By: Iain McClatchie (iain.delete@this.mcclatchie.com), January 7, 2013 12:28 am
Room: Moderated Discussions
David,
It turns out the Vt goes down at higher temperatures (Wikipedia has some coverage). That means the processor can run a little faster as it gets hotter. It also means that subthreshold leakage increases at higher temperatures, which can lead to thermal runaway.
This effect used to be dominated by the higher wire resistance at higher temperature that you cover in your article. Supply voltages have been coming down, and threshold voltages have not been changing much (at least for the leak-stop transistors). As a result, the temperature effect on voltage threshold has been steadily getting more important and it now dominates wire resistance in the latest commodity processes. The high-voltage, HIGH-temperature corner is now the fastest corner!
Just a heads up.
-Iain
It turns out the Vt goes down at higher temperatures (Wikipedia has some coverage). That means the processor can run a little faster as it gets hotter. It also means that subthreshold leakage increases at higher temperatures, which can lead to thermal runaway.
This effect used to be dominated by the higher wire resistance at higher temperature that you cover in your article. Supply voltages have been coming down, and threshold voltages have not been changing much (at least for the leak-stop transistors). As a result, the temperature effect on voltage threshold has been steadily getting more important and it now dominates wire resistance in the latest commodity processes. The high-voltage, HIGH-temperature corner is now the fastest corner!
Just a heads up.
-Iain
Topic | Posted By | Date |
---|---|---|
Article: Cooling and performance/watt | David Kanter | 2011/06/21 12:19 PM |
'temperature' not 'power'? | Paul A. Clayton | 2011/06/21 03:01 PM |
'temperature' not 'power'? | David Kanter | 2011/06/21 03:38 PM |
resistance(temperature) | Moritz | 2011/06/22 04:48 AM |
resistance(temperature) | Adrian | 2011/06/22 05:13 AM |
resistance(temperature) | David Hess | 2011/06/22 08:53 AM |
resistance(temperature) | Adrian | 2011/06/24 02:24 AM |
resistance(temperature) | David Hess | 2011/06/24 02:14 PM |
Article: Cooling and performance/watt | Ed Trice | 2011/06/22 10:57 AM |
Cooling | David Kanter | 2011/06/22 03:26 PM |
Cooling | Ed Trice | 2011/06/22 03:54 PM |
TE-elements | Moritz | 2011/06/23 05:55 AM |
Radiator placement and design | Ricardo B | 2011/06/23 07:34 AM |
TE-elements | EduardoS | 2011/06/23 04:21 PM |
water/air | Moritz | 2011/06/23 10:30 PM |
water/air | Ricardo B | 2011/06/24 02:29 PM |
water/air | bakaneko | 2011/06/24 09:45 PM |
water/air | David Hess | 2011/06/25 04:12 AM |
water/air | Ricardo B | 2011/06/25 06:07 AM |
water/air | ZaZa | 2011/06/25 09:47 AM |
water/air | Ricardo B | 2011/06/25 11:40 AM |
water/air | rwessel | 2011/06/26 03:43 AM |
water/air | ZaZa | 2011/06/26 04:05 PM |
Temperature inversion | Jonathan Kang | 2011/06/22 05:43 PM |
LN2 overclocking | Doug Siebert | 2011/06/25 01:32 PM |
Temperature inversion | Vincent Diepeveen | 2011/06/27 01:01 PM |
Temperature inversion | Anon | 2011/06/28 03:30 PM |
Temperature inversion | Jonathan Kang | 2011/07/05 06:38 PM |
Article: Cooling and performance/watt | (tm) | 2011/06/27 05:51 AM |
Article: Cooling and performance/watt | David | 2011/10/15 06:14 PM |
Article: Cooling and performance/watt | rwessel | 2011/10/15 09:56 PM |
Exponential growth of subthreshold leakage | Konrad Schwarz | 2011/12/15 07:56 AM |
Exponential growth of subthreshold leakage | Rohit | 2011/12/15 01:22 PM |
Exponential growth of subthreshold leakage | David Kanter | 2011/12/15 04:20 PM |
Exponential growth of subthreshold leakage | Iain McClatchie | 2013/01/07 12:28 AM |
Exponential growth of subthreshold leakage | Doug S | 2013/01/07 10:25 AM |
Exponential growth of subthreshold leakage | someone | 2013/01/07 11:12 PM |
Article: Cooling and performance/watt | Robert Pearson | 2021/07/26 09:45 AM |