[citation][nom]silverblue[/nom]For most general computing, current CPUs are definitely "fast enough". It'd make a lot more sense (to me, anyway) to go the APU route (albeit improve the current CPU efficiency level) and maintain or slightly improve on a particular performance level but make the main selling point the energy and thus money you're going to save over the previous generation or competitors' products. I built a spare machine the other day running with an XP3200+, a 77W TDP CPU from 2002 that would be out-performed by an E-450 for a fraction of the power... if only I had one to hand, I'd have a very frugal and quiet machine.From my standpoint, if you keep throwing faster, hungrier CPUs and GPUs at the task, who is ever going to really spend their time developing to push the very limits of a specific architecture when the top CPUs and GPUs have far more grunt than is good for them? "Can it play Crysis?" is a question that shouldn't really be asked when you've got relatively decent hardware from the past four or five years.Sustainable computing. Sure, humanity strives for excellence, but what if we just decided to make do for a change? How much could your machine handle if developers realised that this was the fastest kit available and performance would have to depend more on software? If there wasn't any merit to this, companies like SeaMicro wouldn't have a market.[/citation]
I agree. Most of the time, my desktop's i5 750 CPU goes to waste. (I have no reason to upgrade it in the next few years for sure.) However, I have an E350 laptop, and the CPU actually does bottleneck things like web surfing and general performance noticeably there. For a low-end machine, I think a chip like that (or the E450) is a bit on the slow side, but a Pentium or A4 APU should be fine. It's stupid how some people keep upgrading when they don't need the power. I'm probably going to build a low-end machine like that sometime, and I'm thinking about underclocking it just for efficiency. It's not like anyone who uses it will notice.