Stats

  • Locked Locked
  • Replies 3
  • Subscribers 125
  • Views 14900
  • 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

regarding simultion convergence issues

fazul

Hi everybody!

i have a question regarding  simulation convergence.. I am running a Transient Noise Analysis

with errpreset = conservative. Initially it was having difficultly in converging so i tried the usual techniques

like relaxing abstol, reltol, increasing maxiter, method=gear2only as it was mentioned in the manuals.

Now the simulation seems to run to completion without errors but with a large number of warnings and notices from spectre.

----------------------------------------------------------------------------------------------

WARNING : The bulk-source junction current exceeds `imelt'. The results computed by Virtuoso(R) Spectre are now incorrect because the junction current model has been linearized

WARNING :Vgd has exceeded the oxide breakdown voltage of `vbox'

WARNING : Minimum time step used.  Solution might be in error.

.....

spectre completes with 0 errors, 4672 warnings, and 2409 notices

----------------------------------------------------------------------------------------------

my question is how reliable is the simulation result in this case. At a glance the result seems to be correct but can i really ignore all these warnings and trust the result?

what is the general approach that a CAD engineer takes when he is faced with these kind of warning?

Forgive me if my question is not clear, i am quite new to simulation and havent completely understood  how  spectre/spice simulators work . So any pointers will be very helpful...

Thanks a lot, have a nice day.

Fazul 

 

Parents

  • hello Zach,

    Yes i was running on 64bit Spectre +aps  ..(MMSIM7.2ISR2) ...

     however i am not using IBM PDK.instead its in-house pdk.

    i can understand that there will be some effect on convergence based on what binary i choose to run the simulation

    but   which one do you think is  recommended (32bit or 64bit binary)  when doing simulation with convergence issues?

    my simple understanding is that,

    * when the simulating circuit size is of large use 64bit binary, (accuracy is better, 32bit becomes slower)

    * when trying to plot waveforms of large data size use 64 bit binary

    but what about  from a simulation convergence point of view?

    Any advice ? Sorry this sounds like a new different question .

    thanks Zach for replying.

    Fazul

     

Reply

  • hello Zach,

    Yes i was running on 64bit Spectre +aps  ..(MMSIM7.2ISR2) ...

     however i am not using IBM PDK.instead its in-house pdk.

    i can understand that there will be some effect on convergence based on what binary i choose to run the simulation

    but   which one do you think is  recommended (32bit or 64bit binary)  when doing simulation with convergence issues?

    my simple understanding is that,

    * when the simulating circuit size is of large use 64bit binary, (accuracy is better, 32bit becomes slower)

    * when trying to plot waveforms of large data size use 64 bit binary

    but what about  from a simulation convergence point of view?

    Any advice ? Sorry this sounds like a new different question .

    thanks Zach for replying.

    Fazul

     

Children
No Data

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.