From khurshid at wisc.edu  Sat Feb  2 10:54:12 2013
From: khurshid at wisc.edu (Mushfique Khurshid)
Date: Sat, 2 Feb 2013 12:54:12 -0600
Subject: [Hotspot] Some question regarding temperature distribution and heat
	sink/spreader
Message-ID: <CAKLY5Eka8FckmmDWQoBV7Wj3mNegmSAG2tt2Jf96ubz+kc6r6Q@mail.gmail.com>

Hi,

I was trying to simulate a vertically stacked IC, with totally 18 layers
specified in the LCF file. So, while printing the steady state temperatures
of all layers, I slightly modified the temperature_grid.c to make the
function dump all steady state temperatures of all layers, in
model->last_steady->cuboid[layer no][i][j].

1) are layer numbers zero and one the heat sink and spreader, or are the
last two layers the heat sink and spreader?

2) I simulated my 18 layer (9 layers of silicon and 9 layers of TIM
inbetween) stack as I wrote earlier, and just to experiment, I simply made
all layers uniformly dissipate the same amount of power. I expected that in
each layer the edges would be the hottest since there are less paths for
the heat to propagate to, while the centre would be the coolest, but I was
getting the opposite. I saw concentric circles, with the centre being the
hottest, while the corners are the coolest. Any reason why that is the
case?

I would appreciate if someone could entertain my questions.

Mushfique
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130202/9be6ae52/attachment.html 

From mircea at virginia.edu  Sat Feb  2 11:17:34 2013
From: mircea at virginia.edu (Mircea R. Stan)
Date: Sat, 02 Feb 2013 14:17:34 -0500
Subject: [Hotspot] Some question regarding temperature distribution and
	heat	sink/spreader
In-Reply-To: <CAKLY5Eka8FckmmDWQoBV7Wj3mNegmSAG2tt2Jf96ubz+kc6r6Q@mail.gmail.com>
References: <CAKLY5Eka8FckmmDWQoBV7Wj3mNegmSAG2tt2Jf96ubz+kc6r6Q@mail.gmail.com>
Message-ID: <web-4336362@neon.mail.virginia.edu>

> 2) I simulated my 18 layer (9 layers of silicon and 9 layers of TIM
> inbetween) stack as I wrote earlier, and just to experiment, I simply made
> all layers uniformly dissipate the same amount of power. I expected that in
> each layer the edges would be the hottest since there are less paths for
> the heat to propagate to, while the centre would be the coolest, but I was
> getting the opposite. I saw concentric circles, with the centre being the
> hottest, while the corners are the coolest. Any reason why that is the
> case?

This phenomenon is illustrated in several of the Hotspot papers. Heat *always*
moves from high to low temperature and since heat can only move *from*
where power is dissipated (inside the die/stack) *to* the periphery
(eventually to the heatsink and ambient), the center of the die on average
will tend to be hotter than the periphery so that the heat can move out.
This is the reason 3D cooling is so tricky, the "center" of the stack tends
to be really hot even if no power is dissipated there. Hotspot should
model these effects quite accurately. Hope this helps!

Mircea
> 
> I would appreciate if someone could entertain my questions.
> 
> Mushfique

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 raidalzobaidi at yahoo.com  Sun Feb  3 03:18:04 2013
From: raidalzobaidi at yahoo.com (Raaed)
Date: Sun, 3 Feb 2013 03:18:04 -0800 (PST)
Subject: [Hotspot] Some question regarding temperature distribution and
	heat sink/spreader
In-Reply-To: <CAKLY5Eka8FckmmDWQoBV7Wj3mNegmSAG2tt2Jf96ubz+kc6r6Q@mail.gmail.com>
References: <CAKLY5Eka8FckmmDWQoBV7Wj3mNegmSAG2tt2Jf96ubz+kc6r6Q@mail.gmail.com>
Message-ID: <1359890284.88451.YahooMailNeo@web161301.mail.bf1.yahoo.com>

I think the results are make sense. The edges have vertical conduction to heat sink plus side?literal convection to air which make them cooler :-)

Best of luck,

Ra'ed
?


________________________________
 From: Mushfique Khurshid <khurshid at wisc.edu>
To: hotspot at mail.cs.virginia.edu 
Sent: Saturday, February 2, 2013 6:54 PM
Subject: [Hotspot] Some question regarding temperature distribution and heat sink/spreader
 



Hi,


I was trying to simulate a vertically stacked IC, with totally 18 layers specified in the LCF file. So, while printing the steady state temperatures of all layers, I slightly modified the temperature_grid.c to make the function dump all steady state temperatures of all layers, in model->last_steady->cuboid[layer no][i][j]. 

1) are layer numbers zero and one the heat sink and spreader, or are the last two layers the heat sink and spreader?

2) I simulated my 18 layer (9 layers of silicon and 9 layers of TIM inbetween) stack as I wrote earlier, and just to experiment, I simply made all layers uniformly dissipate the same amount of power. I expected that in each layer the edges would be the hottest since there are less paths for the heat to propagate to, while the centre would be the coolest, but I was getting the opposite. I saw concentric circles, with the centre being the hottest, while the corners are the coolest. Any reason why that is the case? 

I would appreciate if someone could entertain my questions.

Mushfique

_______________________________________________
HotSpot mailing list
HotSpot at mail.cs.virginia.edu
http://www.cs.virginia.edu/mailman/listinfo/hotspot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130203/2ef717fc/attachment.html 

From khurshid at wisc.edu  Tue Feb  5 15:03:43 2013
From: khurshid at wisc.edu (Mushfique Khurshid)
Date: Tue, 5 Feb 2013 17:03:43 -0600
Subject: [Hotspot] Simulating 3d IC with secondary heat transfer path
Message-ID: <CAKLY5EmYQd+j+mq6fsMxSOSeLx=F549QuVTVntZa0qEwc=nXqg@mail.gmail.com>

Hi,

It seems the latest version doesnot support simulation of a 3d IC with
secondary heat transfer path. The bottom-most layer of my 3d ic uses the
most power, and since the heat sink and spreader are at the top, my
bottom-most layer gets most heated up, and I get the error that excessive
temperature increase. I was wondering perhaps if secondary heat transfer
path was simulated, perhaps this could be avoided. I understand that
current hotspot doesnot support this, but could anyone give me some
suggestion to simulate this on a very high level? Did anyone simulate
secondary heat transfer path on a very highlevel by introducing additional
layers in lcf file? Suggestions would be appreciated highly.

-Mushfique
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130205/65a90656/attachment.html 

From vadrao at utu.fi  Tue Feb  5 23:19:29 2013
From: vadrao at utu.fi (Vaddina Kameswar Rao)
Date: Wed, 6 Feb 2013 09:19:29 +0200
Subject: [Hotspot] Simulating 3d IC with secondary heat transfer path
In-Reply-To: <03785726192b4815b1290bc5f22dc97b@EXCH-CAS-04.utu.fi>
References: <03785726192b4815b1290bc5f22dc97b@EXCH-CAS-04.utu.fi>
Message-ID: <CADua_hN4Dwp05YQi14pTzby_TBb1VV+tTgzjtz=qPTEnXJf69A@mail.gmail.com>

Hi Mushfique,

Which version of Hotspot are you using and did you toggle the corresponding
"secondary path" switch in "hotspot.config" file?

Cheers,
Kamesh.

PS: Sorry for multiple copies of the mail Mushfique. Forgot to add the
mailing list to the reply field.


On 6 February 2013 01:03, Mushfique Khurshid <khurshid at wisc.edu> wrote:

>
>
> Hi,
>
> It seems the latest version doesnot support simulation of a 3d IC with
> secondary heat transfer path. The bottom-most layer of my 3d ic uses the
> most power, and since the heat sink and spreader are at the top, my
> bottom-most layer gets most heated up, and I get the error that excessive
> temperature increase. I was wondering perhaps if secondary heat transfer
> path was simulated, perhaps this could be avoided. I understand that
> current hotspot doesnot support this, but could anyone give me some
> suggestion to simulate this on a very high level? Did anyone simulate
> secondary heat transfer path on a very highlevel by introducing additional
> layers in lcf file? Suggestions would be appreciated highly.
>
> -Mushfique
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130206/04484cf9/attachment.html 

From khurshid at wisc.edu  Wed Feb  6 09:01:05 2013
From: khurshid at wisc.edu (Mushfique Khurshid)
Date: Wed, 6 Feb 2013 11:01:05 -0600
Subject: [Hotspot] Simulating 3d IC with secondary heat transfer path
In-Reply-To: <CADua_hN4Dwp05YQi14pTzby_TBb1VV+tTgzjtz=qPTEnXJf69A@mail.gmail.com>
References: <03785726192b4815b1290bc5f22dc97b@EXCH-CAS-04.utu.fi>
	<CADua_hN4Dwp05YQi14pTzby_TBb1VV+tTgzjtz=qPTEnXJf69A@mail.gmail.com>
Message-ID: <CAKLY5E=LY=M4Y9DM6XV7zY5-TDEGJmtrpY90AVMzGoKO1XSQwg@mail.gmail.com>

Hi,

Thanks for the reply. I am trying to simulate "secondary path" for 3d IC. I
dont get that error anymore, but HotSpot tells me that if I specify a
"layer configuration file", this version does not support "secondary heat
transfer path". I am using HotSpot 5.02

On Wed, Feb 6, 2013 at 1:19 AM, Vaddina Kameswar Rao <vadrao at utu.fi> wrote:

> Hi Mushfique,
>
> Which version of Hotspot are you using and did you toggle the
> corresponding "secondary path" switch in "hotspot.config" file?
>
> Cheers,
> Kamesh.
>
> PS: Sorry for multiple copies of the mail Mushfique. Forgot to add the
> mailing list to the reply field.
>
>
> On 6 February 2013 01:03, Mushfique Khurshid <khurshid at wisc.edu> wrote:
>
>>
>>
>> Hi,
>>
>> It seems the latest version doesnot support simulation of a 3d IC with
>> secondary heat transfer path. The bottom-most layer of my 3d ic uses the
>> most power, and since the heat sink and spreader are at the top, my
>> bottom-most layer gets most heated up, and I get the error that excessive
>> temperature increase. I was wondering perhaps if secondary heat transfer
>> path was simulated, perhaps this could be avoided. I understand that
>> current hotspot doesnot support this, but could anyone give me some
>> suggestion to simulate this on a very high level? Did anyone simulate
>> secondary heat transfer path on a very highlevel by introducing additional
>> layers in lcf file? Suggestions would be appreciated highly.
>>
>> -Mushfique
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130206/eb23ab4c/attachment.html 

From wh6p at virginia.edu  Thu Feb  7 07:22:06 2013
From: wh6p at virginia.edu (Wei Huang)
Date: Thu, 7 Feb 2013 09:22:06 -0600
Subject: [Hotspot] Simulating 3d IC with secondary heat transfer path
In-Reply-To: <CAKLY5E=LY=M4Y9DM6XV7zY5-TDEGJmtrpY90AVMzGoKO1XSQwg@mail.gmail.com>
References: <03785726192b4815b1290bc5f22dc97b@EXCH-CAS-04.utu.fi>
	<CADua_hN4Dwp05YQi14pTzby_TBb1VV+tTgzjtz=qPTEnXJf69A@mail.gmail.com>
	<CAKLY5E=LY=M4Y9DM6XV7zY5-TDEGJmtrpY90AVMzGoKO1XSQwg@mail.gmail.com>
Message-ID: <CANYeXBqAMbVw_fGggCQb41DkVZ3yb069t4nMmSyQWgz9vANqkQ@mail.gmail.com>

Hi,

As of version 5.02, HotSpot doesn't support secondary heat transfer path
with multiple silicon layers.

The original intention of including secondary path for package with only
one silicon layer was to evaluate scenarios where heatsink and heatspreader
are extremely weak (oil flow over bare die, for example).

Another reason is that for each additional silicon layer, secondary heat
transfer path needs to also include a layer for on-chip metals and a layer
for C4s/TSVs. Depending on which wafer bonding method is chosen (F2F, F2B),
the dimension and materials and the order in which each layer appears have
a lot uncertainties.

At that time, we didn't have the bandwidth to develop a full-blown support
for both 3D and secondary path simulations. But the infrastructure is
already there -- with some additional hack to the source code
(temperature_grid.c and .h), one can always rip off the default on-chip
metal and C4 layers, and include those layers for each silicon layer in the
LCF file.

Hope this helps.

-Wei


On Wed, Feb 6, 2013 at 11:01 AM, Mushfique Khurshid <khurshid at wisc.edu>wrote:

> Hi,
>
> Thanks for the reply. I am trying to simulate "secondary path" for 3d IC.
> I dont get that error anymore, but HotSpot tells me that if I specify a
> "layer configuration file", this version does not support "secondary heat
> transfer path". I am using HotSpot 5.02
>
>
> On Wed, Feb 6, 2013 at 1:19 AM, Vaddina Kameswar Rao <vadrao at utu.fi>wrote:
>
>> Hi Mushfique,
>>
>> Which version of Hotspot are you using and did you toggle the
>> corresponding "secondary path" switch in "hotspot.config" file?
>>
>> Cheers,
>> Kamesh.
>>
>> PS: Sorry for multiple copies of the mail Mushfique. Forgot to add the
>> mailing list to the reply field.
>>
>>
>>  On 6 February 2013 01:03, Mushfique Khurshid <khurshid at wisc.edu> wrote:
>>
>>>
>>>
>>> Hi,
>>>
>>> It seems the latest version doesnot support simulation of a 3d IC with
>>> secondary heat transfer path. The bottom-most layer of my 3d ic uses the
>>> most power, and since the heat sink and spreader are at the top, my
>>> bottom-most layer gets most heated up, and I get the error that excessive
>>> temperature increase. I was wondering perhaps if secondary heat transfer
>>> path was simulated, perhaps this could be avoided. I understand that
>>> current hotspot doesnot support this, but could anyone give me some
>>> suggestion to simulate this on a very high level? Did anyone simulate
>>> secondary heat transfer path on a very highlevel by introducing additional
>>> layers in lcf file? Suggestions would be appreciated highly.
>>>
>>> -Mushfique
>>>
>>
>>
>
> _______________________________________________
> HotSpot mailing list
> HotSpot at mail.cs.virginia.edu
> http://www.cs.virginia.edu/mailman/listinfo/hotspot
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130207/e4b2a78d/attachment.html 

From franck79 at gmail.com  Fri Feb  8 11:07:03 2013
From: franck79 at gmail.com (Francesco)
Date: Fri, 8 Feb 2013 11:07:03 -0800
Subject: [Hotspot] Simulating secondary heat transfer without heat spreader
	and sink
Message-ID: <CAGiRPj-EiSVyhcji9B1W9mHC98Uk81Zp_pLQo4SNMNWUZGT2_Q@mail.gmail.com>

Hello,

I am using HotSpot 5.0, in particular I am interested in simulating for a
3D stack die what it is mentioned at 2) secondary heat transfer path from
silicon to C4 pads to packaging substrate to solder ball and
printed-circuit board.

I would like to ask you the following questions

1) Is it possible to remove the heat spreader and sink and leave only the
connection to the ambient specifying its heat transfer and thermal
conductivity?
2) how I can set / unset the secondary heat transfer? In particular can I
also set the heat transfer and the thermal conductivity for the connection
to the ambient at the bottom of my system (i.e., below the PCB) ?
3) Can different layers have different width and height? I would like to
simulate different layer materials apart the die


Thanks,
Francesco
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130208/a96cec97/attachment.html 

From majid.06 at gmail.com  Tue Feb 12 01:34:06 2013
From: majid.06 at gmail.com (majid jalili)
Date: Tue, 12 Feb 2013 01:34:06 -0800
Subject: [Hotspot] 3D source level integration
Message-ID: <CALSVAJJ31LFMOw+pyHeDWtOtSH+rfk-VpAUoVrp6YtA1O83P6Q@mail.gmail.com>

Hi

Can anybody help me about source level integration of Hostspot with cycle
accurate tool (My design is 3D (processor+multiple memory layer above))
.sim-template.c is for 2D and I can not find explanation for 3d

tanx
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130212/3d0be30b/attachment.html 

From mcpatspot at gmail.com  Mon Feb 25 02:02:55 2013
From: mcpatspot at gmail.com (Hotspot McPat)
Date: Mon, 25 Feb 2013 10:02:55 +0000
Subject: [Hotspot] how to create gcc.ptrace
In-Reply-To: <CABSqQy55F3VxLL=z9PgRX=PQof-bUTi9jYU73+TZYT3_mzs5dw@mail.gmail.com>
References: <CABSqQy55F3VxLL=z9PgRX=PQof-bUTi9jYU73+TZYT3_mzs5dw@mail.gmail.com>
Message-ID: <CABSqQy7JGQn-UosUxQRZk4ERpLFS0ppPpyQNmKKg+MNCa8WRiQ@mail.gmail.com>

Hi All,

I'm wondering how I can make the gcc.ptrace file if I had the .flp file
from another platform (e.g. ARM, X86, AMD, etc). Does HotSpot have a
command or ability to make this file? Thank you in advance and sorry if
this is a very basic question!

Cheers
Negar


On Sun, Feb 24, 2013 at 11:17 PM, Hotspot McPat <mcpatspot at gmail.com> wrote:

> Hi All,
>
> I'm wondering how I can make the gcc.ptrace file if I had the .flp file
> from another platform (e.g. ARM, X86, AMD, etc). Does HotSpot have a
> command or ability to make this file? Thank you in advance and sorry if
> this is a very basic question!
>
> Cheers
> Negar
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130225/f731190f/attachment.html 

From skadron at cs.virginia.edu  Mon Feb 25 05:13:27 2013
From: skadron at cs.virginia.edu (Kevin Skadron)
Date: Mon, 25 Feb 2013 08:13:27 -0500
Subject: [Hotspot] how to create gcc.ptrace
In-Reply-To: <CABSqQy7JGQn-UosUxQRZk4ERpLFS0ppPpyQNmKKg+MNCa8WRiQ@mail.gmail.com>
References: <CABSqQy55F3VxLL=z9PgRX=PQof-bUTi9jYU73+TZYT3_mzs5dw@mail.gmail.com>
	<CABSqQy7JGQn-UosUxQRZk4ERpLFS0ppPpyQNmKKg+MNCa8WRiQ@mail.gmail.com>
Message-ID: <512B6377.8010408@cs.virginia.edu>

This would be produced by some kind of power model, such as Wattch or 
McPAT, or perhaps a CAD tool.
/K

On 2/25/2013 5:02 AM, Hotspot McPat wrote:
> Hi All,
>
> I'm wondering how I can make the gcc.ptrace file if I had the .flp file
> from another platform (e.g. ARM, X86, AMD, etc). Does HotSpot have a
> command or ability to make this file? Thank you in advance and sorry if
> this is a very basic question!
>
> Cheers
> Negar
>
>
> On Sun, Feb 24, 2013 at 11:17 PM, Hotspot McPat <mcpatspot at gmail.com
> <mailto:mcpatspot at gmail.com>> wrote:
>
>     Hi All,
>
>     I'm wondering how I can make the gcc.ptrace file if I had the .flp
>     file from another platform (e.g. ARM, X86, AMD, etc). Does HotSpot
>     have a command or ability to make this file? Thank you in advance
>     and sorry if this is a very basic question!
>
>     Cheers
>     Negar
>
>
>
>
> _______________________________________________
> HotSpot mailing list
> HotSpot at mail.cs.virginia.edu
> http://www.cs.virginia.edu/mailman/listinfo/hotspot
>

From johann.knechtel at ifte.de  Wed Feb 27 01:59:06 2013
From: johann.knechtel at ifte.de (Johann Knechtel)
Date: Wed, 27 Feb 2013 10:59:06 +0100
Subject: [Hotspot] Modeling 3D ICs
Message-ID: <512DD8EA.9060209@ifte.de>

Hi all,

I am using the HotSpot extension from BU for modeling 3D ICs and got
some questions.

1) I do 3D floorplanning with non-packed, fixed-outline configuration.
Thus, I experience layouts where not every die's outline is defined by
the block arrangement, and I get an error from HotSpot indicating the
layer mismatch. In order to resolve this, I add a dummy block in each
die covering the desired outline w/ zero value for specific heat and
thermal resistivity. This should be reasonable and not affecting the
simulation, right?

2) Is the thermal-leakage loop also working in the 3D BU extension? For
some initial experiments I couldn't see any difference in the results,
but I am not sure whether this is an issue with my settings or an
general issue.

3) During floorplanning-optimization loops, I want to use power blurring
for fast thermal estimation and then subsequently evaluate the results
using HotSpot thermal simulation. For power blurring, I need a thermal
mask (impulse response of layers obtained by applying point heat
sources), which I also generated using HotSpot until now.
However, I am not sure whether the accuracy of the grid-based simulation
is sufficient; for a point source w/ large power values, the thermal
distribution seems very limited to some grid bins, even in cases where
the power source is in an adjacent layer and thus thermal spreading
should occur.
Can somebody comment on that? Is HotSpot in general suitable for
modelling such point sources?

Thanks,
Johann
-- 
http://www.ifte.de/english/staff/knechtel.html
TU Dresden, IFTE, BAR II/47, Phone: +49 351 463 39612

From acoskun at bu.edu  Wed Feb 27 12:38:14 2013
From: acoskun at bu.edu (Ayse Kivilcim Coskun)
Date: Wed, 27 Feb 2013 15:38:14 -0500
Subject: [Hotspot] HotSpot Digest, Vol 77, Issue 2
In-Reply-To: <mailman.23.1361995224.15955.hotspot@mail.cs.virginia.edu>
References: <mailman.23.1361995224.15955.hotspot@mail.cs.virginia.edu>
Message-ID: <512E6EB6.2040607@bu.edu>

Hi Johann,

A few quick responses are below.

Ayse

On 2/27/2013 3:00 PM, hotspot-request at cs.virginia.edu wrote:
>
>
> Hi all,
>
> I am using the HotSpot extension from BU for modeling 3D ICs and got
> some questions.
>
> 1) I do 3D floorplanning with non-packed, fixed-outline configuration.
> Thus, I experience layouts where not every die's outline is defined by
> the block arrangement, and I get an error from HotSpot indicating the
> layer mismatch. In order to resolve this, I add a dummy block in each
> die covering the desired outline w/ zero value for specific heat and
> thermal resistivity. This should be reasonable and not affecting the
> simulation, right?

I am not sure if I understand what your floorplan looks like. If you 
have some whitespace in each layer, then you can put dummy blocks with 
zero power values.

If you have layers with different sizes (which is probably why you are 
getting the layer mismatch error), then you can put dummy blocks. 
Instead of putting zero specific heat & resistivity, I would probably 
put values so as to represent the heat flow to air or to other packaging 
material involved in the design.



>
> 2) Is the thermal-leakage loop also working in the 3D BU extension? For
> some initial experiments I couldn't see any difference in the results,
> but I am not sure whether this is an issue with my settings or an
> general issue.

All the default HotSpot features should be working in general; however, 
we have not tested the currently posted version with the leakage loop. 
You can always add your temperature-dependent leakage model into the 
simulator.

I plan to release an update to our extension within a few weeks (this 
has been in the works for a while now, apologies to those waiting for 
the update). We can comment on how the extension works with the leakage 
model in the updated version.

>
> 3) During floorplanning-optimization loops, I want to use power blurring
> for fast thermal estimation and then subsequently evaluate the results
> using HotSpot thermal simulation. For power blurring, I need a thermal
> mask (impulse response of layers obtained by applying point heat
> sources), which I also generated using HotSpot until now.
> However, I am not sure whether the accuracy of the grid-based simulation
> is sufficient; for a point source w/ large power values, the thermal
> distribution seems very limited to some grid bins, even in cases where
> the power source is in an adjacent layer and thus thermal spreading
> should occur.
> Can somebody comment on that? Is HotSpot in general suitable for
> modelling such point sources?
>


This is more of a general HotSpot question, so others can comment in 
more detail. Quick comment: you can reduce the grid size in HotSpot. I 
don't see a fundamental problem for modeling point sources as long as 
you adjust grid sizes, power numbers, sampling interval, and the like 
accordingly.

Ayse


From wh6p at virginia.edu  Thu Feb 28 11:01:10 2013
From: wh6p at virginia.edu (Wei Huang)
Date: Thu, 28 Feb 2013 13:01:10 -0600
Subject: [Hotspot] Modeling 3D ICs
In-Reply-To: <512DD8EA.9060209@ifte.de>
References: <512DD8EA.9060209@ifte.de>
Message-ID: <CANYeXBrDPHs5csiio4+qFQymJn+HrrOaf5S1gQjxq0wa2xzezw@mail.gmail.com>

Hi Johann,

Please see my comments...
1) your approach seems correct. As long as all layers have the same area,
adding dummy, zero-power blocks is the correct way to do this.
2) I am not sure about the BU package. For HotSpot itself, you may want to
check whether the leakage model is enabled in the config file you are using.
3) HotSpot can simulate "point" heat source only at the granularity of an
individual grid cell. Although I doubt any tool can simulate true point
source, you can always use smaller grid size, at the expense of longer
simulation time. Also, since HotSpot grid model is more accurate when
simulating cell sizes greater than the die thickness, for tiny heat
sources, you may consider slicing the die in z direction into multiple
thinner layers to achieve better accuracy. Of course, this would also add
computation complexity. You can also try other FEM thermal simulators such
as ANSYS or FloTherm.

Hope this helps.
-Wei


On Wed, Feb 27, 2013 at 3:59 AM, Johann Knechtel <johann.knechtel at ifte.de>wrote:

> Hi all,
>
> I am using the HotSpot extension from BU for modeling 3D ICs and got
> some questions.
>
> 1) I do 3D floorplanning with non-packed, fixed-outline configuration.
> Thus, I experience layouts where not every die's outline is defined by
> the block arrangement, and I get an error from HotSpot indicating the
> layer mismatch. In order to resolve this, I add a dummy block in each
> die covering the desired outline w/ zero value for specific heat and
> thermal resistivity. This should be reasonable and not affecting the
> simulation, right?
>
> 2) Is the thermal-leakage loop also working in the 3D BU extension? For
> some initial experiments I couldn't see any difference in the results,
> but I am not sure whether this is an issue with my settings or an
> general issue.
>
> 3) During floorplanning-optimization loops, I want to use power blurring
> for fast thermal estimation and then subsequently evaluate the results
> using HotSpot thermal simulation. For power blurring, I need a thermal
> mask (impulse response of layers obtained by applying point heat
> sources), which I also generated using HotSpot until now.
> However, I am not sure whether the accuracy of the grid-based simulation
> is sufficient; for a point source w/ large power values, the thermal
> distribution seems very limited to some grid bins, even in cases where
> the power source is in an adjacent layer and thus thermal spreading
> should occur.
> Can somebody comment on that? Is HotSpot in general suitable for
> modelling such point sources?
>
> Thanks,
> Johann
> --
> http://www.ifte.de/english/staff/knechtel.html
> TU Dresden, IFTE, BAR II/47, Phone: +49 351 463 39612
> _______________________________________________
> HotSpot mailing list
> HotSpot at mail.cs.virginia.edu
> http://www.cs.virginia.edu/mailman/listinfo/hotspot
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130228/43aa91fc/attachment.html 

From smilingsantanu at gmail.com  Thu Feb 28 17:45:37 2013
From: smilingsantanu at gmail.com (Santanu Sarma)
Date: Thu, 28 Feb 2013 17:45:37 -0800
Subject: [Hotspot] Hotspot Color Map
Message-ID: <CANq=v5eDs9rktxemKBumHYEzV7hSE=51FxADNqiYij9ZAwEMbA@mail.gmail.com>

Hi All,
  I have few queries for the thermal color map generated by the block based
model. I assume that the steady state temperature for 30 blocks (in EV6
floor plan, 5.02 version), should be mapped to 30 color points. However,
there are color variation even with in the blocks. How are the steady state
temperatures mapped to the blocks in the thermal color map? Are there any
additional temperature points computed when generating the thermal color
map?

  Also, can anyone share or point me to the a power trace file other than
the default gcc benchmark for EV6 floor plan as included in the 5.02
versions.

Thanks
Santanu


On Thu, Feb 28, 2013 at 11:01 AM, Wei Huang <wh6p at virginia.edu> wrote:

> Hi Johann,
>
> Please see my comments...
> 1) your approach seems correct. As long as all layers have the same area,
> adding dummy, zero-power blocks is the correct way to do this.
> 2) I am not sure about the BU package. For HotSpot itself, you may want to
> check whether the leakage model is enabled in the config file you are using.
> 3) HotSpot can simulate "point" heat source only at the granularity of an
> individual grid cell. Although I doubt any tool can simulate true point
> source, you can always use smaller grid size, at the expense of longer
> simulation time. Also, since HotSpot grid model is more accurate when
> simulating cell sizes greater than the die thickness, for tiny heat
> sources, you may consider slicing the die in z direction into multiple
> thinner layers to achieve better accuracy. Of course, this would also add
> computation complexity. You can also try other FEM thermal simulators such
> as ANSYS or FloTherm.
>
> Hope this helps.
> -Wei
>
>
> On Wed, Feb 27, 2013 at 3:59 AM, Johann Knechtel <johann.knechtel at ifte.de>wrote:
>
>> Hi all,
>>
>> I am using the HotSpot extension from BU for modeling 3D ICs and got
>> some questions.
>>
>> 1) I do 3D floorplanning with non-packed, fixed-outline configuration.
>> Thus, I experience layouts where not every die's outline is defined by
>> the block arrangement, and I get an error from HotSpot indicating the
>> layer mismatch. In order to resolve this, I add a dummy block in each
>> die covering the desired outline w/ zero value for specific heat and
>> thermal resistivity. This should be reasonable and not affecting the
>> simulation, right?
>>
>> 2) Is the thermal-leakage loop also working in the 3D BU extension? For
>> some initial experiments I couldn't see any difference in the results,
>> but I am not sure whether this is an issue with my settings or an
>> general issue.
>>
>> 3) During floorplanning-optimization loops, I want to use power blurring
>> for fast thermal estimation and then subsequently evaluate the results
>> using HotSpot thermal simulation. For power blurring, I need a thermal
>> mask (impulse response of layers obtained by applying point heat
>> sources), which I also generated using HotSpot until now.
>> However, I am not sure whether the accuracy of the grid-based simulation
>> is sufficient; for a point source w/ large power values, the thermal
>> distribution seems very limited to some grid bins, even in cases where
>> the power source is in an adjacent layer and thus thermal spreading
>> should occur.
>> Can somebody comment on that? Is HotSpot in general suitable for
>> modelling such point sources?
>>
>> Thanks,
>> Johann
>> --
>> http://www.ifte.de/english/staff/knechtel.html
>> TU Dresden, IFTE, BAR II/47, Phone: +49 351 463 39612
>> _______________________________________________
>> HotSpot mailing list
>> HotSpot at mail.cs.virginia.edu
>> http://www.cs.virginia.edu/mailman/listinfo/hotspot
>>
>
>
> _______________________________________________
> HotSpot mailing list
> HotSpot at mail.cs.virginia.edu
> http://www.cs.virginia.edu/mailman/listinfo/hotspot
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20130228/7626bfb2/attachment.html