SV-EC Committee Meeting. Monday December 18 2006 11:00am - 1:00pm PST [Minutes distributed for review, to be approved at next meeting] (1212020201) Day (4815936048) (0000111111) Month (8899001122) (0000000000) Year (6666666666) --------- Attendees ---------- (-AAAAAAAAA) Arturo Salz (--AAA-AAAA) Cliff Cummings (AAAAAAA-AA) Dave Rich (AA-A-AAA-A) Francoise Martinolle (-AAAAAAAAA) Mehdi Mohtashemi (AAAAAAAAAA) Neil Korpusik (AAAAAAAAAA) Ray Ryan (AAAAAAAAAA) Gordon Vreugdenhil (AAAAAA--AA) Steven Sharp (--AAAA-A--) Phil Moorby (---AA-AAA-) Doug Warmke (AAAAAAA---) Stu Sutherland (-AAAA--AAA) Heath Chambers (-AAAAAA-A-) Don Mills (--AA--A---) Jonathan Bromley (--A-------) Logie Ramachandran (----AAAA--) Melvin Cardoza (-----A-AAA) Mark Hartoog (-------A--) Satia (from Intel) (--------AA) Rob Slater (Freescale) ^ |------- non-voting meeting ** Minutes taken by Neil Korpusik and Mehdi Mohtashemi ////////////////// Dec 18, 2006 ///////////////////////// Agenda: 1. IEEE patent policy ref: http://standards.ieee.org/board/pat/pat-slideset.ppt Neil: adopted new policy that will go into effect Move: Cliff - Assume that the patent policy was read Second: Heath Abstain: none Opposed: none passed 2. Review meeting minutes/Notes: a) Review meeting minutes of October 23rd 2006 October 23 2006: http://www.eda-stds.org/sv-ec/Minutes/SV-EC_Meeting_October_23_2006_Minutes.txt Move: Neil - Approve minutes Second: Gord Abstain: none Opposed: none passed b) Review meeting minutes of December 4 2006 (not yet available on the web) 3. Action items review ------------------------------------------------ AI on 890: (and related) Francoise: contact svcc on pli regions Mentioned it to svcc, no concrete proposal yet, asked to forward. bring the new scheduling semantics proposal to the svcc when available. Doug: start with the changes in the 890 mantis (rev 5) Arturo, Cliff : will work on the event scheduling portion. Doug : rewrite the description of #1step in 17.3 (shouldn't be too tough) Steven - previously proposed a change in how to determine which regions to schedule events in. - thinks that the questions sent out by FM cover all the Cadence concerns. Neil - Split off #1step into a separate mantis item (Steven has one already for #1step - reuse it - mantis xxxx) 1473 - svbc - Shalom Mehdi - talk to svac about the change to #1step, svac review it before we vote. AI: 553/1371: Dave Close 553 and place content of 553 in 1371. Action items: Previous meetings AI: follow-up to mantis 978 Gord - re-write section 5.15.4 - only the default argument value makes sense Dave - what order is used for find and find_index (e.g. traversal is left to right of the bounds, associative is min to max) - 5.15.1 - change last sentence in first paragraph. All - can anyone come up with a realistic example? (last example 5.15.4) - not only where you would need it but also what does it mean. 4. Name Resolution sub-group update ------------------------------------ Gord: no update. 5. Continue review and discussion on Mantis 890 and related mantis items -------------------------------------- 890 08-31-06 clarifications in program and clocking blocks (Doug) Cliff: sent a response to Neil's question with respect to proposal. event-scheduling-2005_dec5_2006. Update for section 9- scheduling semantics 9.3 pre-ReNBA, post-ReNBA, post-ReNBA before pre-postponed regions this is two PLI regions. Do we need them both? Arturo: we do not need both of them. Francoise: this is not symmetric to design cycle. Gord - events in the Post-re-NBA could change a module variable - Thinks that there is an ordering issue here. There could be an observable difference between what we have in the current proposal and what we would have if we remove the Post-re-NBA. Not sure if the difference is useful though. If toss one, it should be the Pre-postponed region Rob - May need both if using a golden model written in C. The testbench may want to communicate with a C routine and not let modules know about it. Gord - allows for a detailed level of control, but you need to know what you are doing. Arturo - it will be tricky for PLI writers. AI/Rob - send a summary of the use model to the alias. Neil - Because of table 9-1 we do need to leave in Pre-postponed. Cliff - will use blue for new regions in the diagram. - Dave's feedback - will be done - it was due to a cut and past issue. Cliff - capitalization First letter and also first letter of text after '-'. Cliff is willing to make sure all sections of the LRM are consistent in the capitalization used for the scheduling regions. Francoise - this set of changes need to be approved along with 890 (agreed) AI/Neil - open a mantis item to cover this set of name changes e.g. capitalization only (see mantis 1701) AI/Mehdi - contact svac, svcc about the proposed changes to section 9. 890 - questions from Francoise Arturo - main issue seems to be how to switch scheduling regions Steven - you don't! Gord - will pli be able to do an update irrespective of where running? Can it specify a region? Various vpi possibilities are: 1. schedule as though a regular language task enable 2. specifically scheduled in NBA 3. specifically scheduled in re-NBA Arturo - we could add an explicit ability to switch regions Steven - a qualifier could be added, or just use an initial block. Gord - most likely wouldn't object to such a language extension. Steven - wants scheduling based on the region of current execution. Neil - what about clocking drives? Gord - not affected. Act like an implicit procedural assign to target. - All "normal" constructs don't see any observable change with the change proposed by Steven. - Is prepared to back Steven's suggestion. Arturo - basic change - property of engine in Steven's proposal. Gord - used to have to refer to originating thread. Wasn't well defined. Steven's proposal is more precise. Rob - The proposed change is not backward compatible. - Not sure of the use-model for the program block. - There could be an issue with combinational logic, how verify it? - Verilog testbenches used for his projects - had very few races in the past 2 years. - Doesn't understand why a new region is required for testbench. Gord - Divergence of implementations is part of the problem for users. Steven - What is being discussed now is more complex. Most users may not understand it very well. It doesn't appear to be one language. Arturo - Don't need a separate region to integrate with PLI. Assertions triggered by multiple clocks; need a separate region to deal with this situation. How write a testbench that is waiting for assertions that use multiple clocks? How know if multiple clocks trigger at the same simulation time? AI/All - Cliff wants to know what issues are not backward compatible. AI/Steven - examples of problems - so we can prepare. - doesn't fully understand the use-model, but can comment on language domain issues. Neil - clocking drive - in reactive (from program block) Gord - Driving a net - continuous drive in reactive region. Anything activated due to a value change caused by the clocking drive, will wake up in what region they are executing. For event controls in modules --> propagate values in module regions. If a program block is waiting on that net - it will wake up in the reactive region. Steven - When have multiple drivers what happens? Arturo - Doug's proposal - drive resolution happens right away. Gord - thinks that the combination of Doug's writeup and the changes from Cliff will give enough detail for everyone to fully understand. Arturo - we are only changing a corner case (task calls). Gord - there is enough ambiguity with the existing LRM to argue against requiring full backward compatibility. Rob - Possible agenda for Jan 8 - use model for program block. Rob wants to know what he can do with a program block that he can't do with other constructs. AI/Cliff - send out an example use of the program block. 6. Email bouncing off the reflector --------------------------------------- Since November we have been seeing problems with email bouncing off the reflector. At least the SVEC and SVBC are experiencing these problems. David Smith has been working with Stanford support to address this. There are plans to update the version of the mail tool being used. Cadence and Synopsys addresses seem to be bouncing more than others. Neil has started forwarding bounced email to the Cadence and Synopsys addresses when they bounce. Mentor addresses are also having problems receiving emails but they don't seem to actually get bounced back to the owners of the aliases like the Synopsys and Cadence emails do. Cliff - is able to bring up his email in a web browser and search for stuff. Gord - dropped off 7. Additional mantis items ------------------------------------- Additional mantis items 236 - should be resolved... ##0 239 - will close it 597 - AC ; close this one 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] 553 - Program block exit semantics (Sharp) 551 - Program block interaction with queues (Sharp) 236 - should be resolved... ##0 -- will be covered by 890 close after 890 has been voted 239 - New syntax for cycle delay @clocking block name would be prefereable. Example: ##5 @(posedge clock); // wait 5 posedge clock cycles Neil - @cb_name was mentioned in the F2F as being preferable to the proposed syntax in mantis 239. Cliff - ##5 @cb_name Is this what we were thinking? - using a dot in here would create ambiguity Dave - @cb_name is also an event, how know that ##5 isn't for the default clocking block? Steven - might as well just use repeat since the new syntax is getting so long. Move: Cliff - close mantis 239 - can use other syntax Second: Steven Abstain: none Opposed: none passed 239 closed 1519 - 5.15.3 sum() gives undesired results (Shalom) Move: Dave - close mantis 1519 - covered by 978 Second: Cliff Abstain: none Opposed: none passed 1519 closed 8. Next meetings -------------------------------------- January 8, 2007 Monday, 11:00-1:00pm AI/Mehdi - contact svbc to see if they can also move to Jan 8. ========== ADDITION TO the minutes, not part of regular ========= ========== meeting discussions, informational purposes ========= ========== Action Items updated based on November 20 2006 ==== Action items: November 20, 2006 ------------------------- AI: 890 and related AI: Rob - send a summary of the use model to the alias. AI: Neil - open a mantis item to cover this set of name changes e.g. capitalization only (see mantis 1701) AI: Mehdi - contact svac, svcc about the proposed changes to section 9. AI: All - Cliff wants to know what issues are not backward compatible. AI: Steven - examples of problems - so we can prepare. - doesn't fully understand the use-model, but can comment on language domain issues. AI: Cliff - send out an example use of the program block.