http://news.bbc.co.uk/1/hi/technology/6354225.stm
Probally need it for the next version of windows! :)
Regards
Probally need it for the next version of windows! :)
Regards
future computing
Miscellaneous Forums/General Discussion/future computing And the Terminator is born. :)
Hey, is the company developing this named "Cyberdyne Systems" ??!!
"Mega-core" processors like this would be extremely well suited for game/simulation programming. Just think of the level of detail you could put into physics and AI simulations with ~500 times today's CPU power - amazing. This could even replace video and physics cards eventually.
Maybe someday upgrading your CPU power will even be as easy as upgrading your RAM - simply plug in a new CPU chip and add an extra terra-flop or two :)
Didn't they say it was being developed by Intel?
Maybe someday upgrading your CPU power will even be as easy as upgrading your RAM - simply plug in a new CPU chip and add an extra terra-flop or two :)
Hey, is the company developing this named "Cyberdyne Systems" ??!!
Didn't they say it was being developed by Intel?
Probally need it for the next version of windows! :)
More like the current version! i.e. VistaJohn J.: Terminator? Skynet?
Didn't they say it was being developed by Intel?
Oh..pheww....close one!! Otherwise, we'd be in trouble...
John J.: Terminator? Skynet?
Oh.. I've never seen The Terminator, so that's probably why I didn't recognize "Cyberdyne Systems" :)
Does it automatically split instructions up, or do you have to program 80 threads to get the most out of it? Maybe this can do that real-time raytracing I wanted.
Does it automatically split instructions up, or do you have to program 80 threads to get the most out of it?
But to take advantage of the extra processing power, programmers need to gives instructions to each core that work in parallel with one another.
It would seem you have to program the threads manually somehow.
Maybe this can do that real-time raytracing I wanted.
You're right - if this technology proves to be successful, real-time raytracing may become the new standard in real-time graphics.
Possible uses include artificial intelligence, instant video communications, photo-realistic games and real-time speech recognition, said the firm.
Isn't that the same possible uses as the gHz chips of today? :bIn fact, I swear that belief has existed since practically the beginning of desktop computing.
Hm, deja-vu!
John J: please go out and buy all 3 terminator movies on DVD right now, find a 24 hour video shop if you can. Your life will thank you for it.
Well, any little does help. Now, wonder how long until Cyberdyne Systems gets a hold of it....
John J: please go out and buy all 3 terminator movies on DVD right now, find a 24 hour video shop if you can. Your life will thank you for it.
No movie is that good, is it? :) I'll take your word for it though ...
Isn't that the same possible uses as the gHz chips of today? :b
Artificial intelligence, speech recognition, realistic games, etc. are still far from perfection. Giving computers 500x and maybe 5000x the processing power this way opens up many new real-time computing possibilities (real-time ray tracing, for example) for the average home computer that was previously unimaginable.
*images ways to break down DrawImage() into 80 steps*
*images ways to break down DrawImage() into 80 steps*
If DrawImage() if hardware accelerated, you don't have to - that's exactly what modern video cards do. That's why video cards are so much faster than current CPUs when it comes to rendering, which is inherently a "parallel" process.
Parallel real-time neural networks...oooh.
It'll be interesting to see how long it takes developers to properly utilize that many cores.
It'll be interesting to see how long it takes developers to properly utilize that many cores.
No movie is that good, is it? :) I'll take your word for it though ...
The third one is pants, but the two first ones are prety good. It'll be interesting to see how long it takes developers to properly utilize that many cores.
Well it's taken them nearly two years to not utilize 2 cores, if we give them 2 more, they will probably have it all nailed down. If we assume a linear development of core utilization from this point in time, that would be in about 2167. Ofcourse technically there's no appreciable difference between utilizing 2 and utilizing 80 cores. Then there are the computers out there doing real work like servers, scientific emulations, rendering farms etc. which will efficiently use as many cores as you can throw at them.
which will efficiently use as many cores as you can throw at them.
I'm sure the Cell chip will enjoy a much bigger success in IBMs "BlueGene" Blade Servers than in the PS3.However your assumption is somewhat flawed. There is a significant difference between having one computer with 80 cores (which can use on-die communication) and 80 computers with single (or even 40 computers with dual) processors like most current supercomputer cluster configurations (which need an external bus for communication).