A console price tag, a different kind of computing
The PlayStation 3’s Cell processor was built for demanding game workloads, not as a conventional desktop CPU. It paired a general-purpose core with several smaller specialized processors designed to handle streams of calculations in parallel. That arrangement could be useful for scientific tasks—such as simulations or image processing—where many similar operations can be divided up and run at once.
The catch was that software had to be organized to feed those specialized cores efficiently. A PS3 was not a plug-and-play replacement for an ordinary research workstation; researchers needed programmers who could adapt their code and, often, link many consoles into a cluster.
Why buying consoles could make sense
The economics changed the calculation. Sony sold the PS3 as a consumer product at a price that, according to the fact described here, was about $240 below its cost to make. Researchers could therefore obtain Cell-based computing power for less than comparable purpose-built hardware would cost, even if they had to supply the networking, storage, and engineering work themselves.
That tradeoff favored institutions with workloads suited to parallel processing and staff able to handle the unusual software. Instead of paying more for a polished, general-purpose system, they accepted a less convenient machine to get more specialized processing capacity for the money.
The hidden cost of the bargain
The low purchase price did not make the consoles effortless or universally useful. Their advantage depended on matching the task to the chip: work that could be split into parallel pieces might run well, while software written for a standard processor could not automatically take advantage of the Cell’s design. The bargain was not simply cheap hardware; it was cheap hardware in exchange for specialized programming and system-building.
