You are here

24 Core, 48GB RAM Linux cluster runs on 400W

I just read the story of Helmer, a Fedora 8 linux cluster in an IKEA Helmer cabinet. The story begins

3D computer rendering are very CPU intensive and the best way so speed up slow render problems, are usually to distribute them on to more computers. Render farms are usually very large, expensive and run using ALLOT of energy. I wanted to build something that could be put in my home, not make too much noise and run using very little energy... and be dirt cheep, big problem? :) no computer stuff cost almost nothing these days, it just a matter of finding fun stuff to play with.

Compared with an 8-core/16GB RAM MacPro (a favorite in the 3D community), this computer has 3x the CPU/RAM with approximately the same thermal profile, and cost $3500 in parts, which is about than half the cost of the much less powerful Mac Pro. And with Blender, the cost per node of render licenses is zero, a very affordable price if you are in the content-generation business. Moreover, as the site says, render jobs that took all night, now gets done in 10-12 min!

This is the kind of innovation that open standards (from the screw dimensions to the cabinet width to the power supply sizes to the electrical connectors to the motherboards etc.) make possible. And this is the kind of innovation that TPM, DRM, the DMCA and other restrictive measures either intentionally or unintentionally frustrate. The fact of commodity computing rewards ingenuity by allowing more and more things to be combined in innovative ways. Proprietary interfaces, restrictive legal covenants, and intentional limits on who or how technology should be used means we are stuck at whatever plateau the dominant industry leaders have deemed acceptable.

The Helmers project shows that there is 3x the performance at 1/2x the cost for those willing to innovate. By no small coincidence, this is what I see every day when I visit Red Hat customers who are replacing proprietary systems with open source solutions. It still amazes me how much money some companies will spend on these proprietary solutions year in and year out because they refuse to simply try something new and measure the results. Still, necessity is the mother of invention, and as more and more companies find themselves needing to change their status quo, I'm sure many of them will discover for themselves a whole new price/performance level.