SV-EC Committee Meeting. Monday February 5 2007 11:00am - 1:00pm PST [Minutes distributed for review, to be approved at next meeting] (1212020201020) Day (4815936048825) (0000111111000) Month (8899001122112) (0000000000000) Year (6666666666777) --------- Attendees ---------- (-AAAAAAAAAAAA) Arturo Salz (--AAA-AAAAAAA) Cliff Cummings (AAAAAAA-AAAAA) Dave Rich (AA-A-AAA-AAAA) Francoise Martinolle (-AAAAAAAAAAAA) Mehdi Mohtashemi (AAAAAAAAAAAAA) Neil Korpusik (AAAAAAAAAA-AA) Ray Ryan (AAAAAAAAAAAA-) Gordon Vreugdenhil (AAAAAA--AAAAA) Steven Sharp (--AAAA-A-----) Phil Moorby (---AA-AAA-AAA) Doug Warmke (AAAAAAA---AA-) Stu Sutherland (-AAAA--AAAA-A) Heath Chambers (-AAAAAA-A----) Don Mills (--AA--A---A-A) Jonathan Bromley (--A----------) Logie Ramachandran (----AAAA---A-) Melvin Cardoza (-----A-AAAAAA) Mark Hartoog (-------A-----) Satia (from Intel) (--------AAA--) Rob Slater (Freescale) ^ |------- non-voting meeting ** Minutes taken by Neil Korpusik and Mehdi Mohtashemi ////////////////// Febraruary 5, 2007 ///////////////////////// Agenda: 1. IEEE patent policy ref: http://standards.ieee.org/board/pat/pat-slideset.ppt Move: Cliff - Assume that the patent policy was read Second: Mark Abstain: none Opposed: none passed 2. Review meeting minutes/Notes: --------------------------------------------------- http://www.eda.org/sv-ec/Minutes/SV-EC_Meeting_January_22_2007_Minutes.txt 3. P1800 meeting (and upcoming Feb 20 meeting) ----- LRM review The updated draft is now available. People who wrote the proposals need to review them. SVAC - has two people reviewing each mantis item. SVBC - the person that entered the mantis item and the person that wrote the proposal are doing the reviews. - they are closing them if ok. AI/Mehdi: list of people to do it (said he will do it today) AI/Mehdi: send link and password to Jonathan. AI/Mehdi: Neil, Cliff - Presentation on Mantis 890 to P1800 4. Action items review ------------------------- AI: 890 and related items: Cliff sent out an update to the proposal for the scheduling semantics, with Stu's feedback incorporated. Two versions of figure - Labels - the extra labels in the figure shouldn't say events. Arturo, Neil - prefer first version(version without extra boxes) 5. Continue review and discussion on Mantis 890 and related mantis items ------------------------------------------------------------------------- b) 890 08-31-06 clarifications in program and clocking blocks (Doug) Additional mantis items (there were no discussions on this set) 236 - should be resolved... ##0 597 - [NOTE: look further down in the minutes-- CLOSED] 608 - being clarified. 609 - association of clocking block, need a syntactic form. 1325 - was closed by SVAC - unnamed clocking blocks 1615 - was approved already - allows fork/join_none within functions 564 - Cross-program variable access (Sharp) [recommendation: to close] 551 - Program block interaction with queues (Sharp) Simulation regions versus pli regions - Arturo - they are all simulation regions(not sure what else to call them) Jonathan - #1step may not belong in this section at all. Arturo - leaving it somewhere in this section is probably still useful. Jonathan - is this a complete discussion of scheduling? David - all the user visible things are in here. Arturo - preponed now a PLI region (not quite true since #1step is there) - Some recent SVAC proposals have sampling taking place in the preponed region. mantis 1550, 1473 - time 0 degenerate case (preponed versus postponed) Cliff - #1step - time 0 issues? Doug - there was a discussion in svec that didn't come to resolution Cliff - uninitialized nets are z, variables are x Arturo - preponed - where sim time changes and where sampling takes place Cliff - preponed today is all about PLI in the LRM Arturo - no pli writes allowed in preponed - Proposed (retracted this later in the discussion) preponed (simulation only) pre-active (read-only) Francoise - where can values change? Artuto - Can't write in the following set of regions preponed pre-active <-- problem here with writing values via pli postponed Cliff - pre-active is where HP patented tech would write random values. - would require a patent to be donated. Arturo - there is a predefined behavior for initialization today. Cliff - Lionel has papers to show that bugs at time 0 can be caught. Francoise - there are a bunch of callbacks that can go into Preactive See table at the end of section 9. Arturo - there was no preactive in 1364. Francoise - quoted from the 1364 LRM on how the callbacks get called before the active region events are applied. Arturo - likes the fact that the initial state is a bit ill-defined. Cliff - preponed - make it a simulation event. Jonathan - label for 9.3.3.9 is wrong - mention that sampling in preponed is same as previous postponed Arturo - good idea to split the preponed description (see postponed for an example) Francoise - pli callbacks can occur in the simulation regions (Cliff agreed) Jonathan - 9.3.3.9 readonly pli callback (see 9.3.3.1 wording) Francoise - 9.3.3.1 - remove the note? Neil - move #1step paragraph to preponed section? Cliff, Jonathan - both agreed... Doug - 9.3.1 paragraph on #1step 1st sentence is very long. What are the "equivalent constructs". The part about it being parameterized doesn't work. Jonathan - take out the part about "is a preferred...." Arturo - move the part about being parameterized to the clocking block section. Doug - not really "parameterized" Arturo - take out second sentence as well. Doug - take out "the" at the beginning of the last sentence. Francoise - confusing that we can have pli calls in simulation regions. Arturo - update the description for pli regions. Make it clear that the pli regions are extra regions exclusively for pli callbacks. Jonathan - sentence is tough to undersand (get rid of "the" twice) 9.3.2.4 Arturo - 15.11 "Triggering an event that designates a clk to either an assertion or a cb more than once in the same time slot shall be an error." Francoise - might be tough to detect this. Neil - wasn't too keen on requiring this to always be an error. Doug - 890 is mostly done. Doug has a few questions. Doug reviewed emails and minutes to make sure all addressed. - f2f - #1step issues - not addressed by 890 1473 - #2step allowed? #4.7step? 1550 #1step should always be described wrt postponed. - force or continuous assign to a clockvar? force cb.var1 = 1'b0; Doug decided to make those illegal. - program code can schedule in nba? f2f minutes mentioned this. Steven - those are the "old" rules. Doug - no threads of execution allowed in a package. new() code can run. Arturo - that isn't a thread. - originator of thread not in the package. - call to constructor in package. - Doug had issues with initialization of random numbers, etc.? Doug - declaration initializers? Steven - there is a thread but not a lasting thread for those Dave - declaration initializations are before time 0. - only one thread here - all of the initialization done in it. Doug - not allowed to block Steven - also can't leave any threads around. Doug - 1715 - not in correct form for a proposal. AI/Jonathan - agreed to re-write 1715 Doug - concurrent assert in program scope. Allowed? - the observe region has already passed at this point. Mark - signals in design -- arturo - this makes sense. signals in program action code runs in reactive region Neil - what is the #0 sampling for assertions? Arturo - cb might be looser than assertions section. - bad methodology to put assertions in program, but well defined. Dave - 17.3 assertion sampling - 2nd paragraph Neil - this point was raised in the last f2f AI/Neil - check on what we decided on this (17.3 description of #1step) Doug - 549 - long item on scheduling semantics AI/Cliff - make sure all of 549 is addressed. Doug - An enhancement (which mantis?) - Steven Sharp brought it up. - Steven wasn't actually proposing it.... just reporting it. (wait for a user to request it) Doug - restrictions of stuff in program udf, always added gate instances. (use same rule as for a udp - Arturo) Doug - cycle delay - intra assignment ##1 cb.v <= e; // process blocks (uses default cb) cb.v <= ##1 e; // default or LHS cb for the ##1? - there has been discussion on this, but not agreed to. Steven - no concatenation allowed on lhs (so it is unique) AI/Doug - update of 890 (expected today) 6. Next meetings -------------------------------------- Feb 19 Holiday -- No meeting Feb 20 Face-to-face meeting, at the Mentor site in San Jose 1:30 to 3:30 pm [NOTE: There is a p1800 meeting from 10:00 - 1:00pm] sv-ec can have 1:30 to 3:30pm (it will be same room - Dave) DVCon is during that week Heath - traveling Doug - will be there (hopefully or will be on the phone) Mar 5 Monday, 11:00-1:00pm ========== ADDITION TO the minutes, not part of regular ========= ========== meeting discussions, informational purposes ========= ========== Action Items updated based on February 5 2007 ==== Action items: January 8 2007 ------------------------- LRM and mantis item review, P1800 related items AI: Mehdi: list of people to do it (said he will do it today) AI: Mehdi: send link and password to Jonathan. AI: Mehdi: Neil, Cliff - Presentation on Mantis 890 to P1800 AI: 890 and related items: AI; Jonathan - re-write 1715 AI: Neil - check on what we decided on this (17.3 description of #1step) AI: Cliff - make sure all of 549 is addressed. AI: Doug - update of 890 (expected today)