Home > Community > Forums > Digital Implementation > 2 questions, 1for cdump file and 1for CeltIC

Email

* Required Fields

Recipients email * (separate multiple addresses with commas)

Your name *

Your email *

Message *

Contact Us

* Required Fields
First Name *

Last Name *

Email *

Company / Institution *

Comments: *

 2 questions, 1for cdump file and 1for CeltIC 

Last post Mon, Feb 11 2008 8:38 PM by archive. 7 replies.
Started by archive 30 Jan 2008 02:58 AM. Topic has 7 replies and 2091 views
Page 1 of 1 (8 items)
Sort Posts:
  • Wed, Jan 30 2008 2:58 AM

    • archive
    • Top 75 Contributor
    • Joined on Fri, Jul 4 2008
    • Posts 88
    • Points 4,930
    2 questions, 1for cdump file and 1for CeltIC Reply

    Hi ALL:

    I have 2 questions:

    1. I try to use readgds command to read GDSII file as cdump file, and then using cdump2lef covert cdump to LEF, I got stucked and found that the cdump file is correct, but the generated LEF is wrong.......is cdump2lef having bugs? I am using SOCE52USR3.

    2. When I am running CeltIC, I find the 'set_parm gtol' parameter, this gtol means the filter criteria for aggressors, means glitch tolerance. But I also find 'set_thresh -glitch_peak' can also set such criteria, gtol default is 5% of VDD and glitch_peak is 20% of VDD, what is the difference? Can anybody tell me?

    thank!


    Originally posted in cdnusers.org by eminemshow
    • Post Points: 5
  • Wed, Jan 30 2008 7:56 AM

    • archive
    • Top 75 Contributor
    • Joined on Fri, Jul 4 2008
    • Posts 88
    • Points 4,930
    RE: 2 questions, 1for cdump file and 1for CeltIC Reply

    Hi eminemshow,

    Regarding question #1, readgds and cdump2lef are no longer supported. The recommended way of creating a LEF MACRO from gds is to use the abstract generator which is part of the Virtuoso platform.

    For question #2, I don't know right off personally but hopefully a CeltIC expert can chime in.

    Thanks,
    Bob


    Originally posted in cdnusers.org by BobD
    • Post Points: 0
  • Wed, Jan 30 2008 11:05 AM

    • archive
    • Top 75 Contributor
    • Joined on Fri, Jul 4 2008
    • Posts 88
    • Points 4,930
    RE: 2 questions, 1for cdump file and 1for CeltIC Reply

    Hi eminemshow,

    For #2:

    gtol is the glitch tolerance of a single attacker. This variable to
    used to determine if an attacker can be an individual attacker or if it
    should be grouped into the virtual attacker.

    -glitch_peak is applied to the combined effect of all the attackers on a
    net. If all the attackers that can attack can create more than the
    glitch_peak value amount of noise, then the net is analyzed in more
    detail for glitch. When a net is analyzed in more detail, it means that
    the distributed RCs are used as well as a receiver peak analysis may
    happen.

    thank you,

    Trisha


    Originally posted in cdnusers.org by trishak
    • Post Points: 0
  • Wed, Jan 30 2008 5:42 PM

    • archive
    • Top 75 Contributor
    • Joined on Fri, Jul 4 2008
    • Posts 88
    • Points 4,930
    RE: 2 questions, 1for cdump file and 1for CeltIC Reply

    Posted By trishak on 1/30/2008 11:05 AM
    Hi eminemshow,

    For #2:

    gtol is the glitch tolerance of a single attacker. This variable to
    used to determine if an attacker can be an individual attacker or if it
    should be grouped into the virtual attacker.

    -glitch_peak is applied to the combined effect of all the attackers on a
    net. If all the attackers that can attack can create more than the
    glitch_peak value amount of noise, then the net is analyzed in more
    detail for glitch. When a net is analyzed in more detail, it means that
    the distributed RCs are used as well as a receiver peak analysis may
    happen.

    thank you,

    Trisha
    Glitch tolerance is the criteria for a single aggressor, glitch peak is for total effects of multi aggressors. Will CeltIC ignore the effect of multi aggressors on delta delay (and glitch) at all if their total glitch is below '-glitch_peak'? or just dont make a detail analysis and make a rough calculation instead? Hopes that my points are clear. This is very important for make a correlation between CeltIC and PT.


    Originally posted in cdnusers.org by eminemshow
    • Post Points: 0
  • Wed, Jan 30 2008 6:29 PM

    • archive
    • Top 75 Contributor
    • Joined on Fri, Jul 4 2008
    • Posts 88
    • Points 4,930
    RE: 2 questions, 1for cdump file and 1for CeltIC Reply

    Posted By BobD on 1/30/2008 7:56 AM
    Hi eminemshow,

    Regarding question #1, readgds and cdump2lef are no longer supported. The recommended way of creating a LEF MACRO from gds is to use the abstract generator which is part of the Virtuoso platform.

    For question #2, I don't know right off personally but hopefully a CeltIC expert can chime in.

    Thanks,
    Bob
    Thanks! bob, I will try to use abstract gen. Maybe later I will have questions about it.


    Originally posted in cdnusers.org by eminemshow
    • Post Points: 0
  • Thu, Jan 31 2008 7:02 PM

    • archive
    • Top 75 Contributor
    • Joined on Fri, Jul 4 2008
    • Posts 88
    • Points 4,930
    RE: 2 questions, 1for cdump file and 1for CeltIC Reply

    Posted By eminemshow on 1/30/2008 5:42 PM
    Posted By trishak on 1/30/2008 11:05 AM
    Hi eminemshow,

    For #2:

    gtol is the glitch tolerance of a single attacker. This variable to
    used to determine if an attacker can be an individual attacker or if it
    should be grouped into the virtual attacker.

    -glitch_peak is applied to the combined effect of all the attackers on a
    net. If all the attackers that can attack can create more than the
    glitch_peak value amount of noise, then the net is analyzed in more
    detail for glitch. When a net is analyzed in more detail, it means that
    the distributed RCs are used as well as a receiver peak analysis may
    happen.

    thank you,

    Trisha
    Glitch tolerance is the criteria for a single aggressor, glitch peak is for total effects of multi aggressors. Will CeltIC ignore the effect of multi aggressors on delta delay (and glitch) at all if their total glitch is below '-glitch_peak'? or just dont make a detail analysis and make a rough calculation instead? Hopes that my points are clear. This is very important for make a correlation between CeltIC and PT.



    I need the answer, really!


    Originally posted in cdnusers.org by eminemshow
    • Post Points: 0
  • Thu, Feb 7 2008 7:03 AM

    • archive
    • Top 75 Contributor
    • Joined on Fri, Jul 4 2008
    • Posts 88
    • Points 4,930
    RE: 2 questions, 1for cdump file and 1for CeltIC Reply

    The glitch analysis and delta delay analysis criteria are separate. For delta delay analysis, you should use the set_thresh -delta_absolute command/option. This allows you to put a delta delay value as the criteria for the delta delay analysis. For instance, if you want to analyze all nets with delta delays larger than 5ps, then you would do: set_thresh -delta_absolute 5e-12. For glitch, using -glitch_peak is correct and any glitch less than -glitch_peak will not be analyzed in detail.

    You have a lot of very good questions about CeltIC. You might consider taking a CeltIC training class as these details are discussed in the CeltIC training class in more detail than can be easily written in this forum.

    thank you, Trisha


    Originally posted in cdnusers.org by trishak
    • Post Points: 5
  • Mon, Feb 11 2008 8:38 PM

    • archive
    • Top 75 Contributor
    • Joined on Fri, Jul 4 2008
    • Posts 88
    • Points 4,930
    RE: 2 questions, 1for cdump file and 1for CeltIC Reply

    Posted By trishak on 2/07/2008 7:03 AM
    The glitch analysis and delta delay analysis criteria are separate. For delta delay analysis, you should use the set_thresh -delta_absolute command/option. This allows you to put a delta delay value as the criteria for the delta delay analysis. For instance, if you want to analyze all nets with delta delays larger than 5ps, then you would do: set_thresh -delta_absolute 5e-12. For glitch, using -glitch_peak is correct and any glitch less than -glitch_peak will not be analyzed in detail. You have a lot of very good questions about CeltIC. You might consider taking a CeltIC training class as these details are discussed in the CeltIC training class in more detail than can be easily written in this forum. thank you, Trisha

    Thanks! Trisha, I will consider this.


    Originally posted in cdnusers.org by eminemshow
    • Post Points: 0
Page 1 of 1 (8 items)
Sort Posts:
Started by archive at 30 Jan 2008 02:58 AM. Topic has 7 replies.