Minutes from OVI RTL Synthesis Standards Working Group Meeting (June 16, 1998)
Sudhir Kadkade (Sudhir_Kadkade@SiFR.com)
Tue, 28 Jul 1998 17:38:06 -0700
Minutes from OVI RTL Synthesis Standards Working Group (June 16, 1998)
Venue:
Mentor Graphics, San Jose.
In attendance were:
1. Doug Smith (djsmith@veribest.com) (205) 730 8808
2. Ganapathy Subbaraman (subbu@eng.adaptec.com) (408) 957 6964
3. Ken Coffman (kcoffman@sos.net) (360) 422 7520
4. Edward Arthur (earthur@lucent.com) (508) 303-8885 x217
5. Sudhir Kadkade (Sudhir_Kadkade@SiFR.com) (503) 675 0121
6. Yatin Trivedi (trivedi@seva.com) (510) 249 9085
7. Ambar Sarkar (ambar@viewlogic.com) (508) 303 5493
8. Don Hejna (djhejna@ambit.com) (408) 566 8024
9. J Bhaskar (jbhaskar@bell-labs.com)
Scope of working group
This working group was set up in April 1998. Subsets have been donated by
Synopsys, Cadence and Viewlogic.
Bhaskar proposed that the scope was to review the syntax and semantics
of the proposed synthesis subset of Verilog (1364-1995) for digital design.
This can be found on the web page set up for the group.
Only clarification was to note the standard by including the year 1995.
No objections were raised.
Purpose of the standard
Bhaskar reiterated the purpose of the proposed standard - "portability
across
vendors". No objections were raised.
Working Group Information
WG reflector is vlog-synth@eda.org
Home page is http://www.eda.org/vlog-synth
Administrator is dbishop@kodak.com
Working Group Policies
Bhaskar made a proposal for a set of policies for participation in and
activities of
the Working Group. These will be updated as per the discussion in the
working
group and put on the reflector.
Handover
Ambar Sarkar handed over the chair tasks to J Bhaskar. He also announced
that Synopsys would not participate.
Meeting Schedule
Face-to-face : ICCAD (after BoF)
IVC (after BoF)
Teleconference : August (1st week)
Oct (1st week)
Dec (2nd week)
BoF: ICCAD
IVC
General Issues on Draft 1.0
1. A proposal was floated to break the subset into 4 interim proposals.
MIN - an intersection of all synthesis subsets
SYN - the defacto standard
MAX - an union of all synthesis subsets
WLI - wish list features
This was voted down, in favor of focussing on a single standard and
enumerating features it supported.
2. Task leaders for various sub-groups to review the proposed subset.
Ken Coffman - Semantics
Don Hejna - Syntax
Doug Smith - Attributes and Pragmas
3. On section 5
It was noted that more pragma support exists and may be desirable to
standardize than in the standard. The pragmas discussed were map_to_mux,
state encoding and preservation of hierarchy at the always block level.
In response to the mux inference pragma, a question was raised as to whether
that amounted to standardize the netlist. This was felt to be an extension
of
the scope of the working group and the scope should be restricted to model
portability.
4. A2.1
* pg 21 - A proposal to remove $readmem from the supported list was
approved.
It is to be moved to the wish list.
* pg 25 - what type of NETTYPEs will be supported?
supply0, supply1 moved into supported set. wand, wor will be moved into the
unspecified category.
* Procedural continuous assignments are not supported. Discussions on this
topic will be put off until later.
5. There was a discussion about compilance about what it means to a tool
vendor.
a. Not supported means Not specified. (i.e. vendor tool can do anything,
including
print an error)
b. There is no illegal set.