notion that the line:
Includes
INCLUDES += “-I. -I$(CUDA_INSTALL_PATH)/include -I$(COMMONDIR)/inc”
is incorrect, because it would assign multiple input files, but I do not understand what could go wrong here.
What exactly is a ‘non-link phase’?
Has anybody experience this problem before? And how was it solved?
Thank you very much for your time and advice.
Lukas