Error using Atomic functions

Accelerated Computing CUDA CUDA Programming and Performance
1

Hello everyone,
I made the following changes in makefile to use the atomic functions in my cuda code.

NVCC = /usr/local/cuda/bin/nvcc
CUDAPATH = /usr/local/cuda
NVCCFLAGS = -I$(CUDAPATH)/include -arch=sm_11,
LFLAGS = -L$(CUDAPATH)/lib64 -lcuda -lcudart -lm

file :$(NVCC) $(NVCCFLAGS) $(LFLAGS) -o file file.cu

But i get an error as Value ‘sm_11,’ is not defined for option ‘gpu-architecture’.

I tried giving compute_20 instead of sm_11… but still i am facing the error.
Can somebody help?

2

Don’t stick the comma after sm_11:
change this:
NVCCFLAGS = -I$(CUDAPATH)/include -arch=sm_11,

to this:

NVCCFLAGS = -I$(CUDAPATH)/include -arch=sm_11