why

how come, out of both a CPU and GPU of the same price, the GPU will perform circles around the CPU (assuming the software being used has support for GPU acceleration)

and how come more developers aren't going for GPGPU stuff

just one example:
my 5870 can achieve upwards of 160Mtrips/s in MTY, whereas my i7 980X only gets around 20Mtrips/s

other examples would be Folding@Home, BitCoin mining, Lightworks, MadVR, etc.

cretin detected

oh, forgot to mention, though it should be pretty obvious: a 5870 is way cheaper than a 980x, so at least in this situation, calculations via GPUs are extremely cost efficient by comparison

>>2
hmm?

GPUs suck shit. You don't notice it, and on the contrary, think they are very fast, because there's a lot of tiny little cores doing embarrassingly parallel tasks. Inherently sequential tasks will run slow as fuck on a GPU core.