Graphics to bit-tech staff how you test hardware

Hi folks

I was wondering how you test new hardware, i am particularly interested in the: Zotac GeForce GTX 260 Amp² video card. Especialy the power consumption. Wich i am a liitle bit worried about.

http://www.bit-tech.net/hardware/2008/10/22/zotac-geforce-gtx-260-amp2-216-edition/10

At idle in your test set up it use 190 WATT and in load 337 WATT. Is this the power draw of your test set up?, i've searched other reviews and there are sometimes huge differences between 1 review and another i payed attention to what cpu was used in the power draw mostly overclocked or extreme cpu's all of them had a TDP of 130 if i am not mistaken. Did the cpu went to 100% load when gaming?

Intel Core 2 Extreme QX6850, XFX nForce 780i SLI motherboard, 2x 2GB Corsair XMS2-6400C5, Seagate Barracuda 7200.9 200GB, Enermax Galaxy DXX 1000W PSU, Vista 64.

In other words you did not include a monitor, system fans, dvd-drive? If the answer is yes i should add up 337 + monitor 45WATT + 6 system fans 12 WATT + dvd-drive 15 WATT (when in action) and a pump for watercooling 25 WATT + soundcard 7 WATT + keyboard and mouse 5 WATT = 446 WATT?? Could you please shed some light to these questions?

And if there are people here with GTX 260 in their rig and if its not to much to ask could you post your TOTAL power (including monitor fans etc) draw of your system at idle and load?

Thanks in advance.

 

You don't need to include a monitor for PSU calculations, but you do for the total power consumption.

Iirc, the BiT hardware testing method is with everything in the test setup for simplicity, so CPU cooling fans etc included.

Either way, it's only an approximate science really, BiT's numbers will be good relative to each other, but you'll probably see pretty big differences from one setup to another.

I'm sure Bindi or another Ed. staffer can clear it up.

 

Everything connected to the PSU we measure - so basically what is paid for at the wall, and the PSU is an 80+ Enermax Galaxy 1kW. It's virtually impossible to decipher what the power draw of a PCI-Express card is unless you do some hardcore board modding.

In the GTX 280/260 architecture article it'll tell you the Nvidia spec'd theoretical max, but that's an unrealistic scenario - basically all guns blazing. Games don't do that, we test what the typical max is in-game (Crysis in this example). The "typical idle" values - GTX 260 is about 20W idle and specific load is the difference minus this, so 127W.

 

These guys measured 340W loaded which is pretty damn close to the Bit numbers. That card at load will swamp any differences in the CPU and other components' wattages.

So where are these "huge differences between 1 review and another"?

 

First i want to thank the people who responded

here is one idle 115 load 249

http://www.techpowerup.com/reviews/Zotac/GeForce_GTX_260_Amp2_Edition/24.html

and

Dont know if they included a monitor...they put it in their test setup..but it
seems "low" for the power draw..as for the cpu E8500 overclocked to 3.6ghz does
this means the cpu draws more power? idle 144 load 300

http://benchmarkreviews.com/index.p...k=view&id=238&Itemid=72&limit=1&limitstart=11

And this one its another brand but it has 216 shaders like the Zotac, i do not know if they included a monitor in their TOTAL power draw setup.

http://www.guru3d.com/article/evga-geforce-gtx-260-core-216-superclocked-review/4

http://www.guru3d.com/article/geforce-gtx-260-review/4 (reference model)

So my estimation was wrong? 446 WATT at load is to high? When i
looked more carefully to the reviews perhaps some motherboards are more energy efficient
then others just like the cpu's. I saved my money for a new rig one that last for 3
or perhaps even longer LGA 775 until core i7 came along. Well i was happy about the benchmarks.
But i am worried about the power draw of the i7 cpu and a GTX 260 when gaming, i pay all of my bills. (I did bought my self a very efficient refrigerator old one 130WATT new one A++ 50WATT. And high efficient light bulb 8WATT but with the light of a 60WATT light bulb saved my self 132WATT)
It will be my second build. I did not payed to much attention to the cpu's used in the reviews, (difference in power draw cpu) for that i am sorry for the confusion, (huge difference between reviews) What do you recommend me? Just go for the 9800GT with 1680 x 1050 or 1920 x 1200 monitor, i want to play with good frames. But then again maybe future games will be to hard for the 9800GT in high resolution.

Thanks in advance
PS: sorry for my english.

 

Found some Interesting info about the upcomming GTX 270 and GTX 290. Very promising figures especialy power draw, 50% less power draw then the current GTX260 and GTX 280. I hope this info is true, just have to wait and see. See link:

http://theovalich.wordpress.com/2008/11/11/55nm-gt206-gpu-powers-both-gtx290-and-quadro-fx-5800/

 

This is your lucky day, pilgrim. I can sell you an XFX7200GS that runs on fresh air (almost).

 

You have found your nirvana apparently Uber-pilgrim otherwise you wouldn't sell me things that run on air (almost)....And the rest..... by riding a bike connected to a transformer

Nice going a serious question and a lame answer, moderator....

 

they are a bunch of techno mages, we can never understand there ways

 

For the love of God, didn't we already have the effective GTX270, only they refused to call it that? The 280 is a 240-stream processor model, the 260 has 192 and the "260 Mk.II" has 216...is there even room left in the middle for a GTX 270?

I think nvidia just like keeping us confused :S

 

I don't see why power draw can't be measured more accurately for a graphics card.

Surely run a PC with single card, idle and load and then same PC with (dual card - SLI or Xfire) also idle and load. Then take one from the other. You'd want to run the load test on a well optimised utility/benchmark to make sure the second card was getting good use but I think that's the ideal way.

As for low power cards, if you're particularly concerned about power try the new ATI 4670. It has an excellent performance to power draw ratio.

 

The best way would simply be to rig up a PCI-e extender with a bunch of ammeters on every line. That way you know the current draw for each voltage under load.

Even then, though, you still don't know how much power is wasted in the mobo/PSU circuitry converting enough power for your card, so your true draw is virtually unattainable.

The best way is look at a couple of reviews, get a ballpark figure, and budget well for your PSU. Something well-branded and 100W or so more powerful is probably best.

 

Maybe its because the GTX 270 and GTX 290 is based on a 55nm proces while the old 260 and 280 are based on a 65nm proces, wich in turn should run more cooler its less expensive to build and to buy, at least i hope so, have more room to overclock because its doesnt generate that much heat like the "old"versions so you got room to "play" (overclock) and less stress on the GPU components... less heat. And dont forget lower power consumption the idea sounds great if the link i provided is correct. Basically its the same gtx 260/280 (performance) but with nice advantages, cooler, more efficient etc, hence 55nm. (And maybe it will run even faster then the old models) Atleast thats what i read about on the net, wheter this is true il know for sure when the cards are getting reviewed.

EnglishLion Thanks for the tip about the new ATI 4670 GPU il look into that, never heard of that gpu.

Krikkit yes your right about +- 100 WATT more for a PSU, for me it means more stability when the psu doesnt get stressed to its limit (reach your limit or go over it for amps and wattage and get a unstable pc wich i dont want

Sorry could not multiple quote i'm new with forums.

thanks again for the replys guys.