Subject: ITC 4/26/02 confcall minutes
From: m.vitale@philips.com
Date: Thu May 02 2002 - 08:44:37 PDT
I apologize for the late minutes, but the last confcall was squeezed
between a visit to the dentist and one to the
ER and only now could get back to it. I'm sure I'm missing somebody in the
attendence list, just let me know and
I'll update it.
Best regards,
Maurizio
ITC conference call of 4/26/02
minutes taken by Maurizio Vitale
Participants:
Jason Andrews Axis
Richard Sayde Cadence
Duaine Pryor Ikos
Jan Johnson Ikos
John Stickley Ikos
Linda Prowse Fosler Ikos
Maurizio Vitale Philips
Todd Massey Verisity
Dave ? Verisity
Excused:
Brian Bailey Mentor Graphics
Joseph Bulone ST Microelectronics
Stephane Guerineau EVE
Luc ? EVE
David ? EVE
Vinod Empranthiri CoDesign
PRIORITIZED LIST OF INTERESTS
=============================
Dave will work on his list and send it
Richard will send his list
action Brian&Maurizio:
collect all the lists received so far and put them together for
group approval (by next confcall, as we have to provide a roadmap
of activities at/around DAC)
REFINEMENT OF ISSUES
====================
Richard identified two issues for emulators:
- clock handling
two options are clock from emulator and clock from target.
clocking from the target is more difficult and the group agreed
to concentrate on clock from emulator. A terminology issue has
also been identified: in emulator parlance, target is the system
in which the emulator is plugged in for emulating a specific
IC/IP,
so we need a better term for the target prototyping system.
- reset
how to get the system into a known initial state.
Richard mention that downloading the information is normally
easy
(albeit done in a vendor-specific way) while actually starting
the
emulator from a known initial state may be trickier.
Jason[axis] mentioned a problem in understanding the issues with emulators
for
people outside the emulation companies.
Jan(or Linda) asked if a reference implementation of SCE-MI might be
helpful
and the group agreed. Last time Ikos investigated the issue, they come out
with
an estimate of around $75k for developing such an implementation. Next
meeting
will have in agenda a point for discussing ways for getting that amount of
money.
Maurizio asked for a description of the deliverables that people can
use to justify to their management this type of expense.
Another issue identified [don't have a name for the person who asked.
Jason?]
is related to whether we want to have indipendedent runs be the same cycle
by
cycle.
Duaine answered that if you want top performance you have to give up full
reproducibility across runs, but complete reproducibility can be achieved
if
lower performance is acceptable.
Richard added that Cadence might have proposals in this area.
action Maurizio:
add the identified points to the spreadsheet
The last issue identified in this session is related to the dumping of
debug information. Duaine explained that due to the sheer size of the
dataset and the speed at which things need to be dumped a compression
system (and often a space/time filtering that selects interesting
portions) is mandatory. Each company has their own scheme here and the
easiest thing is to try to find a common API while preserving each
company propietary solution under it.
Duaine said that there're many APIs available (one from Synopsis) and
that would be probably best to select one and have vendors implement
it for their system.
DAC ANNOUNCEMENTS
=================
Verisity and Cadence are still waiting for an answer on whether thei
feel appropriate to officially announce SCE-API plans at DAC.
Axis is ok with a joint announcement, although they won't have a
product of their own ready by that time.
MISC
====
NEXT MEETING
============
the group find a meeting before DAC desireable. A date has not been
set waiting for Brian return. Also a lot of people were absent during
this last call
action Maurizio
find possible dates and propose them to the reflector
AGENDA ITEMS
- group agreement on list of priorities
- fund raising for reference implementation
- decision on DAC announcements
- confirmation of a DAC face-to-face meeting on Monday morning
(room availability etc)
Maurizio Vitale
Philips Semiconductors
System and Software Design - Innovation Team
1721 Moon Lake Blvd. - suite 500 ph: +1 (847) 843-4562
Hoffman Estates, IL 60194
This archive was generated by hypermail 2b28 : Thu May 02 2002 - 08:51:58 PDT