# Thermodynamics- is there a temp limit for computers to work? **URL:** \ **Category:** Factual Questions\ **Created:** [April 24, 2019, 1:21am UTC](https://boards.straightdope.com/t/thermodynamics-is-there-a-temp-limit-for-computers-to-work/832967 "2019-04-24T01:21:42Z")\ **Posts on this page:** 5\ **Page:** 1
**Author:** ![am77494](https://avatars.discourse-cdn.com/v4/letter/a/d2c977/32.png) [@am77494](https://boards.straightdope.com/u/am77494)\ **Post date:** [April 24, 2019, 1:21am UTC](https://boards.straightdope.com/t/thermodynamics-is-there-a-temp-limit-for-computers-to-work/832967/1 "2019-04-24T01:21:42Z")
So in simple thermodynamic terms, computers organize / process data and thus locally reduce entropy while on the larger scale entropy increases per the laws of thermodynamics. As temperature increases (due to radiation effects) temperature (and hence entropy) tends to equalize or the medium gets homogeneous. So is there a thermodynamic temperature limit above which computing is not possible ? ( I am not asking if silicon or germanium semiconductors will work at high temps but if it is thermodynamically possible with any material ?) Since living organisms also operate on the same principle (local entropy reduction / organization), is there a temperature limit to life itself ? ( I am aware of extremophiles living in hot vents and nuclear reactor cores but maybe that’s not hot enough ?) ---
**Author:** ![DPRK](https://avatars.discourse-cdn.com/v4/letter/d/4491bb/32.png) [@DPRK](https://boards.straightdope.com/u/DPRK)\ **Post date:** [April 24, 2019, 2:20am UTC](https://boards.straightdope.com/t/thermodynamics-is-there-a-temp-limit-for-computers-to-work/832967/2 "2019-04-24T02:20:39Z")
The first thing that comes to mind is [Landauer’s principle](https://en.wikipedia.org/wiki/Landauer%27s_principle), which states that erasing one classical bit takes at least kT ln(2) energy, where k is Boltzmann’s constant and T is the temperature. The amount of heat you must be able to dump into the environment therefore increases at least linearly with the environmental temperature… ---
**Author:** ![MikeS](https://avatars.discourse-cdn.com/v4/letter/m/919ad9/32.png) [@MikeS](https://boards.straightdope.com/u/MikeS)\ **Post date:** [April 24, 2019, 11:44am UTC](https://boards.straightdope.com/t/thermodynamics-is-there-a-temp-limit-for-computers-to-work/832967/3 "2019-04-24T11:44:48Z")
> [@DPRK](#): > > The first thing that comes to mind is [Landauer’s principle](https://en.wikipedia.org/wiki/Landauer%27s_principle), which states that erasing one classical bit takes at least kT ln(2) energy, where k is Boltzmann’s constant and T is the temperature. The amount of heat you must be able to dump into the environment therefore increases at least linearly with the environmental temperature… It’s worth noting that this amount of energy is tiny. It’s estimate that [Google, Facebook, Apple, and Amazon have about 1 200 petabytes](https://www.sciencefocus.com/future-technology/how-much-data-is-on-the-internet/) stored. The Landauer limit to erase all of this data (at T = 300 K) works out to about 0.028 joules, which is comparable to the energy of a single peanut M&M after it falls about 4 feet to the floor. And even then, this limit can be (in principle) evaded via [reversible computing](https://en.wikipedia.org/wiki/Reversible_computing), where the information is not “lost” because every computing operation can be undone. Modern computers don’t work this way, because the Landauer limit is so very tiny; but perhaps one day it’ll be a concern. ---
**Author:** ![ftg](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/ftg/32/2801_2.png) [@ftg](https://boards.straightdope.com/u/ftg)\ **Post date:** [April 24, 2019, 12:32pm UTC](https://boards.straightdope.com/t/thermodynamics-is-there-a-temp-limit-for-computers-to-work/832967/4 "2019-04-24T12:32:35Z")
It basically comes down to the limits of the integrity of the materials used. Once those start to fall apart, there goes you computer. Consider the [Tinker Toy Computer](https://www.science20.com/brain_candyfeed_your_mind/great_tinkertoy_computer_tictactoe_domination-46197). It uses classic Tinker Toy parts to play Tic-Tac-Toe. Replace the parts with titanium alloy equivalents, scale up, etc. You have a computer … almost. It uses strings which need fine adjustment. Replacing those with chains or some such made of a high temp alloy may be a problem. There’s a problem of putting it together and the high temps. If you assemble it at low temps then it’s tricky to keep things in place as it is heated up. Assembling at high temps presents a lot of issues. Ditto a modern day Babbage machine. ---
**Author:** ![DPRK](https://avatars.discourse-cdn.com/v4/letter/d/4491bb/32.png) [@DPRK](https://boards.straightdope.com/u/DPRK)\ **Post date:** [April 24, 2019, 7:57pm UTC](https://boards.straightdope.com/t/thermodynamics-is-there-a-temp-limit-for-computers-to-work/832967/5 "2019-04-24T19:57:23Z")
> [@MikeS](#): > > It’s worth noting that this amount of energy is tiny. It’s estimate that [Google, Facebook, Apple, and Amazon have about 1 200 petabytes](https://www.sciencefocus.com/future-technology/how-much-data-is-on-the-internet/) stored. The Landauer limit to erase all of this data (at T = 300 K) works out to about 0.028 joules, which is comparable to the energy of a single peanut M&M after it falls about 4 feet to the floor. > > And even then, this limit can be (in principle) evaded via [reversible computing](https://en.wikipedia.org/wiki/Reversible_computing), where the information is not “lost” because every computing operation can be undone. Modern computers don’t work this way, because the Landauer limit is so very tiny; but perhaps one day it’ll be a concern. All true. The gross picture is that you want your computer to be small, fast, and reliable, and each of those aspects gets degraded at higher temperatures. Abstract thermodynamic limits (like the maximum amount of information a finite volume of space can physically contain) are not going to be as relevant as the fact that your computer is presumably going to have to contain (some) actual matter/materials, but concepts such as thermal noise and Landauer’s principle do illustrate that computation takes more and more energy at higher temperatures, and how are you going to cool all that off if your computer is sitting inside the surface of the Sun? Even so, conventional modern computers may eventually flirt with physical limits such as switching speed even before taking the whole thing and cooking it (and you’ll notice that your desktop or gaming computer requires cooling, and that experimental quantum computers are cooled to even lower temperatures).