[SI-LIST] Re: Hspice vs. Eldo
- From: pritchard_jason@xxxxxxx
- To: <heyfitch@xxxxxxxx>, <si-list@xxxxxxxxxxxxx>
- Date: Tue, 6 Nov 2007 09:49:50 -0500
Part of the problem with Hspice is their inability to keep up with the
times. Most people are operating in the Gbit range whether it is DDR
analysis or high speed serial links. Typical high speed simulations can
be painful in HSPICE so it would be nice to have some different
simulation options. When you have to wait 8 hours to get HSPICE to give
you 200ns of data it starts to become a useless tool.=20
It would be great if HSPICE incorporated alternative simulation options
like Stat-eye or "fast convolution".=20
(NOTE: "Fast Convolution" is a term we use inside EMC. It's not the
literal translation of Fast convolution). "Fast convolution" takes the
same pulse response created for Stat-Eye and you can use convolution to
create a data dependant eye-diagram versus a statistical result. A
statistical result is a worst case simulation in my opinion. A data
dependant simulation is good for lab debug as well as determining what
patterns cause your worst case eye diagram. It also allows you to run
millions of bits in minutes if you so desire.....
Once you get these options it would be nice to have built in DFE
functions. Standards such as SAS and FC are incorporating DFE's for the
upcoming 6G and 8.5G standards. This can be incorporated as direct
building blocks or with IBIS-AMI. HSPICE needs this ability as well.
The good news is that other simulation tools are incorporating these
types of features. The bad news is you need your vendor to give you
models that can be used in those alternative tools.=20
I would encourage you to ask for these types of features in HSPICE if
you think they would be useful.
-Jason
-----Original Message-----
From: si-list-bounce@xxxxxxxxxxxxx [mailto:si-list-bounce@xxxxxxxxxxxxx]
On Behalf Of Vadim Heyfitch
Sent: Friday, November 02, 2007 10:00 PM
To: si-list@xxxxxxxxxxxxx
Cc: Chris.Cheng@xxxxxxxx
Subject: [SI-LIST] Re: Hspice vs. Eldo
Julia, and all:
having been on three sides of this table - systems, silicon vendor -
both
fabless and fabbed, and EDA vendor - I'd like to comment.
Every silicon vendor using TSMC as a foundry, receives spice process
cards
from them in three formats: HSpice, Spectre, and Eldo. All -
unencrypted. (I
don't know about UMC and other foundries.) It's circuit design group
almost
certainly uses just one simulator, let's say HSpice (for example sake).
That
does not preclude other organizations within a silicon company from
taking
the HSpice netlist and running it with other simulators, that understand
HSpice netlist but use other process cards (Eldo or Spectre).
The effort of validating spice models under a different simulator is
rather
small .... as long as you are not trying to map a set of process
parameters
from a one-off, in-house-developed-by-TCAD-department, non-BSIM4.5 . (It
is
not an issue for a fabless semi vendor). And if there are differences
between your simulations , there is not much you can do about it anyway.
To
make a Hspice netlist run under Eldo - is a few minute effort: add a few
short lines at the top and bottom of the netlist, and point to different
libraries. Encryption is as trivial as it is in HSpice: one command at
the
cmd prompt. It takes much less than one part-time apps engineer to
support
on-going releases of IO buffer models in an additional format or two.
If systems companies were more persistent in asking silicon vendors for
IO
buffer models in Spectre and Eldo, they would receive them. Some silicon
vendor already provide such models (Altera Stratix1, for example).
It's is incorrect to assume that all silicon is designed with HSpice.
I've
heard that actually HSpice share of designs has shrunk. Yet I am sure
that
silicon vendors designing with Spectre and Eldo still generate HSpice IO
models for their customers to make them happy. I completely agree with
Chris
that 9 out of 10 times a customer does nothing with the IO spice model
after
receiving it. Yet, having this model available (as a "collateral") makes
a
silicon vendor credible with the customer, thus generating sales.
- Vadim
MetaRAM
On 11/1/07, Chris Cheng <Chris.Cheng@xxxxxxxx> wrote:
>
> In an ideal world, we can have whatever we want.
>
> But in a real world, ask yourself if you are a silicon vendor will you
=3D
> willing to :
> a) pay extra to get an AMS tool that has nothing to do with your
silicon =3D
> design flow
> b) invest in the manpower to qualify your internal SPICE (most likely
=3D
> HSPICE) model with ELDO or whatever AMS standard
>
> HSPICE with S-parameter or W element is a problematic path to say the
=3D
> least. But I can get around it with other tools that can generate =3D
> something HSPICE can feed on and work. That are plenty of free or low
=3D
> cost vector fitting tools that can help you get around the S-parameter
=3D
> or W-element problem with HSPICE.=3D20
>
> Ask yourself this question, with all the hype about IBIS 4.2 with SSO,
=3D
> how many silicon vendor even border to publish that ?=3D20
>
> Tool vendors can promise you a lot of things. At the end of the day,
it =3D
> is your silicon vendor that has to do it.
>
> And coming from experience from both sides of the table, I have seen =
=3D
> plenty of customers that demand certain models in HSPICE, IBIS etc and
=3D
> just sit on it after they are delivered. I have done audit on usage of
=3D
> models after customers received them and surprise surprise less than
10% =3D
> of them are actually simulating anything with them. This is hardly a =
=3D
> statistic that encourage me to invest in generating extra models.
>
> You may be the busy bee out there that work on simulating every net
you =3D
> have with every model you get. But there are plenty of CAD or System =
=3D
> managers out there that demand this and want that but end up sitting =
=3D
> with a $100K+ tool and super complicated model accurate down to 1 pico
=3D
> second and do nothing about them.
>
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx
> [mailto:si-list-bounce@xxxxxxxxxxxxx]On Behalf Of Tom Biggs
> Sent: Thursday, November 01, 2007 1:36 PM
> To: si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Hspice vs. Eldo
>
>
> If a company doesn't have a dedicated SI person, a tool that is most
> useful is not necessarily the 'standard' SPICE. Instead, they will
look
> for a tool that integrates well with the schematic/board layout
process
> and is easy to use for all the designers (who only occasionally use
the
> tool), and has IBIS, SPICE and for the future AMS.
>
> And more than likely there will not be enough budget for more than one
> tool. The sacrifice is that encrypted HSPICE models could become
> useless. Sometimes with lots of pleading, you can sign lots of NDAs
and
> get an unencrypted model. Chip companies want to sell chips, so I'm
sure
> there is going to be more and more pressure on them to provide models
> that don't restrict a customer to a tool they may not have and don't
> have the budget to go out and get.
>
> This has been discussed before: there should be some sort of encrypted
> AMS standard so the industry has something useful to go forward with.
>
> -tom
>
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx
[mailto:si-list-bounce@xxxxxxxxxxxxx]
> On Behalf Of Dan Bostan
> Sent: Thursday, November 01, 2007 11:40 AM
> To: Aubrey_Sparkman@xxxxxxxx; fialka113@xxxxxxxxx;
scott@xxxxxxxxxxxxx;
> weirsi@xxxxxxxxxx
> Cc: si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Hspice vs. Eldo
>
> You get what you pay for.
> Go for the HSPICE as it is the standard in the Spice SI space.
> /dan
>
>
> ----- Original Message ----
> From: "Aubrey_Sparkman@xxxxxxxx" <Aubrey_Sparkman@xxxxxxxx>
> To: fialka113@xxxxxxxxx; scott@xxxxxxxxxxxxx; weirsi@xxxxxxxxxx
> Cc: si-list@xxxxxxxxxxxxx
> Sent: Thursday, November 1, 2007 10:53:06 AM
> Subject: [SI-LIST] Re: Hspice vs. Eldo
>
>
> When Schedules and Time to Market are important, you can justify the
> cost even if the tool is used only a few months a year. When the work
> load is high and constant, the ease of use argument gets weaker as you
> can better benefit from re-use and scripting and the (re-)learning
> curve goes away.
> Aubrey Sparkman
> Enterprise Engineering Signal Integrity Team Dell, Inc.
> Aubrey_Sparkman@xxxxxxxx
> (512) 723-3592
> "Discovery consists of seeing what everybody has seen and thinking
what
> nobody else has thought." - Jonathan Swift=3D3D20
>
> =3D3D20
>
> ________________________________
>
> From: Julia Nekrylova [mailto:fialka113@xxxxxxxxx]
> Sent: Thursday, November 01, 2007 12:35 PM
> To: Sparkman, Aubrey; scott@xxxxxxxxxxxxx; weirsi@xxxxxxxxxx
> Cc: si-list@xxxxxxxxxxxxx
> Subject: Re: [SI-LIST] Re: Hspice vs. Eldo
>
>
> 2) Depends on how constant your work load is. If it is high and
> constant, then yes, I would agree with you that 50 % more cost is
> easily
> justifiable. Btw, Hspice costs 100 % more.
>
>
> ----- Original Message ----
> From: "Aubrey_Sparkman@xxxxxxxx" <Aubrey_Sparkman@xxxxxxxx>
> To: scott@xxxxxxxxxxxxx; weirsi@xxxxxxxxxx
> Cc: fialka113@xxxxxxxxx; si-list@xxxxxxxxxxxxx
> Sent: Thursday, November 1, 2007 8:14:00 PM
> Subject: [SI-LIST] Re: Hspice vs. Eldo
>
> In addition to Steve and Scott's comments, I would like to add two
more
> comments:
> 1) OFTEN, model accuracy is much more important than which simulator
> you
> use.
> 2) If you have a choice of simulators, (you don't with encrypted
> models)
> how important is ease of use? For instance, if one simulator costs
50%
> more, but enables you to get 3 times as much done, do you care?
>
>
> Aubrey Sparkman=3D3D3D20
> Enterprise Engineering Signal Integrity Team=3D3D3D20
> Dell, Inc.=3D3D3D20
> Aubrey_Sparkman@xxxxxxxx=3D3D3D20
> (512) 723-3592=3D3D3D20
> "Discovery consists of seeing what everybody has seen and thinking
what
> nobody else has thought." - Jonathan Swift
>
>
> -----Original Message-----
> From: si-list-bounce@xxxxxxxxxxxxx
> [mailto:si-list-bounce@xxxxxxxxxxxxx]
> On Behalf Of Scott McMorrow
> Sent: Thursday, November 01, 2007 11:32 AM
> To: weirsi@xxxxxxxxxx
> Cc: fialka113@xxxxxxxxx; si-list@xxxxxxxxxxxxx
> Subject: [SI-LIST] Re: Hspice vs. Eldo
>
> I'd agree with Steve. Usually encrypted models is the deal breaker.
> Otherwise, Eldo is technically superior to HSPICE in the handling of
> s-parameters.
>
>
> Scott McMorrow
> Teraspeed Consulting Group LLC
> 121 North River Drive
> Narragansett, RI 02882
> (401) 284-1827 Business
> (401) 284-1840 Fax
>
> http://www.teraspeed.com <http://www.teraspeed.com/>=3D3D20
>
> Teraspeed(r) is the registered service mark of
> Teraspeed Consulting Group LLC
>
>
>
> steve weir wrote:
> > Julia, Synopsys has two almost unassailable market advantages with
> HSPICE:
> >
> > *A lot of models are supplied only as encrypted HSPICE models. If
> you
>
> > don't have HSPICE you can't play.
> > *HSPICE is used for silicon sign-off with a number of fabs.
> >
> > If either of the above issues are material to you, then the
> technical=3D3D3D20
> > merits of one simulator versus another are moot.
> >
> > Steve
> >
> > Julia Nekrylova wrote:
> > =3D3D3D20
> >> Dear SI professionals,
> >> My question is about Eldo vs. Hspice comparisson. With the former
> being roughly 2 times cheaper than the later, I wonder if it offers a
> similar performance. I've heard that Eldo has an improved t-line
> modeling and handles S-parameters in a more advanced fashion. I just
> never had a chance to work with Eldo, but pricewise it's a big
> difference for a small company like ours.
> >>
> >> Thanks again for your help,
> >> Julia Nekrylova
> >>
> >> 74ze Engineering
> >>
> >> __________________________________________________
> >> Do You Yahoo!?
> >> Tired of spam? Yahoo! Mail has the best spam protection =3D
> around=3D3D3D20
> >> http://mail.yahoo.com=3D3D3D20 =
<http://mail.yahoo.com=3D3D3D20/>=3D3D20
> >> ------------------------------------------------------------------
> >> To unsubscribe from si-list:
> >> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject
> field
> >>
> >> or to administer your membership from a web page, go to:
> >> //www.freelists.org/webpage/si-list
> >>
> >> For help:
> >> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
> >>
> >>
> >> List technical documents are available at:
> >> http://www.si-list.net
<http://www.si-list.net/>=3D3D20
> >>
> >> List archives are viewable at: =3D3D3D20
> >> //www.freelists.org/archives/si-list
> >> or at our remote archives:
> >> http://groups.yahoo.com/group/si-list/messages
> >> Old (prior to June 6, 2001) list archives are viewable at:
> >> http://www.qsl.net/wb6tpu
> >> =3D3D3D20
> >>
> >>
> >>
> >> =3D3D3D20
> >> =3D3D3D20
> >
> >
> > =3D3D3D20
>
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List technical documents are available at:
> http://www.si-list.net =
<http://www.si-list.net/>=3D3D20
>
> List archives are viewable at: =3D3D3D20
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
> =3D3D3D20
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List technical documents are available at:
> http://www.si-list.net =
<http://www.si-list.net/>=3D3D20
>
> List archives are viewable at: =3D3D20
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
> =3D3D20
>
>
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around=3D3D20
> http://mail.yahoo.com=3D3D20
>
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List technical documents are available at:
> http://www.si-list.net
>
> List archives are viewable at: =3D3D20
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
> =3D3D20
>
>
>
>
>
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam? Yahoo! Mail has the best spam protection around=3D3D20
> http://mail.yahoo.com=3D3D20
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List technical documents are available at:
> http://www.si-list.net
>
> List archives are viewable at: =3D3D20
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
> =3D3D20
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List technical documents are available at:
> http://www.si-list.net
>
> List archives are viewable at: =3D20
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
> =3D20
>
>
> This email and any attachments thereto may contain private, =3D
> confidential, and privileged material for the sole use of the intended
=3D
> recipient. Any review, copying, or distribution of this email (or any
=3D
> attachments) by others is strictly prohibited. If you are not the =3D
> intended recipient, please contact the sender immediately and =3D
> permanently delete the original and any copies of this email and any =
=3D
> attachments thereto.
> ------------------------------------------------------------------
> To unsubscribe from si-list:
> si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
>
> or to administer your membership from a web page, go to:
> //www.freelists.org/webpage/si-list
>
> For help:
> si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
>
>
> List technical documents are available at:
> http://www.si-list.net
>
> List archives are viewable at:
> //www.freelists.org/archives/si-list
> or at our remote archives:
> http://groups.yahoo.com/group/si-list/messages
> Old (prior to June 6, 2001) list archives are viewable at:
> http://www.qsl.net/wb6tpu
>
>
>
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
List archives are viewable at: =20
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
=20
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List technical documents are available at:
http://www.si-list.net
List archives are viewable at:
//www.freelists.org/archives/si-list
or at our remote archives:
http://groups.yahoo.com/group/si-list/messages
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
Other related posts:
