Cuda Parallelism question

Hey I have a question about Cuda parallelism, I am doing about 1,000 tests on 500 data sets and all the tests are done on each of the data sets and they are the same tests.

Could that be done in parallel computing?
All may tests are done in pandas and numpy.