Originally Posted by JinXed
I said right now.. multi-threaded program availability isn't quite there yet. By the time it is, Yonah will be out already
|and about "roughly the same" not really.. Well, it really depends on the spectrum.. a 2.13 P-M will beat a 3500+ in almost everything, even though the 3500+ is clocked higher.. (2.2 right? )|
and don't give me gamepc or tomshardware propaganda.
|if you go to extreme cooling, the A64s that break 4ghz don't calculate Pi 1M as fast as the Dothans clocked at that speed.. in terms of raw 1 threaded processing power, right now, Dothan is king,|
|the problem is that for desktops they're stuck on a dead platform. All that will change when Conroe releases for desktop though. P-Ms love tight timings, I can't wait to see what will happen when Intel integrates the mem controller on the die..|
Intel does not integrate memory controller on die for a very good reason, that it still relies on the old FSB architecture, which won't be replaced by the CSI until the arrival of the Whitefield; and there is no telling when or if CSI will replace FSB on mobile chipsets.
|If intel does the right thing, and broadens the pipelines instead of lengthening them, then AMD might want to start investing more in R&D.. the problem is that the longer the pipes, the higher the clock speed, and higher clocks can be marketed much better to the avrg Joe.. so I'm afraid Intel will go with marketing instead of performance|
You do know that AMD 64 has roughly the same number of stages as the PM, right?
-- 12-15 stages (estimated, since intel won't release spec sheets)
increased from 10 of the P-III
AMD 64 (K8)
-- 12 stages
increased from 10 of the Athlon XP (K7)