[ibis-macro] Re: Package power models and on-die Power models
- From: "Muranyi, Arpad" <Arpad_Muranyi@xxxxxxxxxx>
- To: IBIS-ATM <ibis-macro@xxxxxxxxxxxxx>
- Date: Wed, 27 Jun 2012 14:41:39 +0000
Walter,
We ran out of time in the ATM meeting yesterday so I didn't
have a chance to ask a question about your email below and
respond to your comment on my presentation, both of which
are very related.
Here are a couple of snippets from the minutes. After the
discussion on your email I asked:
- Arpad: Are these model variants supposed to be:
- separate models to be selected by a "model selector" type arrangement or
- a single model which supports "compression" or "expansion" capabilities
like swathing, to cover the various levels of simulation conditions?
- Walter: The way we do it you select a subcircuit to use for power
- The user makes the choice of which subcircuit to use for given conditions.
- Opportunity for IBIS is to utilize MCP.
After I was done with my presentation, you commented:
- Walter: My email focused only on power distribution on die.
- Signaling is an independent problem not addressed in my example email.
- I think your scheme might work for a single package model.
- In practice it won't work or will be very cumbersome for multiple models.
From this, I am still not sure I understand how MCP works.
If you say that the user (or EDA tool) selects a subcircuit
for the package model, then how is this subcircuit going to
accommodate the varying number of pins and pads you outlined
in your email for the same package model used at different
simulation frequencies? Would the subcircuit have the maximum
number of connection terminals and at lower frequencies some
would be left unconnected? How does this fit in with the fixed
number of pins listed under the [Pin] keyword?
Or would we have a different subcircuit for each simulation
condition, with a different number of connection terminals?
Regarding your comment on my presentation, if the subcircuit
you were talking about has a fixed number of connection terminals
for all simulation conditions, I don't see why this MCP could
work with the syntax proposed in my presentation.
I guess my problem is that I am still unclear on how MCP works.
Maybe this would be the time to ask Brad Brim to give us an
overview (presentation) in the next ATM meeting...
Thanks,
Arpad
====================================================================
From: ibis-macro-bounce@xxxxxxxxxxxxx [mailto:ibis-macro-bounce@xxxxxxxxxxxxx]
On Behalf Of Walter Katz
Sent: Tuesday, June 26, 2012 1:12 PM
To: Muranyi, Arpad; IBIS-ATM
Subject: [ibis-macro] Package power models and on-die Power models
All,
I hope the following example will explain why Power Distribution Models need to
be handled as a Model Connection Wrapper around the power subckts.
P are Power Silicon Bump Pads
P are Power Pins and Pads
B are Buffer Silicon Bump Pads, and IBIS Pins
IBIS file has 6 Power Pins, 8 Buffer Pins
Silicon has 24 Power Bump Pads and 8 Buffer Bump Pads
The Package Pins:
P
B B
P
B B
P
P
B B
P
B B
P
The Silicon Bump Pads
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
For the following Case
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
Package Power Distribution Circuit could have:
1 Power Pin Node, and 1 Power Pad Node
6 Power Pin Nodes, and 24 Power Pad Nodes
2 Power Pin Nodes, and 6 Power Pad Nodes
Or any combination of 1-6 Power Pin Nodes and 1-24 Power Pad Node
On die power distribution can have
1 Power Pad Node, and 1 Power Buffer Node
24 Power Pad Nodes, and 8 Power Buffer Nodes
Or any combination of 1-24 Power Pad Nodes and 1-8 Power Buffer Nodes
Power Distribution can be done for slices:
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
PPPP
B B
PPPP
B B
PPPP
Walter Katz
wkatz@xxxxxxxxxx<mailto:wkatz@xxxxxxxxxx>
Phone 303.449-2308
Mobile 303.335-6156
Other related posts:
