Re: [sv-ec] 890 scheduling algorithm questions

From: Clifford E. Cummings <cliffc_at_.....>
Date: Mon Apr 16 2007 - 18:04:33 PDT
Hi, Francoise -

We discussed your email in the SV-EC meeting today. Arturo answered 
most of the questions in his previous email response.

Concerning the feedback arrow from the Post-Observed PLI Region. We 
added this based on the idea that the PLI committee might want to do 
a callback that would write a variable that could trigger additional 
activity. We also added the note:
NOTE - The PLI currently does not schedule callbacks in the 
Post-Observed region.

We are going to leave the diagram as is as part of Mantis 890 for 
inclusion in Draft 3, but if the CC committee decides that there 
should be no feedback from the Post-Observed Region to the Active 
Region, feel free to file a post-draft 3 Mantis item to have it removed.

Regards - Cliff

At 10:44 AM 4/11/2007, Francoise Martinolle wrote:
>The sv-cc discussed the 890 scheduling algorithm picture today and
>Michael Rohleder asked the following questions:
>
>1) There is a loop back to the active region from the pre-NBA, NBA 
>and post-NBA
>regions. This presumes that when a PLI callback wakes up in the 
>pre-NBA region for ex,
>the pli user code may schedule an event into the active region queue.
>The interpretation from the picture lead us to understand that if 
>that it is the case (an event
>has been scheduled from the pli pre-NBA region in the active 
>region), the active region queue
>is processed once the pre-NBA events have been processed, before 
>moving to processing the NBA region
>events.
>
>Michael wants to get the confirmation that this is the correct 
>interpretation and that has not changed since 1364.
>
>2) same issue for the loop back for the NBA and post-NBA regions
>
>3) There is a loop back from the observed and post-observed regions 
>to the active region
>There is no loop back from the pre-observed to the active region and 
>the text explicitly states that it is illegal to write
>values or schedule an event from that region:
>
>9.3.3.5 Pre-Observed region
>
>"it is illegal to write values to any net or variable or to schedule 
>an event within
>
>the current time slot."
>Why this discrepancy?
>If those regions are "read only" as the text states , why is there a 
>loop back to active?
>
>Extract from mantis 890:
>
>9.3.2.5 Observed events region
>
>The Observed region is for evaluation of property expressions when 
>they are triggered.
>
>During property evaluation, pass/fail code shall be scheduled in the 
>Reactive region of
>
>the current time slot. PLI callbacks are not allowed in the Observed region.
>
>9.3.3.6 Post-Observed PLI region
>
>The Post-Observed region provides for a PLI callback control point 
>that allows PLI
>
>application routines to read values after properties are evaluated 
>(in the Observed or an
>
>earlier region).
>
>Francoise
>        '
>
>--
>This message has been scanned for viruses and
>dangerous content by <http://www.mailscanner.info/>MailScanner, and is
>believed to be clean.

----------------------------------------------------
Cliff Cummings - Sunburst Design, Inc.
14314 SW Allen Blvd., PMB 501, Beaverton, OR 97005
Phone: 503-641-8446 / FAX: 503-641-8486
cliffc@sunburst-design.com / www.sunburst-design.com
Expert Verilog, SystemVerilog, Synthesis and Verification Training


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.
Received on Mon Apr 16 18:05:02 2007

This archive was generated by hypermail 2.1.8 : Mon Apr 16 2007 - 18:05:31 PDT