Daily Archives:November 12th, 2011

Silicon heating

In the study/shed at the end of my garden, I’ve just turned off an elderly Pentium-based PC which consumed about 200-300W.

The result is that I now need to turn up the heating.

Is this foolish? I mean, the vast majority of energy used by CPUs, RAM, hard disks and PSU emerges in the form of heat in the room in which they’re located. Will my underfloor heating be more efficient? In the same way, I’ve been replacing my nice old filament lightbulbs with expensive and less-pretty LEDs. I know it makes sense in the summer, but in the winter…?

Of course, the problem with the PC was that it was indiscriminate in its heating tendencies – for the last few months I’ve often been sitting in here with the door or window open – whereas the real heating is thermostatically controlled. On the other hand, the PC could actually do useful work as a side effect – run backups, for example.

I’ve often thought that in houses of the future, computing power should be combined with the heating and air-conditioning systems: it would make much more sense to have them in the same room and then distribute pixels and heat from the same source. (Which is partly why I’m looking forward to future DisplayLink chips with full ethernet support).

In the meantime, how about a thermostatically-controlled PC? Has anyone done this? It would be off on hot summer days, come on as the temperature dropped, and you could schedule certain tasks – backups, downloads, compressing the TV programs you’ve captured, ripping DVDs, re-indexing document collections, uploading photos to the cloud – to run at times based at least to some degree on the usefulness of the energy consumption involved.

I feel a patent coming on…

© Copyright Quentin Stafford-Fraser