Research topic for parallel programming

I have to publish a small/local research paper on parallel programming as a course requirement.I have knowledge of both MPI and Open CL.I want to do some comparison type thing.Like write some program and run it on both.I want to complete my work in the next 3 to 5 days.(I am free all day).Can you make some suggestions?
The problem should be easy and small enough to be implemented in the next 3 to 5 days but good enough to get a publication.
Please give me some suggestions.
My idea is Proving open cl is better than MPI.If I am in the right direction then what type of things should I do to compare both.If I am writing programs then what type of programs should I write to compare both MPI and Open CL.Please help me.
Or If you have a better suggestion then please help me. I would prefer working on my own idea.Please tell me how do I approach it to make it publishable?

I think if this is part of a course, then selecting the algorithm to demonstrate is part of the graded material too - because the answer to the proposition of “proving opencl is better than mpi” is not true or false, but depends on the problem.

You should probably ask your supervisor for direction and not the internet.