One test with multiple sequences or multiple tests with single/multiple sequences in UVM?

I wanted to know what is a more efficient methodology/guidelines for UVM when it comes to block-level verification?

  • one test with multiple sequences
  • multiple tests with single/multiple sequences.

In my past projects, I used to have both approaches to align with legacy codes but in general, I want to know if there are any pros and cons of both methods and which one is more suited.

In reply to supal:

There is no clear answer. It depends on your application and how your sequences has been coded. If you have a lot of small sequences it might be a good idea to run several sequences in 1 test. If you are running multiple test with small sequences you might end-up in low verification quality becaus eyou do not verify the big picture but only small snippets.