Stats

  • Locked Locked
  • Replies 2
  • Subscribers 125
  • Views 13063
  • Members are here 0
This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

MC simulation results of copied-but-equal cells are different

HoWei

Hi,

I do have a design (testbench+blocks) and performed MC simulations.

Doing the same simulation again will result with same simulation results.

Now I had to rename the design (testbench+blocks), so I simply copied all cells to a new name.

Nothing internally changed. The config-view was adopted to the new names and the maestro view as well.

But performing the same MC simulation shows different results.

Is this expected ?

Will the rename of a design cause a different seed for the MC simulation and thus results in different simulation results ? 

  • The seed remains the same (unless you set it, the seed in ADE is fixed at 12345). If you're doing process only, I would expect no difference. If mismatch is included (either just mismatch, or process+mismatch) then there's a possibility that something during the copy caused the database order to alter, which means that the order in the netlist might be different.

    Since Spectre generates the random numbers for mismatch for each subckt instance in the design, either having different numbers of instances or them in a different order can lead to different results. Of course, the fact that the results are different shouldn't really matter, since the distribution should be the same - sometimes worrying about having precisely the same results is a false concern.

    One thing you can do to check iwould be to set the option sortinstance=yes (Simulation->Options->Analog, type this in the Additional arguments field at the bottom of the Miscellaneous tab). You'd have to do this on both runs (i.e with the original and new design) to compare them.  This would then ensure that the order in the netlist doesn't matter if you're really concerned about whether the difference is significant or not. Note that I think the instance names would need to remain the same for the sort order to match.

    Regards,

    Andrew.

  • in reply to Andrew Beckett

    Perfect !

    I tried your proposal and the results are exactly the same !

    This gives me confidence that the designs are the same.

    Of course the different result would not change the designs performance, but I was just wondering what caused the difference in results !

    Thx !

Community Guidelines

The Cadence Design Communities support Cadence users and technologists interacting to exchange ideas, news, technical information, and best practices to solve problems and get the most from Cadence technology. The community is open to everyone, and to provide the most value, we require participants to follow our Community Guidelines that facilitate a quality exchange of ideas and information. By accessing, contributing, using or downloading any materials from the site, you agree to be bound by the full Community Guidelines.