launch_main.sh

#!/bin/bash

### Options to run the test on the Marzola cluster ###

#SBATCH --job-name=homework-2-run
#SBATCH --output=output.out
#SBATCH --error=error.err

#SBATCH --partition=edu5
#SBATCH --nodes=1 
#SBATCH --gres=gpu:1
#SBATCH --ntasks-per-node=1
#SBATCH --cpus-per-task=1



# DESCRIPTION : 
# "homework-2" transpose the input matrix with transpose_blocks_gpu() or transpose_blocks_gpu_coalesced() kernels & shows the related results

# OUTPUT :
# prompts details about the run on the terminal and computes 2 metrics :
#  - execution time : time to compute the transpose() function
#  - effective bandwidth : estimated bandwidth of transpose() function


### User Variables ###

method="blocks_coalesced"       # chose which version of transpose to run between "transpose_blocks_gpu()" and "transpose_blocks_gpu_coalesced()" by setting :
                                #   method="blocks_naive" : run transpose_blocks_gpu() 
                                #   method="blocks_coalesced" : run transpose_blocks_gpu_coalesced()

N=12  # defines the matrix size : ( 2^N x 2^N )

block_size=5  # in method transpose_blocks() defines the block size : ( 2^block_size x 2^block_size )

th_size_x=3  # defines threadBlock shape ( 2^th_size_x x 2^th_size_y )
th_size_y=5

./bin/homework-2 --method=$method --N=$N --block_size=$block_size --th_size_x=$th_size_x --th_size_y=$th_size_y