This is a couple of very cool video's from the Mythbusters it explains a CPU and GPC in something easy to understand.
The second video Jamie takes one for us to see!
I hope you enjoyed the videos.
Happy Holidays from all of us at the OnComputers show.
I have been folding proteins for the last six months. Got an Intel 8-core up to #1 at Rosetta@Home. But I've been looking at GPU-based parallel computing for the past six months.
ReplyDeleteFirst, nVidia is currently ahead in technology for Folding@Home (FAH). The CUDA drivers for nVidia 8000 series and newer cards are more mature than the recently updated AMD ATI 8.13 code for 4000-series graphics cards. I get 5x performance on an 8800GTX than a 4870x2, which should not be. So hold in there, ATI fans, new FAH code based on latest ATI drivers will be out soon.
The huge number of floating point engines on GPUs relative to ordinary CPUs is what is driving the scientific community interest in GPGPU technology. With the recent switch to high-level programming interfaces such as CUDA, programming is easier and so more projects will be attempted. Also, more than a few eyebrows were raised to see the combined computing power of 300,000 Sony Playstations beating the rest of the computing world, hands down, in protein folding.
So, I expect we'll see a lot of maturation in GPGPU computing over the next five years, and there will be some important scientific advances as a result of harnessing the vast potential of graphics computing engines.