hcarter108 / cisc372_proj Goto Github PK

Instructions for compiling and running code

Assumes the user has gcc and pgi compilers

-------------------------------------------------------------------

For OpenACC multicore or gpu, do at least the following to compile:

pgcc -ta=<target_accelerator> acc_single_gpu.c

e.g. for multicore

pgcc -ta=multicore acc_single_gpu.c

e.g. for GPU

pgcc -ta=tesla acc_single_gpu.c

Then to run the compiled code with the proper arguments do:

#Where output.file is where we want to store output
./a.out size1 size2 num_iterations output.file

e.g. with particular sizes and number of iterations 
./a.out 256 256 10000 out.txt

To change the number of threads to use to n, run the following:
export ACC_NUM_CORES=n

-------------------------------------------------------------------

For OpenMP multicore at least do this to compile:

gcc -fopenmp heat_omp.c

Then to run the compiled code it would be the same as above, e.g.
./a.out 256 256 10000 out.txt

To change the number of threads to use to n, run the following:
export OMP_NUM_THREADS=n