I'm a little new to this BIRD (I *have* read it <wink>), but I am not
clear on why the Tx/Rx reserved parameters need be segregated -- and
mandated to be checked as such. The EDA platform should know which (Tx/Rx)
it is intending the calls to the executable to be for, and use or ignore
Reserved_Parameters as they apply.
My primary objection to explicit checking lies in the stated fact that the
checklist is a "living table" which must be updated with every new reserved
parameter BIRD forever and ever, world without end. If this is really
necessary, a more explicit explanation of the reason why in the BIRD's
preamble would be useful.
A secondary motivation for not over-constraining the utility of I/O
executables lies in my interest in Tx-Rx interaction (e.g the "BCI" BIRD we
are discussing). As the tech lead of an AMI modeling team, the possibility
of combining the Tx and Rx functions into not only one executable *but one
model structure* (made possible by a re-entrant AMI_Init if we actually go
there in the BCI BIRD) would allow end-to-end optimization without the need
for the EDA platform to act as the mailman in cases where the Tx and Rx
already live together in the same flat. I could see the usefulness of a
defined *Model_Specific* parameter "Direction" (In|Out) so the EDA platform
can tell the executable what it wants it to be when it calls it (especially
if the model structure is shared).
I realize that this may be an unwelcome diversion to this BIRD's true
intent, and if so I apologize and (virtually) sit down. But it intrigues
me. We seem to be beating around the same bush.
Regards,
Bob
On 04/29/15 08:39, Mike LaBonte wrote:
This is now posted to the IBIS-ATM work archive:
*DATE* AUTHOR
<http://www.eda.org/ibis/macromodel_wip/archive-author.html> ORGANIZATION
<http://www.eda.org/ibis/macromodel_wip/archive-org.html> TITLE
<http://www.eda.org/ibis/macromodel_wip/archive-title.html> FORMATS
28-APR-2015 Michael Mirmak Intel Corporation AMI Reserved Parameters for
Buffer Directionality BIRD draft 3 (zip
<http://www.eda.org/ibis/macromodel_wip/archive/20150428/michaelmirmak/AMI_Reserved_Parameters_for_Buffer_Directionality_BIRD_draft_3.zip>
)(docx
<http://www.eda.org/ibis/macromodel_wip/archive/20150428/michaelmirmak/AMI%20Reserved%20Parameters%20for%20Buffer%20Directionality%20BIRD%20draft%203/ami-directionality-birdxxx-draft3.docx>)
On 4/28/2015 6:55 PM, Mirmak, Michael wrote:
As discussed in today’s IBIS-ATM Task Group meeting, the AMI
Directionality BIRD has been revised. Draft 3 is enclosed.
This draft removes the AMI_Model_Type Reserved Parameter as unnecessary.
It also makes explicit the need and rules for consistency checking between
[Model] Model_type, [Algorithmic Model] Direction, and Reserved Parameter
directions. Reserved Parameters approved for post-6.0 IBIS have been
added. The BIRD is now considerably shorter.
Comments and suggestions are welcome. Mike LaBonte, please post to the
archive when convenient. Thanks!
- MM
