Please read the instructions and convert my program to MPI with parallel file IO
The task is to convert this sequential program to MPI along with adding parallel file io to it.
The parallel file io is a MUST and can ONLY use MPI.
I tried converting the program myself to mpi(non parallel file io) and made it twice as slow. So the parallel file io is important.
The program overall takes in a deck of cards (52) as .pgm files rotates them at an arbitrary angle and writes them to the output folder. Bilinear interpolation needs to be used for the holes.
I have provided my bad c code.
I will have access to a cluster with 1-15 nodes. So hopefully we can get speed up. My linux machine takes ~120 seconds to read/rotate.write the cards
Categories:
Tags:
Comments are closed