Many papers discuss about the benefit cost of using OVM/UVM field macros all coming to the conclusion of that it should not be used. But, many of these papers were written in 2011, has there not been any improvements since then to say that it has been fixed where users can use them?
In reply to GCHAN310:
Although the UVM had a few performance improvements put in around the time of this paper, we have not seen any improvements to the field macros that would change our recommendations.
In reply to GCHAN310:
cliff cummings “UVM transactions” (2014) dives into this topic pretty heavily ( more detail than most users probably need ); he benchmarks the field macros vs do methods. He mentions that the Cadence user guide recommends using the field macros. Mentor has the opposite recommendation.