SV-EC Committee Meeting. Face-to-Face Tuesday February 20 2007 1:30pm - 3:30pm PST [Minutes distributed for review, to be approved at next meeting] (12120202010202) Day (48159360488250) (00001111110000) Month (88990011221122) (00000000000000) Year (66666666667777) --------- Attendees ---------- (-AAAAAAAAAAAAA) Arturo Salz (--AAA-AAAAAAA-) Cliff Cummings (AAAAAAA-AAAAAA) Dave Rich (AA-A-AAA-AAAAA) Francoise Martinolle (-AAAAAAAAAAAAA) Mehdi Mohtashemi (AAAAAAAAAAAAAA) Neil Korpusik (AAAAAAAAAA-AAA) Ray Ryan (AAAAAAAAAAAA-A) Gordon Vreugdenhil (AAAAAA--AAAAA-) Steven Sharp (--AAAA-A------) Phil Moorby (---AA-AAA-AAAA) Doug Warmke (AAAAAAA---AA-A) Stu Sutherland (-AAAA--AAAA-A-) Heath Chambers (-AAAAAA-A----A) Don Mills (--AA--A---A-AA) Jonathan Bromley (--A-----------) Logie Ramachandran (----AAAA---A--) Melvin Cardoza (-----A-AAAAAA-) Mark Hartoog (-------A------) Satia (from Intel) (--------AAA---) Rob Slater (Freescale) (-------------A) Alex Gram (Mentor) ^ |------- non-voting meeting ** Minutes taken by Neil Korpusik and Mehdi Mohtashemi ////////////////// Febraruary 20, 2007 ///////////////////////// Agenda: 1. IEEE patent policy ref: http://standards.ieee.org/board/pat/pat-slideset.ppt Move: Dave - Assume that the patent policy was read Second: Neil Abstain: none Opposed: none passed 2. Review meeting minutes/Notes: --------------------------------------------------- http://www.eda.org/sv-ec/Minutes/SV-EC_Meeting_February_5_2007_Minutes.txt Move: Arturo - approve the meeting minutes of February 5th, 2007. Second: Jonathan Abstain: none Opposed: none passed 3. P1800 meeting update (today, 2.20.2007) ----------------------------------- Mehdi gave a brief overview of the changes contained in mantis 890 to the P1800 meeting that was held earlier in the day. There were a total of 3 mantis items that had problems in the new draft. Input was sent to Stu. 4. 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) 1717 - 1715 - NOTE: We went through various issues during discussion on 890, Doug was making changes to the proposal as discussion moved on. The result of those changes has been posted as version 7 of the proposal for mantis 890 - part 1. Cliff's changes will become part 2. Stu agreed that it would be ok to have to separate pdf files for a single mantis proposal. Gord - Sent email before the meeting pointing out that when an input clockvar has #0 skew, the value isn't updated until the Observed region. This seems to imply that @(cb) isn't always the same as @(posedge clk). When a cb is used within a module there is a visible difference between the two. Arturo - #0 is a special case When @(cb) occurs, could be different, depending upon whether there was a #0 input skew specified. He has been assuming that you wouldn't want to mix input clockvars with #0 skew and input clockvars without #0 skew in the same cb. Gord - wants it to be well defined. - whether the cb event occurs before or after the NBA region is an observable difference. Jonathan - We could simply say that the cb event always occurs in the Observed region. - Is there a coverage issue? Gord - We just need to make sure that everything is well defined. - "The event triggered by the cb occurs during the Observed region". - We also need to continue to mention that that clockvars are updated before the cb event occurs. (15.12) Jonathan - assertions in a cb - now execute in the Observed region. This would cause a race condition. Gord - he is ok with that. Mantis 1550 - from the svac. It describes the semantics for $past at time 0. <= ##delay // the ##delay on the RHS uses cb of LHS (page 222). Take out the reference to transport delay. There is very little information in either LRM on this topic. Arturo - when there is a cb in an interface - how do we know which of the two NBA regions will be used? - virtual interfaces also need to be taken into account. Dave - the target doesn't determine if NBA or Re-NBA region - the location of the cb determines this. Gord - there could be dynamic scheduling issues. - whoever is blocking on it determines when it wakes up. Arturo - can't decide if NBA or Re-NBA based on the thread. clocking blocks are allowed in programs, modules, interfaces. Gord - cb in an interface - always NBA? - can't assign to a clockvar within an interface. - the drive must occur in the context of the location of the cb. Arturo - we only have an issue with interfaces. Dave - some users put RTL in interfaces. Gord - driving in the NBA versus driving in the Re-NBA can cause a visible difference. A program could react to a change in value made during the Re-NBA (before going back around the big loop). Jonathan - can we have a cb always drive values in the Re-NBA? This idea was agreed to. Gord - has a big concern with continuous assignments within an interface. AI:Neil - send the set of changes needed for the scheduling semantics to Cliff. 5. Next meetings -------------------------------------- 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 20 2007 ==== Action items: February 20 2007 ------------------------- AI: 890 and related items: AI; Neil - send the set of changes needed for the scheduling semantics to Cliff.