RE: [SI-LIST] : Hspice: Windows vs Unix

About this list Date view Thread view Subject view Author view

From: Burns, Douglas ([email protected])
Date: Wed Apr 11 2001 - 09:52:38 PDT


Most of our data extraction is performed directly from the TR0 file, but
when we need to see the result, we've been using AWAVES. From our
experience, AWAVES is more stable on UNIX than on the PC. This is
especially true for large files (as you point out).

 Doug

-----Original Message-----
From: Offer Kaye [mailto:[email protected]]
Sent: Wednesday, April 11, 2001 1:41 AM
To: Burns, Douglas
Cc: [email protected]
Subject: RE: [SI-LIST] : Hspice: Windows vs Unix

Hmm..
Doug, I have a question also- how would you compare the data
visualization
tools (e.g. waveform viewers) for Hspice in PC vs. Unix? I've used
Hspice
on both platforms, and have found the Unix solutions to be much superior
to the PC solution- just try loading a 1.5GB results file to your
waveform viewer on a PC...

Offer Kaye
[email protected]

On Tue, 10 Apr 2001, Burns, Douglas wrote:

> Anthony,
>
> I work for Signal Integrity Software Inc (SiSoft) where we have been
> developing software to support our high-speed design methodology. This
> methodology encompasses both pre and post layout waveform and timing
> analysis for GHz+ signaling environments. We seamlessly support IBIS
and
> SPICE models on both PC and UNIX based HSPICE. We use both compute
> environments and have launched long complicated simulations
successfully
> on both platforms.
>
> PC Solution
> 1) Does not support queuing. Jobs must be submitted in a serial
fashion.
>
> 2) PC_HSPICE lives on the PC. If you get a node-locked license, Only
the
> person with the key can access HSPICE. You can get a floating license
> and install HSPICE on all user's PC's, but you are still limited to
the
> number of jobs that can be launched simultaneously. Also you would
need
> to do separate installs/updates/support on all machines. Third, you
> could install HSPICE on a PC server and give your users access to the
> server.
> 3)PC hardware is less expensive with similar performance to the UNIX
> platforms
>
> The Unix Solution
> 1) Supports queuing. Hundreds of jobs can be submitted easily
> 2) Multiple users can simultaneously log onto 1 machine and submit
jobs
> 3) Easy scripting/data manipulation
> 4) New CMI models support (though not on all UNIX's)
>
> In practice, user access and data manipulation capabilities should be
> primary concerns. We believe that the UNIX platform presently provides
> the best access and flexibility.
>
> Doug Burns
> Principal Consultant
> Signal Integrity Software, Inc.
> www.SiSoft.com
> [email protected]
>
> -----Original Message-----
> From: Mohan [mailto:[email protected]]
> Sent: Tuesday, April 10, 2001 12:18 AM
> To: Anthony Davidson; [email protected]
> Subject: Re: [SI-LIST] : Hspice: Windows vs Unix
>
>
> Hi
>
> It depends on your needs......If you need to run small circuits, it is
> okay
> to use the PC version. Otherwise, Unix is preferable anyday !! I have
> had
> many instances of PC-version crashing when running long and accurate
> simulations ( eg. PLL ). Version 2000.2 (?) has problems with the
> 'Update'
> feature in PC. I cannot track a waveform while simulation is running.
(
> There could be some fixes for these ! ).
>
> As regards the links to 3rd party tools : the links help you to invoke
> Hspice simulator other than the default simulator. For example, if you
> are
> using Cadence's Analog artist, you would be able to use Hspice as a
> simulator instead of CDSpice or Spectre.. the link provides the
license
> to
> do so. You could then analyse the results in the 3rd party's waveform
> viewer. Of course, you need to have Hspice binaries too....i am not
> aware of
> any Cadence or Mentor tools on PC. Mentor's NT support is probably
close
> to
> what you would want on a PC.
>
> Mohan Kumar
>
>
>
> ----- Original Message -----
> From: "Anthony Davidson" <[email protected]>
> To: <[email protected]>
> Sent: Tuesday, April 10, 2001 1:55 AM
> Subject: [SI-LIST] : Hspice: Windows vs Unix
>
>
> > We are looking at Star-Hspice from Avanti! as a signal integrity
> analysis
> > tool to evaluate/analyze high speed interconnects. We have decided
to
> > purchase some licenses but now we have to decide to buy a Windows
> version
> or
> > a Unix (Solaris, I believe) version.
> >
> > According to Avanti!, Windows Hspice is a subset of Unix Hspice. The
> > information I have is this:
> >
> > CMI is not supported on Windows. CMI is a development library
allowing
> one
> > to create MOSFET models in C. This will prevent us from using any
> vendor's
> > MOSFET model that was generated with CMI, however, this may not be a
> problem
> > because most manufacturers use the BSIM3 model nowadays. Does
anybody
> know
> > if BSIM3 models are better/worse than models generated with CMI?
Will
> I be
> > limited in anyway by not having CMI?
> >
> > Avanti! says that some 3rd party output formats are only supported
on
> Unix,
> > such as CadenceLink and MentorLink. Does anybody have any
experiences
> they
> > can describe using such links? What do these links allow you to do?
> >
> > Finally, has anybody used Hspice on both Windows and Unix platforms?
> Would
> > you say that you prefer one platform over the other? Why?
> >
> > Any other comments would be greatly appreciated.
> >
> > Thanks in advance for your help!!
> >
> > Anthony Davidson
> >
> > **** To unsubscribe from si-list or si-list-digest: send e-mail to
> > [email protected]. In the BODY of message put: UNSUBSCRIBE
> > si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> > si-list archives are accessible at http://www.qsl.net/wb6tpu
> > ****
> >
> >
>
>
> **** To unsubscribe from si-list or si-list-digest: send e-mail to
> [email protected]. In the BODY of message put: UNSUBSCRIBE
> si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu
> ****
>
> **** To unsubscribe from si-list or si-list-digest: send e-mail to
> [email protected]. In the BODY of message put: UNSUBSCRIBE
> si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
> si-list archives are accessible at http://www.qsl.net/wb6tpu
> ****
>
>

**** To unsubscribe from si-list or si-list-digest: send e-mail to
[email protected]. In the BODY of message put: UNSUBSCRIBE
si-list or UNSUBSCRIBE si-list-digest, for more help, put HELP.
si-list archives are accessible at http://www.qsl.net/wb6tpu
****


About this list Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Jun 21 2001 - 10:11:31 PDT