Use OpenVAF instead of ADMS to implement Verilog-A support ？

[weiji]

OpenVAF is a Next-Generation Verilog-A compiler. It is open source and provides state-of-the-art Verilog-A support. Recent releases of NGspice have implemented OpenVAF integration and started replacing ADMS. OpenVAF showed better performance in some tests. Is it possible for Xyce to support Verilog-A model by using OpenVAF instead of ADMS? What are the differences and difficulties in the implementation? 

[xyce-users]

Thanks for your interest.  We are aware of OpenVAF and have had some interactions with the OpenVAF team.  As part of that interaction we provided some feedback to their OSDI interface.   We'd like to get it working with Xyce but we haven't had time to make that happen yet. 

So to answer your question,  yes, it should be possible to make it work.   

At an abstract level our needs would be very similar to nearly any other circuit simulator and I would imagine the work required would also be similar.  At a more detailed level, Xyce's data structures and device API are a bit different than SPICE-derived codes.  So we'd need to implement our own OSDI support.

thanks,

Eric

[Murat H Eskiyerli]

I think it also allows Verilog-A support without requiring GCC is also an important consideration. Verilog-A would also be usable on Windows boxes.

Murat H Eskiyerli

---

From: xyce-users <xyce-...@googlegroups.com>
Sent: Tuesday, February 21, 2023 9:04:54 AM
To: xyce-users <xyce-...@googlegroups.com>
Subject: [xyce-users] Re: Use OpenVAF instead of ADMS to implement Verilog-A support ？

 

--
You received this message because you are subscribed to the Google Groups "xyce-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to xyce-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/xyce-users/09f19e3f-1996-44d5-9d11-325932f87337n%40googlegroups.com.

[Markus Müller]

Hi, Markus from SemiMod/OpenVAF team here.

As mentioned we did have some very productive exchanges on the OSDI interface with the Xyce team.

I believe the Xyce integration is very doable, in fact, we studied both Xyce and Ngspice interfaces during

development.

I sincerely hope that either Xyce Devs at Sandia or SemiMod can find funding for this integration.

And yes: OpenVAF removes GCC dependency, and you will likely see a significant speed-up of the models compared to ADMS.

This is because we employ some neat compiler techniques to remove any unnecessary operations.

[Mehmet Cirit]

Google should be able to provide some funding, along with the compact modeling group. Did you contacted them? 

To view this discussion on the web visit https://groups.google.com/d/msgid/xyce-users/2691be14-6ed8-4ff4-9e0c-1b845d97c48dn%40googlegroups.com.

--

http://www.linkedin.com/in/mehmetcirit

Dr. Mehmet A. Cirit                    Phone:  (408) 647-6025

Library Technologies, Inc.        Cell:       (408) 647-6025

19959 Lanark Lane                   http://www.libtech.com

Saratoga, CA 95070                 Email: m...@libtech.com

[Mehdi Saligane]

Hi Marcus, Eric and all -

Sorry for hijacking this thread.

Thanks for all the work. We have invited Ken and Alessandro to our next AWG meetings (28th Feb and next).

So please make sure to attend. 

Thanks

Mehdi

On Thu, Feb 23, 2023 at 2:35 PM Markus Müller <metro...@gmail.com> wrote:

I did have a few conversations with Google on this topic and asked for funding. Unfortunately, I am not sure if they want to fund this.

At least so far, nothing has happened. But I am still positive, we are just at the beginning.

Once we have PDKs with advanced models coming, the need for an ADMS replacement will be seen by others.

Attached are two of our recent presentations for your interest.

To view this discussion on the web visit https://groups.google.com/d/msgid/xyce-users/2339181e-77e2-44df-a081-a72a3fe05562n%40googlegroups.com.

--

Mehdi Saligane

Assistant Research Scientist

University of Michigan, Electrical and Computer Engineering

1301 Beal Ave, 2417G EECS

734.272.5288

me...@umich.edu