MeetingMinutes20090805 < VerilogAMS

VerilogAMS Web>AmsAssertions>RequirementsGatheringGroup>MeetingMinutes20090805 (2009-08-05, AnandHimyanshu) (raw view)EditAttach
Attendees:

   * 00000000000000010000101111 Qamar Alam
   * 11111111111111111110101111 Himyanshu Anand
   * 01111111100111111011101000 Kenneth Bakalar
   * 11111010101101101110110000 Prabal Bhattacharya
   * 00000000000000110000101001 Sri Chandra
   * 01111011111111111101101011 Eduard Cerny
   * 11101110000101101011101111 Scott Cranston
   * 00000000000000010000000000 Dave Cronauer
   * 00000000000011000000000111 Dejan Nickovic
   * 11011101111001000000000000 Mike Demler
   * 00000000000000000000000000 Surrendra Dudani
   * 11100000001111111111110011 John Havlicek
   * 11100011001000000000000000 Kevin Jones (RGG Leader)
   * 00000001111111111011101111 Jim Lear
   * 00000000000011101110000000 Top Lertpanyavit
   * 11111101101111111111111110 Scott Little
   * 00000000000001000000000000 Erik Seligman
   * 10100000000000000000000000 David Sharrit
   * 00000001000000000000000000 Murtaza
   * 00000001000000000001011001 Martin O'Leary

Decisions:

   1. Target for the vote on the requirements two weeks from now (August 19th, 2009)

Action Items:

   1. Research multiple clocked assertion evaluation. [John Havlicek, Scott Cranston] 

Details:

SC: Was waiting for the write up from John.

JH: I will send the document now. Is the mailing list setup?

HA: Yes, sort of. Lynn did setup the mailing list as Scott requested, but we do not have know who is on the list or what is the process to add more people. Scott tried to subscribe from another email and the request was rejected. So it is not fully functional and that is why we have not announced it yet. 

JH: In that case I will send it to a bunch of people through email.

JL: I have something to discuss based on my offline interaction with Scott. Scott has created a workaround around data acquistion. We have the workaround for most of the stuff, but not for everything. It forced a duality between a VAMS and SV. You will need have tools and pieces connected together. Its like a glue. If you can't fix everything, can we have a libraries around the language, which will allow the data exchange?

HA: The larger committee is working on merging VAMS and SV. Whatever work we do here is going to be useful to the larger committee as we will face many of the same issues on the interface of VAMS and SV as they will. So, when we have a unified language, the connections will work smoothly. This is not to say you cannot have the libraries. You could still have those in addition to the unified language.

JL: So, then it is not going to be a split solution? 

HA: Yes, it will not. In the last meeting we discussed on separating the accuracy information from the assertion language. Most everyone was there on the call, except Martin.

ML: Can we summarize that requirement.

DN: The accuracy information inside the assertion is not good. The control should not be inside the language. The assertion should remain just the high level language and how the data should be separated from the accuracy information, otherwise there will be semantic problems for assertions at various different levels. The directives can be given to simulator to modify the accuracy. You could also say that the simulator is aware of assertions, and in that case there would be some kind of automatic/semi-automatic way to tell the simulator to increase/decrease the accuracy based upon the assertion. That would direct the simulator to do finer calculations. It is still unclear how to incorporate this accuracy modification within the language and its interaction with the assertion (which does not contain the accuracy information)

ML: Is this document available on the Twiki?

JL: Yes, Dejan posted this on Twiki, under the meeting minutes.

HA: Ok, so we are still waiting for Ken to refine his section of the requirements. Once, we have those, we should be able to take a vote on which requirements we need to spend more time on. I am hoping we should be able to get the vote in about two weeks from now. Is that a fair amount of time for people to think a little bit more about the requirements?

JL: Yes, that is fair. So, does that mean that Ken will combine all the requirements?

HA: My assumption is that Ken will refine his section of the requirements. John's requirements and your requirements will stay as they are. So, in essence we will have three sub-sections of requirements.

JL: Ok, so does the bind still stay as a requirement since this is going to be targeted towards the future integration of SV and VAMS?

HA: I think yes. bind is there in SV today but is not there in VAMS. So, in order to bind to work across both VAMS and SV, VAMS will have to add it as an extension. Without that extension, even when SV and VAMS merge, it could be that bind only works on the digital portion and not across both VAMS and SV. So, we need it as a requirement.


-- Main.AnandHimyanshu - 05 Aug 2009
Topic revision: r1 - 2009-08-05 - 14:47:13 - AnandHimyanshu
VerilogAMS