Re: [Fwd: Minutes of Verilog-AMS meeting - 31 Jan 2008]

Geoffrey,

With regards to no mention of $random * $rdist_ functions in the 
minutes, we reviewed only the changes that were done as part of LRM2.3 
post draftA. All other changes were reviewed as part of individual 
chapter reviews when the owners modified the particular section. I think 
this section did not get modified during the individual chapter edit or 
during the draft revisions, tho' there were some points which have been 
poorly defined which needs to be addressed.

I agree that given that type_string is optional there should be a 
mention of what the behavior should be if no value is specified. This is 
missing and I am not exactly sure what it should refer to when the 
string is omitted. I think it should be "global".

With regards to the "instance" also I agree that its very vague. I 
remember that this issue has come once earlier and I am trying to search 
for it in the reflector. The reference to the instance here (I believe) 
is, when we have a paramset which has a $rdist function, then each 
instance of the paramset will get a new value for the distribtion 
function. Would it be more appropriate to say 'then an unique value is 
generated for each instance of the paramset which has a distribution 
function'?  (as per example in 6.4.1). This really makes sense only for 
paramsets i think. If anybody has a better wording to clarify this 
particular statement we can include it for the next draft version.

Shalom,

Is there any plans to include constant seed argument (parameter) to 
P1800 standards? This was added to LRM2.2 as part of the device modeling 
enhancements to allow them to be part of the parameter initialization, 
and ideally would not want to remove the feature. We need to figure out 
how it will work when we do a P1800 dot standard with AMS integration.

Regards,
Sri

Bresticker, Shalom wrote:
> 1800 says, "The seed argument shall be an integral variable".
> 
> Shalom 
> 
>> I don't see any mention in the minutes of the $random and 
>> $rdist functions being reviewed.  In particular, the whole 
>> business of the type_string argument is poorly explained in 
>> the current LRM.
>> (What is the default behavior, if no type_string is given?  
>> What does it mean "each instance that references this value"?)
>>
>> It also appears that the text that was added to allow the 
>> seed argument to be a parameter (which is needed if you want 
>> to use $rdist_ to set a parameter, eg the threshold voltage or oxide
>> thickness) may violate P1800's sematics for $random.
>>
>> -Geoffrey

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.