From polim at ssl.snu.ac.kr Wed Feb 3 21:57:21 2010 From: polim at ssl.snu.ac.kr (=?ks_c_5601-1987?B?wNPHyr/B?=) Date: Thu, 4 Feb 2010 14:57:21 +0900 Subject: [Hotspot] [HotSpot] Where can I get the HS3D for 3D thermal estimation? Message-ID: <000b01caa55e$ead69240$c2792e93@polimnew> Hi, everyone. I need HS3D tool for 3D thermal simulation for my work. I tried to find it on the linke which was shown in the HotSpot Homepage. But it said there's no web-page for it. Could you let me know where can I download the HS3D tool for my work? or Is there anything which can I use for 3D thermal simulation? Thank you! -polim -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20100204/08143a94/attachment.html From skadron at cs.virginia.edu Thu Feb 4 06:03:03 2010 From: skadron at cs.virginia.edu (Kevin Skadron) Date: Thu, 04 Feb 2010 09:03:03 -0500 Subject: [Hotspot] [HotSpot] Where can I get the HS3D for 3D thermal estimation? In-Reply-To: <000b01caa55e$ead69240$c2792e93@polimnew> References: <000b01caa55e$ead69240$c2792e93@polimnew> Message-ID: <4B6AD397.6040805@cs.virginia.edu> It seems that the authors of HS3D at Penn St. have taken it down or done something that broke the link. I've emailed Greg Link, the primary author, to see what we can do. I should point out that we only link to that tool as a service, so other than trying to find out whether it's still available, there's not much we can do. You might want to contact Greg directly. He's now at York College of Pennsylvania. --Kevin ??? wrote: > > Hi, everyone. > I need HS3D tool for 3D thermal simulation for my work. > I tried to find it on the linke which was shown in the HotSpot Homepage. > But it said there's no web-page for it. > > Could you let me know where can I download the HS3D tool for my work? > or Is there anything which can I use for 3D thermal simulation? > > Thank you! > > -polim > > > > ------------------------------------------------------------------------ > > _______________________________________________ > HotSpot mailing list > HotSpot at mail.cs.virginia.edu > http://www.cs.virginia.edu/mailman/listinfo/hotspot From shervin at ucsd.edu Thu Feb 4 12:55:07 2010 From: shervin at ucsd.edu (Shervin Sharifi) Date: Thu, 4 Feb 2010 12:55:07 -0800 Subject: [Hotspot] Modeling mobile embedded systems thermal behavior Message-ID: Hi, I'm using HotSpot to simulate the thermal behavior of mobile embedded systems where using heat sinks and good packaging is not possible due to space and cost limitations. I try to mimic such heatsink and packaging by making the heat sink and spreader very thin and making r_convec large. I have two questions: 1. Is there a better way to model packaging and cooling of such systems in HotSpot? 2. Using very thin heatsink and heat spreader, I expect to see much faster changes in the temperature of the heatsink. But the time constant of the heatsink seems to be in the order of 10s of seconds which seems very high to me. Should I change other parameters as well to get correct results? Am I doing something wrong? Thanks Shervin -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20100204/a89654f1/attachment.html From shervin at ucsd.edu Fri Feb 5 12:18:34 2010 From: shervin at ucsd.edu (Shervin Sharifi) Date: Fri, 5 Feb 2010 12:18:34 -0800 Subject: [Hotspot] Estimating convection capacitance Message-ID: Hi, I need to estimate the convection capacitance of a heat sink (or the package in there is not heat sink). This is provided to the HotSpot through the configuration file. How should I estimate it? Shervin -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20100205/f4d6cad2/attachment.html From acoskun at cs.ucsd.edu Fri Feb 5 12:36:27 2010 From: acoskun at cs.ucsd.edu (Ayse Kivilcim Coskun) Date: Fri, 05 Feb 2010 15:36:27 -0500 Subject: [Hotspot] [HotSpot] Where can I get the HS3D for 3D thermal estimation? In-Reply-To: <4B6AD397.6040805@cs.virginia.edu> References: <000b01caa55e$ead69240$c2792e93@polimnew> <4B6AD397.6040805@cs.virginia.edu> Message-ID: <4B6C814B.2040402@cs.ucsd.edu> I guess Kevin or other HotSpot developers can comment on this in a more concrete way, but I believe Greg Link's HS3D is already included in HotSpot by a large extent (if not fully). When I contacted Greg with the same question a few years ago he advised me to use HotSpot, which is what I have been doing for 3D thermal simulations. -Ayse Kevin Skadron wrote: > It seems that the authors of HS3D at Penn St. have taken it down or done > something that broke the link. I've emailed Greg Link, the primary > author, to see what we can do. I should point out that we only link to > that tool as a service, so other than trying to find out whether it's > still available, there's not much we can do. You might want to contact > Greg directly. He's now at York College of Pennsylvania. > > --Kevin > > > -- Ayse K. Coskun Assistant Professor Electrical and Computer Engineering Department Boston University 8 Saint Mary's St. Boston, MA 02215 Tel: (617) 358-3641 E-mail: acoskun at bu.edu Web: http://people.bu.edu/acoskun From wh6p at virginia.edu Fri Feb 5 12:48:41 2010 From: wh6p at virginia.edu (Wei Huang) Date: Fri, 05 Feb 2010 15:48:41 -0500 Subject: [Hotspot] Estimating convection capacitance In-Reply-To: References: Message-ID: Hi, if you can find the specific heat, the density and the volume of the heatsink (or package), thermal capacitance is the product of the three. Regards, -Wei On Fri, 5 Feb 2010 12:18:34 -0800 Shervin Sharifi wrote: > Hi, > > I need to estimate the convection capacitance of a heat sink (or the > package in there is not heat sink). > This is provided to the HotSpot through the configuration file. How should > I estimate it? > > Shervin From wh6p at virginia.edu Fri Feb 5 12:54:37 2010 From: wh6p at virginia.edu (Wei Huang) Date: Fri, 05 Feb 2010 15:54:37 -0500 Subject: [Hotspot] Modeling mobile embedded systems thermal behavior In-Reply-To: References: Message-ID: Please see replies below (leading by ...) On Thu, 4 Feb 2010 12:55:07 -0800 Shervin Sharifi wrote: > Hi, > > I'm using HotSpot to simulate the thermal behavior of mobile embedded > systems where using heat sinks and good packaging is not possible due to > space and cost limitations. > I try to mimic such heatsink and packaging by making the heat sink and > spreader very thin and making r_convec large. > > I have two questions: > > 1. Is there a better way to model packaging and cooling of such systems in > HotSpot? ... if you mean natural convection, there is such a mode in HotSpot 5.0. But I would suggest use caution with your settings, because natural convection has much worse thermal characteristics and it is very easy to result in thermal runaway if there is something unrealistic in your configuration. > 2. Using very thin heatsink and heat spreader, I expect to see much faster > changes in the temperature of the heatsink. But the time constant of the > heatsink seems to be in the order of 10s of seconds which seems very high >to > me. As another email you sent suggested, you were considering reducing the convection capacitance -- I think that is the right approach. Glad you could figure this out yourself before we reply and sorry for the late reply :) -Wei > Should I change other parameters as well to get correct results? Am I > doing something wrong? > > Thanks > Shervin From vijay at cse.psu.edu Fri Feb 5 13:02:25 2010 From: vijay at cse.psu.edu (Vijaykrishnan Narayanan) Date: Fri, 05 Feb 2010 16:02:25 -0500 Subject: [Hotspot] [HotSpot] Where can I get the HS3D for 3D thermal estimation? In-Reply-To: <4B6C814B.2040402@cs.ucsd.edu> References: <000b01caa55e$ead69240$c2792e93@polimnew> <4B6AD397.6040805@cs.virginia.edu> <4B6C814B.2040402@cs.ucsd.edu> Message-ID: <4B6C8761.1010807@cse.psu.edu> Yes, this is correct. Hotspot is more up to date than HS3D and does incorporate capabilities of HS3D. We will recommend using latest versions of Hotspot Vijay Narayanan Penn State Univeristy > I guess Kevin or other HotSpot developers can comment on this in a more > concrete way, but I believe Greg Link's HS3D is already included in > HotSpot by a large extent (if not fully). When I contacted Greg with the > same question a few years ago he advised me to use HotSpot, which is > what I have been doing for 3D thermal simulations. > > -Ayse > > > Kevin Skadron wrote: > >> It seems that the authors of HS3D at Penn St. have taken it down or done >> something that broke the link. I've emailed Greg Link, the primary >> author, to see what we can do. I should point out that we only link to >> that tool as a service, so other than trying to find out whether it's >> still available, there's not much we can do. You might want to contact >> Greg directly. He's now at York College of Pennsylvania. >> >> --Kevin >> >> >> >> > > > From sitos.lin at gmail.com Sat Feb 6 03:42:52 2010 From: sitos.lin at gmail.com (Sitos Lin) Date: Sat, 6 Feb 2010 19:42:52 +0800 Subject: [Hotspot] Package in HotSpot 5.0 Message-ID: Deal all: I have several questions about the "package" in HotSpot 5.0. I have read the source code of package. To the best of my knowledge, the package only affects the value of r_convec (the thermal resistance). Is it right? So, setting the parameters in package would result in the same outcome if I set the r_convec accordingly. Besides that, because I am interesting in how the power consumption of fans affects the thermal resistance. Do any one know how to know the relationship between rpm of a fan and its power consumption? Thank a lots. Regards, Sitos From wh6p at virginia.edu Tue Feb 9 16:47:13 2010 From: wh6p at virginia.edu (Wei Huang) Date: Tue, 09 Feb 2010 16:47:13 -0500 Subject: [Hotspot] Question about scaling floorplans In-Reply-To: <4B71D04C.2070807@virginia.edu> References: <4B71D04C.2070807@virginia.edu> Message-ID: Hello Juan, Sorry for the delay. In the future, if you want quicker response, I would recommend sending an email to the mailing list (hotspot at cs.virginia.edu). In that way, you can reach more people and might get faster response. I am trying to answer one of the questions in your previous email to Prof. Stan, please see below. Other HotSpot users/developers might be able to answer your other questions. Regards, -Wei > I'm trying to integrate HotSpot into Opal (GEMS) and the first step i > took was to create a valid description file for the HotSpot > floorplanner to build the floorplan. After some time i've almost > decided how to make an equivalent description file, but i also need to > scale down the areas in the description files, from the provided ev6 > 130nm file to a 32nm technology, and also the heatsink thermal > resistance, and the modify the areas and units to match our processor > design... a quick starting point to estimate heatsink convection thermal resistance is to use the total power consumption of your chip of interest. For example, if your chip burns ~50W, and the average chip temperature you have in mind is around 60C, and the ambient temperature is 25C, that leads to Rth=(60-25)/80= 0.4375. So this would be a good initial guess, and you can further fine-tune it to meet your specific needs. > > Reading the mailing list archive i found that you said that "Hotspot > has beed validated for several floorplans and various technology > nodes.". Are any of those files available somewhere so i can compare > against my description files and check if we made the right decisions > creating the description files? > > Thank you in advance. > > Best, Juan Manuel. From fsamie at gmail.com Thu Feb 11 12:09:05 2010 From: fsamie at gmail.com (farzad samie) Date: Thu, 11 Feb 2010 20:39:05 +0330 Subject: [Hotspot] zero power Message-ID: Dear all, I produced a power trace file with zero values for all components and using default config values, run Hotspot. As I expected all the transient temperature values reported by hotspot was equal to ambient temperature (60 degree of Celsius). But steady state temperatures was strange. All the components have a 45 C steady state temperature. Is there some things wrong? if not, what is the reason? -- Farzad -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20100211/0e5071e6/attachment.html From wh6p at virginia.edu Thu Feb 11 14:50:36 2010 From: wh6p at virginia.edu (Wei Huang) Date: Thu, 11 Feb 2010 14:50:36 -0500 Subject: [Hotspot] zero power In-Reply-To: References: Message-ID: I think this might be caused by the initial temperature setting. Be default, the initial temperature is 60C, and the ambient is 45C. So if you haven't done a long transient simulation, all transient temperatures should be around 60C. But if you run long enough, they should converge towards 45C if there is absolutely no power dissipated. Please let me know what you find out. Regards, -Wei On Thu, 11 Feb 2010 20:39:05 +0330 farzad samie wrote: > Dear all, > I produced a power trace file with zero values for all components and >using > default config values, run Hotspot. As I expected all the transient > temperature values reported by hotspot was equal to ambient temperature >(60 > degree of Celsius). But steady state temperatures was strange. All the > components have a 45 C steady state temperature. Is there some things >wrong? > if not, what is the reason? > > -- >Farzad From polim at ssl.snu.ac.kr Wed Feb 17 12:05:58 2010 From: polim at ssl.snu.ac.kr (=?ks_c_5601-1987?B?wNPHyr/B?=) Date: Thu, 18 Feb 2010 02:05:58 +0900 Subject: [Hotspot] How can I generate the floorplans for the 3D layers easily? Message-ID: <000601caaff3$79cde1d0$c2792e93@polimnew> This is my first time to use the *lcf file for 3D simulation. First I made 2 floorplans for each layers. ( used hotfloorplan by 2 desc. files ) Second I made sample.lcf file with copying the example.lcf file but modified the floorplan file names with the 2 above floorplans. Third I run the hotspot with the lcf file with my ptrace file. It's not working. and the error message was "error: width and height differ across layers". I think I may have to adjust the layers' width and height equal on the lcf file's floorplan. Am I right? How can I generate the floorplans with same width and height easily? I think it's not easy to estimate the width and height of layers in advance from the first steps in *desc files. Is there any better way to do so? Please help me. (I tried to find same topic on the mailing lists, but there's no such threads.) -polim -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20100218/1dfa7d5f/attachment.html From anandmay22 at gmail.com Fri Feb 19 03:55:12 2010 From: anandmay22 at gmail.com (Anand J stephen) Date: Fri, 19 Feb 2010 14:25:12 +0530 Subject: [Hotspot] reference need for temperature measrement Message-ID: <46be924c1002190055h5228682di9d707025811d3db8@mail.gmail.com> Hi to all can any one suggest me a reference material to learn "how to measure temperature of a 2D flooplan"... For information, i am very new to this area. -- Best Regards Anand. J.Stephen Research Scholar National Center for Advance Research in Discrete Mathematics (Sponsored by Dept of Science and Technology,Govt of India) Kalasalingam University TamilNadu India www.ncardmath.com -------------- next part -------------- An HTML attachment was scrubbed... URL: http://www.cs.virginia.edu/pipermail/hotspot/attachments/20100219/d02f5010/attachment.html