Stats

  • Locked Locked
  • Replies 5
  • Subscribers 61
  • Views 15971
  • 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

Multi corner workaround with RC?

archive

Hi,

I need advice related to WC and WCL corners.  Thanks.

At very small geometry, a std cell delay can be slower with decreasing temperature.  As a result, we have a new corner named WCL. It is the same as WC except that temperature is at -40 degree and not 125 degree. Compare between WC and WCL, a std cell is not always slower or faster at one corner.  As delays are modelled as 7x7 table for rise/fall/rise_transition/fall_tranisiton, a cell can be faster/slower in some of the entries at WC or WCL.

My question is how do I get the best possible result in term of timing, leakage, dynamic power and area with RC.  The easy answer is to use multi-corner approach that is available in SoC. But multi-corner is not available in RC and I think runtime will increase.

Any workaround you can recommend ?

Regards,
Eng Han


Originally posted in cdnusers.org by EngHan
Parents

  • Hi EngHan,

    power should certainly not suffer. The power_library attribute should certainly guarantee that. It basically tells RC to use a different library to evaluate power. I also saw in the low power document that there is support for DVFS hence you could address your multi-mode issues that way. Seems that would be much cleaner than a merged dotlib. If you try that, I would be curious to hear how it goes. I suspect there is a runtime penalty but other than that, I think this would be quite powerful.

    good luck,
    gh-


    Originally posted in cdnusers.org by grasshopper
Reply

  • Hi EngHan,

    power should certainly not suffer. The power_library attribute should certainly guarantee that. It basically tells RC to use a different library to evaluate power. I also saw in the low power document that there is support for DVFS hence you could address your multi-mode issues that way. Seems that would be much cleaner than a merged dotlib. If you try that, I would be curious to hear how it goes. I suspect there is a runtime penalty but other than that, I think this would be quite powerful.

    good luck,
    gh-


    Originally posted in cdnusers.org by grasshopper
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.