Got a YouTube account?

New: enable viewer-created translations and captions on your YouTube channel!

English subtitles

← Implementation Approach cs348 unit5new

Get Embed Code
1 Language

Showing Revision 1 created 04/13/2013 by Cogi-Admin.

  1. The correct answer is number two, checking at transaction level. We have a high
  2. level TestBench, it's class-based, so we already have transactions. We can
  3. easily observe using monitors, what went in and out of a device, and we could
  4. write our checking based on that. The first choice would require us to write
  5. our TestBench at the bit level. So, we would basically be repeating the
  6. hardware model in our TestBench. This creates a lot of work and a lot
  7. complexity, and we don't gain anything from that. So this this first approach
  8. is not efficient compared to the second one. The last one log packets into
  9. files and use the Shell to check whether the inputs and outputs are correct.
  10. this really moves the problem out of our TestBench, so we then would have to
  11. interact and write Shell scripts or Perl scripts to checked log files and we
  12. also would have a lot of memory overhead. With our transaction level TestBench,
  13. we can check every packet which comes out immediately and if it was correct,
  14. delete it, and have no further overheads. So, the second choice is really the
  15. most efficient one.