G70 Vs X1800 Efficiency

[Nite_Hawk]

GT also has a quad disabled.

And yet still appears to generally be faster than the XL with lower power consumption...

Nite_Hawk

 

[KimB]

Continuing from aaron's comments, do we know that the 7800GTX is pushing process clock limits as high as the X1800XT?

It doesn't appear to, given that it seems to overclock so well.

 

[trinibwoy]

It doesn't appear to, given that it seems to overclock so well.

Dave dropped some cryptic comment earlier about the max voltages recommended by TSMC for the 110nm process. If somebody knows what this upper limit is, maybe we can use it to extrapolate how much extra performance could be squeezed out of the G70 should Nvidia want to push it further.

 

[ERK]

Trying to remember my schooling...
Power goes like f*V^2, right?

 

[KimB]

Trying to remember my schooling...
Power goes like f*V^2, right?

Er, for a plain resistor. It really depends, though, as there are a number of factors that influence power consumption in a complex circuit like a microprocessor. I would say that the power consumption going as V^2 is a very, very rough first approximation to how a microprocessor behaves.

 

[3dcgi]

I was under the impression that caches are used to optimize bandwidth usage ie. not fetch the same texels twice. Although caches can also be used to store prefetched data it would seem to me that its smarter to use resources for keeping more threads in flight to work around the latency than to prefetch speculatively and add lots of caches so that the prefetches don't waste a lot of bandwidth.

Cache is used to optimize bandwidth. It serves a dual purpose. If you support 100 threads in flight you need more cache than if you only support 50 threads. Hense, cache and more threads are helping to hide the latency.

 

[3dcgi]

Do you think that the difference between cas 9 and cas 11 memory clocked at the same speed would be enough to significantly tip the scales toward the video card with the cas 9 memory? It seems that if the card with cas 11 memory has been designed to hide that latency as much as possible that it may play a small role, but not terribly significant. I seem to remember that back in the old days going from cas 3 to cas 2 yielded something like a 10% increase... Going from CAS 11 to CAS 9 on a board that is designed to hide CAS 11 latency seems like it might yield rather smaller benefits.

Nite_Hawk

I'm not really sure what kind of a difference it would make. It's been a while since I've looked at DRAM in any detail.

 

[nobody]

Er, for a plain resistor. It really depends, though, as there are a number of factors that influence power consumption in a complex circuit like a microprocessor. I would say that the power consumption going as V^2 is a very, very rough first approximation to how a microprocessor behaves.

No, it's indeed a very good approximation of the power consumption of CMOS devices.
A transistor is nothing else than a switched resistor, only during the switching process there is some current flow from positive supply voltage to ground. And this current flow is proportional to the supply voltage -> P ~ V^2.