Talk:SPICE

From Wikipedia, the free encyclopedia

This is the talk page for discussing improvements to the SPICE article.
Article policies

Hi, the article states "The original SPICE program was released under a restrictive license, which makes it difficult for others to improve upon the original software." and yet I have read in print ( I think it was in the IEEE Spectrum ) that SPICE was originally "public domain" and I think they implied it was the first "open source" program. Can someone source the assertion about the original license, and change it if necessary ? --Rob

Question moved from article:

Does anyone know the revision level of the first C implementation of SPICE?

If you know the answer add it to the article in the appropriate paragraph. dave 10:02, 1 Sep 2003 (UTC)

not currently known, but check the following for clues... Waveguy

Found it. SPICE 2G.6 (1983) is the last FORTRAN version. SPICE 3 is C. -- RTC 20:08, 9 Sep 2003 (EDT)

Hello, I am Paolo Nenzi, ngspice developer. Ngspice is not covered by the GPL license, it is still covered by the old BSD license. We asked Berkeley's Regents some years ago to change spice license to new BSD (no obnoxious clause) but we got no answer. Thanks for citing ngspice in wikipedia!

I feel that this article could also include information about the SPICE elements such as their function, computational methods, parameters, etc. -- Peter 1:53 AM, 2 Apr 2006 (EST)

Contents

[edit] XSPICE merged

The brief text from the page XSPICE has now been merged to this article. The XSPICE page has been changed to a simple redirect. DFH 08:05, 19 October 2006 (UTC)

[edit] restrictive license?

SPICE was released under what amounts to the BSD license (or Don Pederson's force of will, depending on how you look at it). There's a line in the article (poorly placed) which asserts BSD is restrictive. It's not gnu, and in many senses it's less restrictive than gnu. Can we remove this line? A note about licensing is probably appropriate, since SPICE was probably the first open source program, but the emphasis should be switched to point out how the nascent IC industry benefitted from a freely available open source circuit simulator. 66.69.212.211 15:10, 4 November 2006 (UTC) Steve

This point is moot; Berkeley went to BSD licensing per their download page. YoungGeezer 17:31, 10 November 2007 (UTC)

[edit] SPICE Acronym Meaning

The explanation of the SPICE acronym has been changed to match standard usage (previously C was 'circuits', now 'circuit'). 25 Jan 2007, Steve

[edit] External/commercial links

I propose eliminating the links to commercial simulators. This has gotten out of hand and has been flagged by Wikipedia as excessive and in need of cleanup. To replace them, I think it would be sufficient to simply say, "There are many commercial circuit simulators, some of which are direct descendants of SPICE2," with a possible mention of PSPICE and HSPICE as being two of the earliest and most popular. Steve, 66.69.212.211 14:13, 25 January 2007 (UTC)


I've moved the Macspice link from 'Open Source' to 'Commercial', as it's not open source. However, it is free (as in beer), so it doesn't really belong there. The thing is there are so many commercial spice implementations, and it's not really as if a few dominate the market so it's got to be all or nothing.144.173.6.75 11:39, 2 February 2007 (UTC)

[edit] Open source

SPICE has been referred to for many years as one of the first open source programs. The open source article mentions source code available to the general public with relaxed or nonexistent IP restrictions; as I remember, the SPICE license was essentially BSD, which is certainly one of the less restrictive. However, now I can't find the original language in the source code or elsewhere, and the current software agreement is about as open as can be found. I suggest the language in the article about restrictions and an acknowledgment clause can be removed. YoungGeezer 05:04, 29 June 2007 (UTC)

This point is moot; Berkeley went to BSD licensing per their download page. YoungGeezer 17:31, 10 November 2007 (UTC)

[edit] Compact model council purpose

The CMC mainly exists to make it easy to port model parameters from one simulator to another. This requires standard models and standard model parameters. The reason this is important is primarly for fabless (or fab-lite) semiconductor companies, who need to take the model parameters given by the foundry and use them in various simulators. While there are still syntax differences between the different simulators, a standard model will at least ensure that a given set of model parameters will produce a known result on each of the simulators (assuming the implementation follows the reference implementation faithfully, which is again somewhat iffy but at least there's a standard to use to beat on the implementations that deviate).

All that said, the CMC website mentions this obliquely through the first sentence in their vision statement: "Standardized compact models for all major technologies so that customer communication and efficiency can be enhanced." "Customer communication" basically comes down to sets of model parameters to run into standard models.

As it turns out, the "standard interfaces" thing dropped out. Eight or nine years ago a couple of standard interfaces were proposed, but neither caught on, because of their implicit assumptions about simulator architecture. A better approach is being pursued now, using Verilog-AMS as an executable specification and a program like ADMS to translate the Verilog-AMS into simulator code (with some intervening steps and a bit of work).

So I'm reverting the change regarding the purpose of the CMC.

YoungGeezer 18:02, 12 November 2007 (UTC)

Thanks for the explanationBrews ohare 18:10, 12 November 2007 (UTC)

[edit] Vaccardi link does not work

I undid this link because it didn't work Brews ohare (talk) 01:10, 24 November 2007 (UTC)

[edit] Tina-TI?

There seems to be a toggling of the TI simulator reference from TISPICE to Tina-TI and back. The paragraph in question is about industrial circuit simulators. TISPICE is a SPICE derivative which is written, supported and used in TI. Tina-TI is a give-away for marketing purposes, was not written in TI, and as far as I know is not used (except by marketeers) in TI. The proper reference for the paragraph is to TISPICE. YoungGeezer (talk) 14:14, 24 December 2007 (UTC)

[edit] Lists of compact models

I (somewhat regretfully) removed the recently added list of 'alternative compact models'. The article is about SPICE; the GEIA Compact Model Council and standard models is somewhat related, but I don't think the SPICE article is the correct place for a list of all possible compact models -- there are sooo many of them, everyone has their favorites, and most have only limited usage. Perhaps the transistor models article would be a better place. YoungGeezer (talk) 19:11, 26 January 2008 (UTC)

[edit] perhaps some discussion of gnucap could be added?

Perhaps there are complicated reasons for not doing so, but it seems that adding some reference to gnucap as an open source project largely inspired by the SPICE program family might be in order? <gnucap_website> 71.197.225.209 (talk) 00:13, 31 May 2008 (UTC)