In my university, I got some lessons about CUDA. However, all students are testing their code on the same GPU (GeForce GTX 580), so we get about 20 tests at the same time. When no one tests, my code works well but when everyone test, it’s a kind of mess.
Is it possible that the GPU do some bad processing when everyone test on the same GPU?
Is there some controls on the execution which come from different users?
Thanks for your answers