SV-EC Committee Meeting Monday March 27, 2006 11:00am - 1:00pm PST [Minutes distributed for approval at next (sv-ec) committee meeting] (102112 ) Day (993337 ) (100000 ) Month (211233 ) (000000 ) Year (566666 ) --------- Attendees ---------- (AAAAAA ) Arturo Salz (-AAA-A ) Cliff Cummings (A-AAAA ) Dave Rich (AAAAAA ) Francoise Martinolle (AAAAAA ) Mehdi Mohtashemi (AAAAAA ) Neil Korpusik (AAAAAA ) Ray Ryan (--AAAa ) Surrendra Dudani (-AAAAA ) Gordon Vreugdenhil (-A--AA ) Phil Moorby (AA---- ) Doug Warmke (---AAA ) Stu Sutherland (----AA ) Steven Sharp (-----a ) Don Mills (Aa---- ) Chris Spear (------ ) Brad Pierce (------ ) Karen Pieper ^^------------ Considered ONE MEETING. ** Minutes taken by Neil Korpusik and Mehdi Mohtashemi ////////////////// March 27, 2006 ///////////////////////// 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: Gord Abstain: none Opposed: none passed 2) Review Meeting minutes, of March 13th, 2006 meeting http://www.eda.org/sv-ec/Minutes/SV-EC_Meeting_March_13_2006_Minutes.txt Move: Neil - Approve minutes March 27, 2006 Second: Dave Abstain: none Opposed: none approved 3. Continue with Mantis items: 3a) Items from Dave's email left from last meeting 1306 foreach iterates over unpacked arrays Gord - last sentence, refers to changing the dimensions of a dynamically sized array. Dave - can get an invalid index and should be treated that way Neil - this mantis isn't changing functionality, simply further clarifying what is already there. Dave - last statement was added, all else are just clarifications. Steve - what if no loop variables are specified? would there be any statement execution? (i.e. once) Gord - there is no formal definition of skipped Friendly amendment - add a period to end of last sentence. Move: Dave - Approve 1306 with friendly amendment Second: Gord Abstain: none Opposed: none approved AI: Dave to update the proposal with the amendment. AI: Steven: create a new mantis item for the case foreach where no variable has been specified. AI: Ray a bug note to this mantis item. 1315 User specified default on associative array access Dave - returning the default initial value applies to all index types. Gord - if user specifies a default it gets really noisy if a warning is generated. Dave - reading a default value is not an invalid entry Specifying a default value doesn't actually create all entries. Can still use the exists() method to check for an associative array entry. Arturo - removed the part about a warning when returning a default value Neil - typo in last sentence. Defined --> Defining Cliff - typo effect --> affect Ray - invalid index - invalid value or out of range? Arturo - invalid index and no default specified - "may issue a warning" Legacy issue with out of range values not issuing a warning Gord - should a warning be required? Ray - usually this type of thing is implementation specific Stu - there are other places where a warning is required. Arturo agrees modifications: warning may --> warning shall friendly amendments: change: warning may be issued --> warning shall be issued Defined --> Deining effect --> affect Move: Dave - Approve 1315 with the above friendly amendments Second: Cliff Abstain: none Opposed: none approved AI: Dave update the proposal to update. 1330 Missing BNF for parameterized virtual interfaces AI: Dave update the proposal, create an example. 3b): 1393 03-23-06 Typo in set membership constraint example Move: Dave - Approve 1393 Second: Cliff Abstain: none Opposed: none approved 1357 Instance-specific default coverage goal should be 100 (coverage) Move: Arturo - Approve 1357 Second: Cliff Abstain: none Opposed: none approved 1279, clarify covergroup/point/cross Ray: will complete the proposal by next time 884 incorrect interpretation of expression in constraint example AI: Francoise, get the propsal and upload database, we will vote on it next time 270: enhancement (keep it as enhancement) Neil - enhancement request Mehdi - leave it as an enhancement request 262 randcase unclear (assigned) & 261 randcase width rules (assigned) Neil - no proposal Ray - treat weights as being unsigned - was discussed before Steve - treat all evaluations as unsigned? is signedness determined up-front or at the end? Arturo - each expression is self-determined Gord - RHS remains signed and treated as unsigned at the end Steve - (-1 >> 30) = 3 or a big number? - what is the desired behavior? - there have been some cases that were different from normal verilog rules, so we need to define this very carefully. Gord - aren't we in agreement? Ray - not really an operator but similar Steve - like a big super-compare operator Gord - if any of them are unsigned then all are treated that way - today it says each is self-determined - this is well specified - how combine them? (not well specified) Steve - Verilog assumes you coerce early on and therefore has no rules for doing coercion at the end. Arturo - do each expression separately? (see mantis 261) - sum is done using standard precision - need to specify that there is an implied assignment to a temporary Gord - why are the rules different here from Verilog? Ray - a lot of times the weights will be a literal integer Arturo - this whole discussion is definitely about corner cases Need to combine the two mantis items 261 and 262 on 261: Steve - what if all weights added together equals 0? - how are x-values handled? (individually or at end?) Gord - 2-state values for final weight values makes sense - compute all temps as 4-state - if sum is x - issue an error Arturo - if see an x - could choose one randomly. Can't compute distribution if there are x-values. Gord - could use a cast to an int to cast-away any x-values - let's make it a hard error if it is really such a rare case Steve - since we aren't sure what is the correct behavior we should issue an error if there is an x-value (even though x values usually are ok) Arturo - it is well-defined in constraints. It fails if there is an x-value Gord - expressions retain their 4-stateness until look at the sum. If x-value then it is non-computable. Steve - constraining a 4-state value versus getting a value out of randomizer Gord - would like to have a known alg. (even if it is arbitrarily decided) - 4-state temps - if sum has an x - it is an error Arturo - prefers to just select one if there is an x value Ray - even distribution? (doesn't like that idea) - for constraints you don't change the value when there is an error Neil - prefer first one or none chosen when see an x value - we don't want different implementations to do different things Steve - doing none is probably better Steve - his summary: - x anywhere the sum goes to x and no paths taken - each expression is self-determined - each assigned to a temporary of the same width and then summed Gord - overflow - on your own - a user error There was general agreement if there is an x-value that none are selected. Need to combine the two mantis items 261 and 262 AI: Steven, create a new proposals for both of these mantis items. Should address both 261 and 262. and vote on it next time. 247 -- enhancement Ray: on this static randc, rand variable static, is the seed and cyclicality shared ? among all instances of the class random number generator would have their own state is cyclic state specific to an instance or static. Arturo: it should be static, where is the seed coming from, any particular object. Francoise: do we allow static rand. Ray: rand will take effect on all instances. Arturo: this mantis item may need to be split. Mehdi: leave as an enhancement for now. Ray: static randc - what if multiple instances? Cyclic state also static? Neil, Mehdi: assume everything is shared. Arturo: this is not well specified. AI: to create mantis static randc. Ray - create new mantis with a proposal for this aspect of it. - create a new mantis for the stuff at the bottom (enhancement) If it is necessary. Motion: Ray - close 247 and open a new mantis on static randc issue Second: Arturo Oppose: none Abstain: none Passed unanimously 211 built-in scope randomize (assigned) Ray: class randomize and scope randomize related issues is that any different than class_randomize.variable. same restrictions about variables. the example in the proposal is broken. will defer for next meetings with the randomization items Gord and Ray are working on, more detail proposal will come about. 134 4 state values as constraint (enhancement) Mehdi - keep this one as an enhancement. Miscellaneous: -------------- 1308 abstract classes clarifications Gord - what is the type of 'this' - where the override happens - the correct way to visualize is to think about what the this is referring to. AI: Neil - create a new mantis item for this issue on 7.13 AI: Mehdi - set up an email vote for this. 4. Next meeting proposal: April 10, 2006 =================== Resolved Mantis items ============================= Resolved Mantis items (by dates) -------------------------------------------------------------------- March 27, 2006 (state) -------------------------------------------------------------------- 1306 approved acknowledged 1315 approved acknowledged 1393 approved acknowledged 1357 approved acknowledged 247 Closed --> closed 270 remains as enhancement 134 reamins as enhancement -------------------------------------------------------------------- March 13, 2006 (state) -------------------------------------------------------------------- 714 approved assigned 892 approved assinged 1242 approved assigned [ one abstain] 1243 approved assigned 1244 approved assigned 1263 approved assinged Reviewed items from other committees 1325 SV-AC February 13, 2006 -------------------------------------------------------------------- 138 approved assigned enhancement 251 - no action - assigned enhancement 253 approved closed 1178 approved closed 1237 approved assigned 1239 approved assigned 1240 approved assigned January 9, 23 2006 items -------------------------------------------------------------------- 344 closed 644 closed 683 approved closed 712 approved closed 722 approved acknowledged 725 approved closed 1236 approved assinged 1238 merged with 1237 closed ======================================================================== ==== ========================================================================= ================== Action Items ============================ Action items: March 27 2006 ---------------------------- AI: 1306: Dave: to update the proposal with the amendment. Steven: create a new mantis item for the case foreach where no variable has been specified. Ray: Add a bug note to this mantis item. AI: 1315 Dave - update the mantis proposal AI: 1330 Dave - Create an example for this mantis AI: 1279 Ray - still needs a proposal AI: 884 Francoise - create a proposal AI: 270 Mehdi - stays as enhancement AI: 261/262 combined Steven, put the proposals together, should address both 261 and 262. and vote on it next time. AI: 247 Ray - close 247, create a new mantis item for static randc issue AI: 134 Mehdi- keep this as an enhancement AI: 1308 Neil - create a new mantis item for this issue on 7.13 AI: 1308 Mehdi - set up an email vote for this. Action items: March 13, 2006 ------------------------------ AI: 1325 - (sv-ac) Neil to send input to the SVAC on sv-ec feedback AI: new item - Gordon to create a new mantis item for randomize() method call item AI: 1306 Dave - redo the proposal for next time AI: 1330 Dave add example to the proposal AI: Mehdi ask Karen about the new state for the editor to view. Action items: February 13, 2006 -------------------------------- AI: 714 Neil to close, Mehdi to find out about which state it needs to have 714 to be added into the LRM. AI: 1236 Ray - add in the text change for gc-->cg for this mantis AI: 1237 Ray to update with friendly amendment AI: 1239 Ray to make the updates to the proposal. AI: 253 Gord to close the mantis AI: 138 Mehdi move this into enahancement AI: 1279 create some proposals on these issues. Send out some ideas for discussion. Create a few examples for discussion AI: 1313 Gord to put the proposal for bin expression AI: 1178 Mehdi to close this mantis AI: 1308 Dave to update the proposal AI: Ray - add a new Mantis item for the visibility issue. Action items: January 9, 2006, January 23, 2006 AI: 683 Gord to close AI: 712 Gord to close AI: 725 Gord to close AI: 1236 Ray to update the proposal with friendly amendment AI: 1237 Ray to create the proposal, to be discussed later on and voted on in the next meetings AI: Gord to create mantis item for type expressions, comparisons in covergroup AI: 1238 Ray to close (and merge) with 1237 once the proposal is there AI: 1239 Ray & Gord to place a new proposal based on 1/23/06 discussions AI: 1240 Ray; update with clarifications for what has affects. AI: 714 Neil, re-open, make the corrections in the proposal