Blitzkommando
Dec 20, 2006, 03:01 AM
The entire manufacturing process for both firms will have switched over to 65µm. This is a stark difference from the 130µm parts that were being produced just 18 months ago. Even so, by 2008 a 45µm process will be introduced and should be fully implemented by 2009.
End users should expect decreased overall temperatures and potential for increased clockspeeds over the older 90µm process. It will not be a magic change, but it should boost overall efficiency and reduce voltages even further. The reduction in manufacturing process should also help in increasing the number of transistors per core as well as increase number of cores per die.
AMD has in fact provided engineering samples to a number of hardware review groups of 65µm parts designed to run off of a mere 65 watts where before the same parts in a 90µm process would peak at 110 watts and 85 watts for Energy-Efficent (EE) models. The reduction should theoretically increase the number of parts in the ultra-efficient 35 watt envelope as well.
However, even with this it will be difficult for AMD to pull ahead or even equal Intel's success with the Core 2 Duo and soon to be Core 2 Quad series. The roles have essentially reversed for the two semiconductor giants, with Intel having the more efficient and more powerful parts at lower clockspeeds and the AMD parts being less efficient and forced into higher clockspeeds in attempt to equal performance. The AMD Ahtlon 64 FX-64 will mark AMD's first foray into the 3GHz+ range of speeds as a result of this showing the age of the four-year-old K8 architecture against the less-than-one-year-old Core 2 Duo architecture.
However, if AMD has their ideas of stream-processing with tremendous floating point calculation prowess proves successful they might very well pull ahead of Intel technologically for a number of years starting in 2009 with the introduction of their hybrid-APU architecture being first introduced into the ever-growing mobile market, another market that has been dominated by their nemesis in the industry, Intel.
-Norvekh December 20, 2006 03:00
End users should expect decreased overall temperatures and potential for increased clockspeeds over the older 90µm process. It will not be a magic change, but it should boost overall efficiency and reduce voltages even further. The reduction in manufacturing process should also help in increasing the number of transistors per core as well as increase number of cores per die.
AMD has in fact provided engineering samples to a number of hardware review groups of 65µm parts designed to run off of a mere 65 watts where before the same parts in a 90µm process would peak at 110 watts and 85 watts for Energy-Efficent (EE) models. The reduction should theoretically increase the number of parts in the ultra-efficient 35 watt envelope as well.
However, even with this it will be difficult for AMD to pull ahead or even equal Intel's success with the Core 2 Duo and soon to be Core 2 Quad series. The roles have essentially reversed for the two semiconductor giants, with Intel having the more efficient and more powerful parts at lower clockspeeds and the AMD parts being less efficient and forced into higher clockspeeds in attempt to equal performance. The AMD Ahtlon 64 FX-64 will mark AMD's first foray into the 3GHz+ range of speeds as a result of this showing the age of the four-year-old K8 architecture against the less-than-one-year-old Core 2 Duo architecture.
However, if AMD has their ideas of stream-processing with tremendous floating point calculation prowess proves successful they might very well pull ahead of Intel technologically for a number of years starting in 2009 with the introduction of their hybrid-APU architecture being first introduced into the ever-growing mobile market, another market that has been dominated by their nemesis in the industry, Intel.
-Norvekh December 20, 2006 03:00