By using this site, you agree to our Privacy Policy and our Terms of Use. Close
NJ5 said:
The Cell is a great processor, no doubt about that... But is it a great processor for making games with?

That's the question which Sony should have researched better IMO.

Sony researched it fine, but they originally intended to give it a dual core CPU in addition to the SPEs. The cost of the PS3 was prohibitive, so they cut whatever they could from the system and they decided that was one thing to be cut. They knew it would be an issue, because the Cell's SPEs are very clearly not suited well toward the majority of algorithms game engines employ (I'm talking AI, phsyics, collusion detection, animation, rendering, etc) -- they do need good floating point performance, but they also need good general CPU powerformance for searches and things like that.

Sony knew the new version of the Cell wouldn't be optimal for games, but they also knew it would be valuable for embedded applications and they could still convince people it was very powerful by showing some theoretical benchmarks or best case nothing-to-do-with-games scenarios.

You have to realize that Sony started development of the Cell with its partners well before the 360 was released. They originally believed that (1) the cell would have more than one general purpose core -- and it still can -- and that (2) the 360 wouldn't be as powerful as it is. Sony is no stranger to esoteric architectures, but this is one where there isn't really a way to win. Factor 5 couldn't find it, after all.

Sony and its partners couldn't very well throw away the billions they spent developing the Cell, could they?  Like a terrible but costly CG scene in a movie (most of King Kong), they leave it in there because they had to justify the cost of it and it would be more difficult to go back and say it was a costly mistake.