AMD's 3GHz K10 to break 30,000 3DMark06

[sacha35]

K10: meet (Beverly) Krusher to the 10th degree

Source:http://www.theinquirer.net/?article=41970

 

[K404]

theres a s**t-storm on XS about this Its not impossible, but I think its a pretty big claim, especially at the GPU clocks.

Its possibly tri-Crossfire? That would be a coup in itself for AMD/ATI, but jurys very still out

EDIT: Sampsa (Who works for ATI IIRC) has said the bench was 3DM05 NOT 3DM06 which a pretty amateur mistake to make on Theos part.

Having said that, its still a good performance jump

 

[bazx]

with amd making the cpu and the gfx, chipset it would not be hard to fix the 3d scores

if they tunned the hardware and bios to produce good results in the main benchmarks

they do have alot of control over the there setup

 

[Master_G]

Wouldnt surprise me too much given the clockspeeds, they did say a while back that the design only really went well over 2.6 or something like that. I want to know how easy it was to get the CPU that high, and how representative it is of CPUs (i.e. a one off, or can a large chunk do the same)

It is all very well to say it performs great at 3GHz, but they are only being released at 1.9GHz atm, so the amazing performance only comes when you increase the clockspeed by half. Pretty much like saying OMG look how fast this Penryn is at 4.5GHz.

IMO AMD's designs have always been good at getting a high level of work per clock, but they have always struggled to get the clockspeeds high enough. If they could release a K10 at 2.5+ GHz that would be very good, but 1.9 just aint high enough.

Maybe some light at the end of the tunnel though for AMD.

G

 

[maverik-sg1]

Wouldnt surprise me too much given the clockspeeds, they did say a while back that the design only really went well over 2.6 or something like that. I want to know how easy it was to get the CPU that high, and how representative it is of CPUs (i.e. a one off, or can a large chunk do the same)

It is all very well to say it performs great at 3GHz, but they are only being released at 1.9GHz atm, so the amazing performance only comes when you increase the clockspeed by half. Pretty much like saying OMG look how fast this Penryn is at 4.5GHz.

IMO AMD's designs have always been good at getting a high level of work per clock, but they have always struggled to get the clockspeeds high enough. If they could release a K10 at 2.5+ GHz that would be very good, but 1.9 just aint high enough.

Maybe some light at the end of the tunnel though for AMD.

G

with amd making the cpu and the gfx, chipset it would not be hard to fix the 3d scores

if they tunned the hardware and bios to produce good results in the main benchmarks

they do have alot of control over the there setup

Agreed on both counts, the other AMD achilles heal is batch performance, on which I found on AMD64 only a small amount of certain batches had excellent overclocking potential, with these cores having adjustable volts on each core (or pair of cores I dont recall which) the potantial for batch preference coming into it could be more prominent than ever.

Still interesting times ahead.

Mav