Power Consumption: ATi vs NVidia

[chavvdarrr]

http://www.overclockers.ru/news/newsitem.shtml?category=2&id=1092690267
on russian, probably soon english version will be posted at xbitlabs.com

 

[madshi]

Thanks! That's very interesting for me. Especially the low power consumption of the plain 6800. I planned to buy a X800 Pro, but the 6800 looks mighty tempting after seeing this chart. If only it had gamma corrected 6xAA...

P.S: I'm a bit disappointed that Beyond3D still does not yet have power consumption information in its reviews.

 

[PatrickL]

I don't get it. Why do you compare 6800 NU and X800 pro power consumption ? They are not exactly in the same performance range.

 

[incurable]

Pretty interesting that the NV40U doesn't even use 10 W more than an XT PE under load.

(low-K really seems to pay off in idle situations)

 

[Dave Baumann]

P.S: I'm a bit disappointed that Beyond3D still does not yet have power consumption information in its reviews.

We're not electronics experts and we don't have the kit to do it.

Plus, how do we know whats being drawn over the AGP bus? With those that have auxiluary connectors, without some device for measuring the load across the AGP bus, you can only make assumptions whats being drawn across it.

(low-K really seems to pay off in idle situations)

IMO, thats more likely to be the design of the chip rather than Low-k itself. R420 was designed with all of the laptop power svaing mechanisms.

 

[madshi]

I don't get it. Why do you compare 6800 NU and X800 pro power consumption ? They are not exactly in the same performance range.

I'm thinking in power consumption ranges, not in performance ranges.

 

[madshi]

We're not electronics experts and we don't have the kit to do it.

Plus, how do we know whats being drawn over the AGP bus? With those that have auxiluary connectors, without some device for measuring the load across the AGP bus, you can only make assumptions whats being drawn across it.

Please have a look at how techreport does it:

http://www.techreport.com/reviews/2004q2/radeon-x800/index.x?pg=27

While not being perfect, it's easy to measure and at least a good indication of how high the power consumption in idle/burn is.

 

[Lezmaka]

Plus, how do we know whats being drawn over the AGP bus? With those that have auxiluary connectors, without some device for measuring the load across the AGP bus, you can only make assumptions whats being drawn across it.

http://www.xbitlabs.com/articles/video/display/ati-powercons_3.html

So, it’s easy to find the total power consumption of a graphics card from the additional power connector.



Well, this is only a portion of all the consumed power. Besides the 5v and 12v circuits that go through the additional power connector(s), there are also 3.3v, 5v and 12v power circuits going through the AGP slot. It’s more difficult to measure the current flowing in these circuits – you can’t plug an adapter with the shunts into the AGP slot!

What to do? I took it easy and just insulated the appropriate pins of the AGP connector on the graphics card with stripes of adhesive tape (pin A1 – 12v; pins B2, B3 – 5v; pins A9, B9, A16, B16, A25, B25, A28, B28 – 3.3v) and applied those voltages directly to the graphics card, through the prepared shunt, taking the voltages from the PSU.



This gave me more trouble with each of the cards – I had to mess around with scissors and tape searching for any points to attach 3.3v, 5v and 12v lines to and soldering wires to them – but this allowed calculating the total power consumption of the card with more precision. Moreover, some of the graphics cards don’t have an additional power connector, so the powering through the shunt, bypassing the AGP slot, was the only way of measuring their power consumption.

No way to know for sure how accurate that is, but that's probably the best way without spending who knows how much money on the equipment to find out for sure.

 

[Dave Baumann]

http://www.techreport.com/reviews/2004q2/radeon-x800/index.x?pg=27

While not being perfect, it's easy to measure and at least a good indication of how high the power consumption in idle/burn is.

I did actually look for a plug to measure it at the socket but couldn't find one. However, here again we see that there is a 20W difference between the XT PE and 6800 Ultra in these tests and only 10W with the results here.

http://www.xbitlabs.com/articles/video/display/ati-powercons_3.html

And you just highlighted exactly why we won't be doing that! Plus, they appear to be bypassing AGP power, putting the power through other circuitry that probably wasn't designed for that load - how does that affect things?

 

[chavvdarrr]

afaik they put the power through same traces/circuits through which agp-slot-power would flow.

 

[madshi]

I did actually look for a plug to measure it at the socket but couldn't find one.

http://kaufen.conrad.de/energy_check_3000.asp

(Germany, though)

However, here again we see that there is a 20W difference between the XT PE and 6800 Ultra in these tests and only 10W with the results here.

Yeah, I noticed that, too. But I think you can't do much wrong when using the techreport measurement method. So I trust that more.

 

[incurable]

However, here again we see that there is a 20W difference between the XT PE and 6800 Ultra in these tests and only 10W with the results here.

Yeah, I noticed that, too. But I think you can't do much wrong when using the techreport measurement method. So I trust that more.

Well, the difference is quite easily explained by the inefficiencies of the PSU's and the motherboard's VRMs that you do measure when using TechReport's method, but leave out when measuring at the socket/power connector of the card. And IMHO, the latter method is better. (i.e. closer to the truth)

cu

incurable

 

[Maverick]

Well, it wouldn't get you the *absolute* power consumption, but couldn't you measure the power consumption on the ATX connector to the motherboard, and calulate the difference between various cards?

 

[Neeyik]

X-bits method is pretty reasonable; the shunts make it quick to record with a decent voltmeter. I wonder if they factored in the power dissipation of the shunts though?

 

[LeStoffer]

Why does the 6800 Ultra draw 30 percent more power than the 6800 GT when the difference in clock speed is only 400 mHz vs 350 mHz and the memory 550 mHz vs 500 mHz?

 

[Cryect]

There is normally something similar to an exponetial increase in power usage and heat output when you clock a chip higher. So not too suprising to see such a large jump.

 

[thop]

Yes very interesting, and the power consumption is becoming a big factor for me. I'm quite impressed how well the 6800 does, and how much the X800 improved over the 9800 (in idle mode).

Does anyone have any graphs on older cards? Such like GF2/3/4, or the original Radeon and Radeon 8500.

 

[rendezvous]

There is normally something similar to an exponetial increase in power usage and heat output when you clock a chip higher. So not too suprising to see such a large jump.

Just increasing the clock would give an linear increase in power consumption. But to increase the the clockspeed you need to increase the voltage since the propagation delay is (roughly) inversly proportional to the voltage.

This would yield a quadratic increase in power consumption.

 

[plat]

Why does the 6800 Ultra draw 30 percent more power than the 6800 GT when the difference in clock speed is only 400 mHz vs 350 mHz and the memory 550 mHz vs 500 mHz?

from what i'v seen in the forums, the GT runs at 1.3v and the ultra runs at 1.4v. people who flash their GT's with ultra bios's usually get more out of their cores.

 

[fallguy]

There have been several of these tests. I have not seen any two with the same results. Hard to try and figure out which is accurate.