I always find this sort of thing disappointing. They're not going for quality but quantity, and there's nothing special about that considering the budget that is poured down NASA's throat every year. Instead of building a giant network of processors working in tandem, it might help if they put more research into building better more applicable technology. But hey, that's just me. I can't afford to put twelve processors together in a computer, but I would get excited if they managed to get the speed of twelve processors into one that would one day become available to me at a reasonable price.
you can plase some of the people some of the time....
whats the point in a government funded agency thats had it's budget questioned many times doing expensive research into a more efficent processor? theres the obvious merits (sale, etc.) but it would take months, years to build a "12 in 1" type thing that would be better and more efficent. bundling a load of cheap, powerful processors together is a (albeit expensive) great way of doing it fast on the fly.
I see your point empawe, but it's not NASA's job to research better super computers and components. it's their job to use them.
Here is where I request that NASA donate it's old super computer to me <!--emo&:)--><img src='http://www.unknownworlds.com/forums/html/emoticons/smile-fix.gif' border='0' style='vertical-align:middle' alt='smile-fix.gif' /><!--endemo-->
Wow, they might even be able to run Doom3 on that <!--emo&:0--><img src='http://www.unknownworlds.com/forums/html/emoticons/wow.gif' border='0' style='vertical-align:middle' alt='wow.gif' /><!--endemo-->
<!--QuoteBegin-DragonMech+Jan 9 2005, 12:04 AM--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (DragonMech @ Jan 9 2005, 12:04 AM)</td></tr><tr><td id='QUOTE'><!--QuoteEBegin--> This + 200 Terabytes of RAM <!--QuoteEnd--> </td></tr></table><div class='postcolor'> <!--QuoteEEnd--> The average computer has about 512 MB of RAM... so, going on that average, if this system has 10,240 processors, then 5 Terrabytes of RAM would be more realistic.
Bah, NASA is pointless. Other than putting satellites in orbit, what else do they do? NASA gets a bunch of money every year to build spaceships and put people in orbit but for what reason, just to say they did it? Give NASA more money when they send people to Mars or something.
Pardon my ignorance if NASA actually does something. If they do, please inform me.
<!--QuoteBegin-emperor awesome+Jan 8 2005, 10:07 PM--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (emperor awesome @ Jan 8 2005, 10:07 PM)</td></tr><tr><td id='QUOTE'><!--QuoteEBegin--> I always find this sort of thing disappointing. They're not going for quality but quantity, and there's nothing special about that considering the budget that is poured down NASA's throat every year. Instead of building a giant network of processors working in tandem, it might help if they put more research into building better more applicable technology. But hey, that's just me. I can't afford to put twelve processors together in a computer, but I would get excited if they managed to get the speed of twelve processors into one that would one day become available to me at a reasonable price. <!--QuoteEnd--> </td></tr></table><div class='postcolor'> <!--QuoteEEnd--> Those types of improvements (speeding up what happens in the same amount of space on a processor) will come very slowly now. Transistors can't be made much smaller to speed them up, so most of those improvements come in the form of changing the architecture of the processor, which can only help so much. Unless someone finds a completely different basis for making computers (something other than silicon transistors, or other materials that use the same concept), the only way to have significantly more powerful computers will be to combine components together, as was done with the NASA computer. At least that's how I understand it. Someone correct me if I'm wrong.
<!--QuoteBegin-TheFrostmourne+Jan 9 2005, 03:41 PM--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (TheFrostmourne @ Jan 9 2005, 03:41 PM)</td></tr><tr><td id='QUOTE'><!--QuoteEBegin--> Wait could someone explain to me their choice of the Itanium over the Ultra-High end Xeons or Pentiums. Last I heard Itaniums were a 64-bit real flop. <!--QuoteEnd--> </td></tr></table><div class='postcolor'> <!--QuoteEEnd--> They probably got the whole cache for extra cheap. They need the 64 bit processing because they need to be able to do calculations with extreamly high #'s of didgets of accuracy, thus being a big buisness they probably just hammered out a mutually benificial deal with another big buisness (intel)
<!--QuoteBegin-Delarosa+Jan 9 2005, 06:53 AM--></div><table border='0' align='center' width='95%' cellpadding='3' cellspacing='1'><tr><td><b>QUOTE</b> (Delarosa @ Jan 9 2005, 06:53 AM)</td></tr><tr><td id='QUOTE'><!--QuoteEBegin--> i have actually touched a 100 terabyte drive... the things you can afford when your hosting GM and Ford websites... <!--QuoteEnd--> </td></tr></table><div class='postcolor'> <!--QuoteEEnd--> Can i touch you?
I think they should use the new power to help them either:
A) Create true AI B) Create smaller, faster, more powerfull processors
May I point out that typically single threaded things like games don't benefit at all from MP systems? The less 'parallellizable' the task the more you need good single core performance.
Now that dual cores are comming to the desktop though I bet we will see games starting to come out which take decent advantage of MP systems. But dual-core prescotts, eeewww, dual core pentium M is much more like it.
Comments
= ....
= .....
I'm sorry, there is no word for how beautiful something like that would be. :')
whats the point in a government funded agency thats had it's budget questioned many times doing expensive research into a more efficent processor? theres the obvious merits (sale, etc.) but it would take months, years to build a "12 in 1" type thing that would be better and more efficent. bundling a load of cheap, powerful processors together is a (albeit expensive) great way of doing it fast on the fly.
Here is where I request that NASA donate it's old super computer to me <!--emo&:)--><img src='http://www.unknownworlds.com/forums/html/emoticons/smile-fix.gif' border='0' style='vertical-align:middle' alt='smile-fix.gif' /><!--endemo-->
The average computer has about 512 MB of RAM... so, going on that average, if this system has 10,240 processors, then 5 Terrabytes of RAM would be more realistic.
Pardon my ignorance if NASA actually does something. If they do, please inform me.
Those types of improvements (speeding up what happens in the same amount of space on a processor) will come very slowly now. Transistors can't be made much smaller to speed them up, so most of those improvements come in the form of changing the architecture of the processor, which can only help so much. Unless someone finds a completely different basis for making computers (something other than silicon transistors, or other materials that use the same concept), the only way to have significantly more powerful computers will be to combine components together, as was done with the NASA computer. At least that's how I understand it. Someone correct me if I'm wrong.
They probably got the whole cache for extra cheap. They need the 64 bit processing because they need to be able to do calculations with extreamly high #'s of didgets of accuracy, thus being a big buisness they probably just hammered out a mutually benificial deal with another big buisness (intel)
why intel? sun SPARC or Cray systems should run better
Can i touch you?
I think they should use the new power to help them either:
A) Create true AI
B) Create smaller, faster, more powerfull processors
Now that dual cores are comming to the desktop though I bet we will see games starting to come out which take decent advantage of MP systems. But dual-core prescotts, eeewww, dual core pentium M is much more like it.