Kab
04-29-2007, 01:24 PM
If fabrication process is understood, a P4 Prescott had high level transistor current leakage when off, a higher staged pipeline aswell than it's predecessors but this was attempted to compensate with higher clocks.
That is also what in effect screwed it up before reaching their 10GHz by 2006 forecasts, leakage, heat, power and stability.
IPC and in affect "it/s" were dealt with a combination of measures ensuring to 'do a single instruction faster' rather than 'do two at a slower speed'.
This is at a very simple level. ;)
Transistors are what it's effectively about on a microprocessor nowadays. Hence why AMD had done such a sweet feat by being able to process a dual core to fit existing higher clocked single core processor sockets on motherboards.
Anyway, onto the facts: http://www.rochester.edu/news/show.php?id=2585
The team has already had some luck in fabricating a prototype. The ballistic transistor is a nano-scale structure, and so all but impossible to engineer just a few years ago. Its very design means that this "large" prototype is already nearly as small as the best conventional transistor designs coming out of Silicon Valley today. Feldman and Diduck are confident that the design will readily scale to much smaller dimensions.
There's one hurdle the team isn't quite as confident about: "We're talking about a chip speed measured in terahertz, a thousand times faster than today's desktop transistors" Diduck says. "We have to figure out how to test it because there's no such thing as a terahertz oscilloscope!"
So you see, higher clock speeds is the future.. :D as well as higher costs. :(
Well that is a given on many dependencies actually.
That is also what in effect screwed it up before reaching their 10GHz by 2006 forecasts, leakage, heat, power and stability.
IPC and in affect "it/s" were dealt with a combination of measures ensuring to 'do a single instruction faster' rather than 'do two at a slower speed'.
This is at a very simple level. ;)
Transistors are what it's effectively about on a microprocessor nowadays. Hence why AMD had done such a sweet feat by being able to process a dual core to fit existing higher clocked single core processor sockets on motherboards.
Anyway, onto the facts: http://www.rochester.edu/news/show.php?id=2585
The team has already had some luck in fabricating a prototype. The ballistic transistor is a nano-scale structure, and so all but impossible to engineer just a few years ago. Its very design means that this "large" prototype is already nearly as small as the best conventional transistor designs coming out of Silicon Valley today. Feldman and Diduck are confident that the design will readily scale to much smaller dimensions.
There's one hurdle the team isn't quite as confident about: "We're talking about a chip speed measured in terahertz, a thousand times faster than today's desktop transistors" Diduck says. "We have to figure out how to test it because there's no such thing as a terahertz oscilloscope!"
So you see, higher clock speeds is the future.. :D as well as higher costs. :(
Well that is a given on many dependencies actually.