From sbae at cse.psu.edu  Mon Sep 10 12:15:31 2007
From: sbae at cse.psu.edu (sbae)
Date: Mon, 10 Sep 2007 15:15:31 -0400
Subject: [Hotspot] Can I model a chip with bare silicon die without heat
	spreader on it?
Message-ID: <47245D66FAD5496CB4F2A3FE90C514A0@cse.psu.edu>

Hi

 

Can I model a chip with bare silicon die without heat spreader on it?

 

Thank you,

 

Sungmin

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20070910/8e1eb9eb/attachment.html 

From mrs8n at cms.mail.virginia.edu  Wed Sep 12 13:21:39 2007
From: mrs8n at cms.mail.virginia.edu (Mircea R. Stan)
Date: Wed, 12 Sep 2007 16:21:39 -0400
Subject: [Hotspot] Can I model a chip with bare silicon die without
 heat	spreader on it?
In-Reply-To: <47245D66FAD5496CB4F2A3FE90C514A0@cse.psu.edu>
References: <47245D66FAD5496CB4F2A3FE90C514A0@cse.psu.edu>
Message-ID: <web-15593179@neon.mail.virginia.edu>

On Mon, 10 Sep 2007 15:15:31 -0400
  "sbae" <sbae at cse.psu.edu> wrote:
> Hi
> 
> 
> 
> Can I model a chip with bare silicon die without heat spreader on it?
> 
Sungmin,

Right now the Hotspot models assume a spreader (which is
actually hardcoded in the model). It would be relatively easy
to remove the spreader from the code but there would be no guarantee
that the results would remain meaningful. We spent a lot of time validating
the existing models, and this effort would have to be replicated
for the case w/o the spreader in order to trust the compact models.
We are not planning anything along this line (as far as I know),
but if you can make the case that this is something that the user
community needs we could reconsider. Thanks for the inquiry!

Mircea
> 
> 
> Thank you,
> 
> 
> 
> Sungmin
> 

Mircea Stan

The views expressed in this email represent the personal
views of the sender and do not represent the official
position of the University of Virginia

From skadron at cs.virginia.edu  Fri Sep 21 18:05:57 2007
From: skadron at cs.virginia.edu (Kevin Skadron)
Date: Fri, 21 Sep 2007 18:05:57 -0400
Subject: [Hotspot] Announcing HotSpot version 4.0
Message-ID: <46F44045.9010603@mail.cs.virginia.edu>

We are pleased to announce the release of HotSpot version 4.0.  It 
includes some changes to enhance accuracy for high-aspect-ratio blocks 
in the block model, modification of the first-order transient solver of 
the grid model into a fourth-order adaptive step-size Runge-Kutta solver 
for increased accuracy, improved heatsink modeling, and a complete 
re-write of the grid model for better readability and extensibility. To 
read more about these changes, see our WDDD 2007 paper: 
http://www.cs.virginia.edu/~skadron/Papers/wddd07-UVA-hotspot-final.pdf

HotSpot 4.0 can be downloaded at:
http://lava.cs.virginia.edu/HotSpot/

In general, and especially for those who have already downloaded version 
4.0 since we posted it recently, we'd like to know how you're finding 
HotSpot, and whether you have encountered any questions, problems, or 
shortcomings.  We have an NSF CRI grant to continue supporting HotSpot 
and our users' needs, so it is important to us to find out what our 
users' needs are.

Some of our current tasks are to better support closing the 
thermal-leakage feedback loop and improved modeling of the package and 
cooling solution.  Longer term we will provide better integration with 
HotLeakage, better support for integration into full-system models, and 
faster solvers.

--Kevin