( DAC 04 Item 35 ) --------------------------------------------- [ 02/09/05 ]

Subject: Sequence Columbus RF, Simplex Fire&Ice, Mentor Calibre-xRC

WHERE'S PENNY? -- Ever since Penny Hersher (the old CEO of Simplex who had
become the head of Cadence marketing) has left, all the old Simplex tools
inside of Cadence have been disappearing from the user mindshare.  You now
see a paltry handful of 2 liners about Fire&Ice vs. Sequence Columbus users
writing volumes.  Check out ESNUG 433 #2 along with the comments below.


    Simplex I use, but it's complex.  Simplex ElectronStorm and Fire&Ice.

        - [ An Anon Engineer ]


    Fire&Ice

        - [ An Anon Engineer ]


    Currently, I have done the correlation for RC extraction and delay
    calculation for Sequence tools.

    I have compared the RC results produced by Sequence Columbus with
    Simplex Fire&Ice and Avanti Star-RC.  They are close.

    I have compared the delay calculation result produced by Physical
    Studio with Avanti HSPICE and Cadence MDC.  It seems that the
    Sequence result is closer to HSPICE.

        - Frank Zhou of S3 Graphics


    We primarily do high-speed mixed signal & microwave analog design using
    IBM SiGe BiCMOS processes.  We use Cadence design and simulation SW.
    We had a primary interest in a tool that could accurately extract
    parasitic inductance, as well as capacitance and series resistance in
    wiring interconnects.  Inductance extraction is the most difficult job
    of the extractor -- capacitance and resistance extraction are easier
    technically.  There are not that many tools out there that can do the
    inductance extraction job effectively and it seems that the only people
    who really care are the analog designers doing high frequency work.

    Because we're a small company with limited resources we did not evaluate
    the other parasitic extraction tools, but we did test Columbus-AMS in
    some detail before we bought it.  We had heard positive reviews of
    Columbus-AMS from other analog designers and IBM recommends it for use
    with their SiGe BiCMOS design kits.

    It might seem logical for us to have evaluated Cadence Fire&Ice since we
    use Cadence, but the hard truth is that we have had so much negative
    experience with countless and continual bugs in Cadence design software 
    that we just didn't want yet another piece of buggy software to have to
    deal with.  Our experience (at least with the analog side of Cadence) is
    that the quality of Cadence simulation/analysis software is incredibly
    poor, relative to the cost of using it.  The only reason we use Cadence
    is that we're forced to because the IBM SiGe process design kits are
    offered primarily for Cadence.  That's not to say that the other design
    software out there my not have similar quality problems.

    Now on to Sequence Columbus RF.  In general, inductance in anything but
    the simplest configurations, such as a straight wiring line over a
    ground plane, is very difficult to compute exactly because the
    inductance of a specific wiring geometry depends on the ground return
    current paths but the ground return current paths in turn depends on the
    inductance impedance.  So the exact solution to the inductance problem
    of a real life circuit is a simultaneously solution of a complex array
    of coupled equations which no inductance extraction software would
    attempt.  Hence, by its nature, inductance extraction is an art of
    estimating inductance using some reasonable assumptions about the ground
    return current paths.

    For this function, we have been very happy with Columbus.  In my initial
    evaluations, I ran some specific tests of its inductance computation
    accuracy on several wiring interconnects for which I had already
    calculated inductance manually using approximate analytical equations.

    In most cases the two inductances agreed within the accuracy limitations
    expected for these approximate analyses.  In terms of user interface, we
    found that Columbus RF installed relatively easily, and is very easy to
    use by the average circuit designer who does not want to wade into the
    complexities of customizing his inductance extraction -- the default
    extraction parameters seems to be well designed.  On the other hand,
    Columbus offers a significant extraction flexibility and control for the
    power user who wants to do more complex or customized extractions.

    One weakness of Columbus RF (that we have observed in Cadence-based
    parasitic capacitance extraction as well) is that it is rather difficult
    to view a layout of your design with the parasitic elements back-
    annotated into the layout view.  The back-annotated elements are very
    small and hard to see and their position in the layout view yields
    little information about where they are actually connected.

    But this seems to be a generic problem with other extraction software
    as well, based on our experience with Cadence.  On the other hand,
    Columbus has included an inductance calculator that allows the user to
    manually probe the inductance of a user-selected wiring interconnect.
    This allows the designer to rapidly get a feel for the impact of
    specific layout configurations and is a great aid analysis.

    Thus far (we have owned the tool for about 4 months) we have been very
    pleased with the capabilities and user interface of Columbus RF.

        - Mike Hoskins of Q-dot, Inc.
    Sequence Design sells "ExtractionStage", their new name for their
    Columbus tools, which they claim is within 5% of measured silicon,
    can extract 1 million nets per hour, and has the industries first
    loop and mutual inductance capability.

    Magma almost sounds like they've taken a page form Cadence's
    acquisition book, except that unlike Cadence they are acquiring
    technology they don't already have.  They bought Random Logic Corp,
    whose QuickCAP tool is the de facto standard for field solvers.
    Most normal extraction tools will quote how closely they correlate
    to QuickCAP.  Note that "quick" is only relative to older tools in
    the field -- any field solver will still be extremely slow compared
    to a normal extraction tool and they are not suitable for full chip
    extraction unless you have a mighty small chip or a 10 year design
    schedule.

        - John Weiland of Intrinsix Corp.


    For digital extraction you propably want to use the same vendor as
    for your P&R tool.  For analog I have most faith in Calibre-xRC.

        - [ An Anon Engineer ]

Index   
Next->Item







   
 Sign up for the DeepChip newsletter.
Email
 Read what EDA tool users really think.


Feedback About Wiretaps ESNUGs SIGN UP! Downloads Trip Reports Advertise

"Relax. This is a discussion. Anything said here is just one engineer's opinion. Email in your dissenting letter and it'll be published, too."
This Web Site Is Modified Every 2-3 Days
Copyright 1991-2024 John Cooley.  All Rights Reserved.
| Contact John Cooley | Webmaster | Legal | Feedback Form |

   !!!     "It's not a BUG,
  /o o\  /  it's a FEATURE!"
 (  >  )
  \ - / 
  _] [_     (jcooley 1991)