[ibis-macro] Re: issue raised by Walter in last ATM
- From: <radek_biernacki@xxxxxxxxxxx>
- To: <wkatz@xxxxxxxxxx>, <fangyi_rao@xxxxxxxxxxx>, <Arpad_Muranyi@xxxxxxxxxx>, <ibis-macro@xxxxxxxxxxxxx>
- Date: Mon, 15 Jul 2013 16:30:40 -0600
Walter/Arpad,
The currently proposed arguments to AMI_Resolve() and AMI_init() differ more
than what you say. The "corner" and "model_name" are not arguments to AMI_init,
and that's the reason for this discussion. Perhaps you anticipated removal of
these two arguments from AMI_Resolve() since the new reserved parameters would
be of usage In and would be passed to AMI_Resolve() anyway - I think that's the
right thing to do.
I agree, the name "Corner" may be too generic, and perhaps "Model_corner" would
be a better and a more self-explanatory name. Nevertheless, in the context
proposed by Fangyi, it should be a reserved parameter to establish precise
rules for setting its value. For model specific parameters of usage In or
InOut, a free user selection is expected. If the "Model_corner" parameter were
a model specific parameter its selection would not be that free, independent of
the actual corner of the IBIS model. Therefore, the most appropriate handling
of this piece of information is for the EDA tool to follow the user selection
of the IBIS model corner. For that, the Reserved_Parameter branch is the right
place.
Radek
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Walter Katz
Sent: Monday, July 15, 2013 2:52 PM
To: RAO,FANGYI (A-USA,ex1); Arpad_Muranyi@xxxxxxxxxx; ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: issue raised by Walter in last ATM
Fangyi, Arpad,
I do agree with the new optional reserved parameter " Model_Name". I have two
problems with the reserved parameter " Corner". First, we have a leaf named
Corner. Although a parameter name could be Usage, Description, Type, ... I
think we should refrain from making reserved parameters these names.
Second, I think Arpad's solution is quite acceptable. Other alternatives that a
model maker might make are:
(my_corner (Usage In) (Corner 0 -1 1) (Type Integer) (Description "Simulation
corner."))
I am not sure I understand Fangyi's comment " That's because EDA tools must
make sure that the corner value in the input parameter string to AMI_Init is
consistent with the legacy IBIS model that is being used.".
Please now note that the signature (C calling argument list) of AMI_Resolve and
AMI_Init are now identical, except that AMI_Init all includes a pointer to an
Impulse Response Matrix. What Adge suggested originally was to treat the
AMI_Init call as the AMI_Resolve call when the impulse response matric pointer
(*impulse_matrix) was net to Null.
Walter
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of fangyi_rao@xxxxxxxxxxx
Sent: Monday, July 15, 2013 5:36 PM
To: Arpad_Muranyi@xxxxxxxxxx; ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: issue raised by Walter in last ATM
Arpad;
That's because EDA tools must make sure that the corner value in the input
parameter string to AMI_Init is consistent with the legacy IBIS model that is
being used.
Fangyi
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of Muranyi, Arpad
Sent: Monday, July 15, 2013 2:33 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] Re: issue raised by Walter in last ATM
Fangyi,
I am not sure why a new reserved parameter for corner is needed. Wouldn't the
following work in the Model_Specific
section:
(corner (Usage In) (Corner "Typ" "Min" "Max") (Type String)
(Description "Simulation corner."))
Thanks,
Arpad
=========================================================
-----Original Message-----
From: ibis-macro-bounce@xxxxxxxxxxxxx
[mailto:ibis-macro-bounce@xxxxxxxxxxxxx] On Behalf Of fangyi_rao@xxxxxxxxxxx
Sent: Monday, July 15, 2013 4:25 PM
To: ibis-macro@xxxxxxxxxxxxx
Subject: [ibis-macro] issue raised by Walter in last ATM
Hi, All;
Walter raised an issue during the AMI_Resolve API discussion at last ATM.
Since Dep parameters are not included in the input parameter string to
AMI_Init, if AMI_Init needs to use a Dep parameter value, it'll have to repeat
the calculation in AMI_Resolve. If the parameter depends on corner and/or model
name, AMI_Init cannot access them in current spec. In fact, this issue exists
regardless AMI_Resolve is used or not.
There are two potential solutions.
Solution 1: Require all Dep parameters to be included in the input parameter
string to AMI_Init.
Pro: minimum change to the spec
AMI_Init doesn't need to repeat the calculation in AMI_Resolve
Con: lots of Dep parameters are not designed to be used by AMI_Init
AMI_Init still has no access to corner and model name
Solution 2: Introduce two new AMI reserved parameters, Corner and Model_Name
Required: No
Usage: In
Type: String
Format: Value
Allowed values of Corner: typ, min, max
Allowed values of Model_Name: any string
Their values specified in the .ami file are irrelevant. EDA tools must make
sure that their values in the input parameter string to AMI_Init are consistent
with the legacy IBIS model that is being used.
Pro: AMI_Init has access to corner and model name regardless whether
AMI_Resolve exists.
Con: Two new reserved parameters
AMI_Init repeats the calculation in AMI_Resolve
Let's discuss at tomorrow's ATM.
Thanks,
Fangyi
---------------------------------------------------------------------
IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe
---------------------------------------------------------------------
IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe
---------------------------------------------------------------------
IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe
---------------------------------------------------------------------
IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe
---------------------------------------------------------------------
IBIS Macro website : http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector: //www.freelists.org/list/ibis-macro
To unsubscribe send an email:
To: ibis-macro-request@xxxxxxxxxxxxx
Subject: unsubscribe
Other related posts:
