Thesis CPU + GPU: which problem to solve?

Question

IT education

Good day. I plan to write my thesis soon. The topic is related to writing something parallel on the tandem CPU + GPU (NVidia). Plus, the condition that the system must be distributed is also taxed, that is, a network of N computers with powerful video cards and processors. I worked separately with both CUDA and OpenMP, but I still don’t know how to connect it (especially since the system is still distributed). I'm actually interested in this topic, but I'd like some advice on a task (or type of task) that can do it well.

Answer 1

Factorization in integers.