• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

If anyones wondering what 3x 7970s draw in a watercooled rig...

Associate
Joined
21 Feb 2012
Posts
200
Location
Blackpool. UK
Hi everyone,
I have read all sorts of power figures for a trifire 7970 setup in mags and on websites so decided to buy my own meter and measure it myself to ensure PSU was not causing any bottlenecks here, so the answer is....

991wMedium.jpg


Tests were:

Heaven 2.5 @ 1920 x 1080 with everything at full and that never saw in excess of 850w.

3D mark 11 on extreme didnt exceed 935.

3D mark vantage interestingly was the one that will repeatedly peaked at almost a kw on the return to jerico scene.



Rig of course is totally relevant, so here it is:
Mobo: Asus Maximus Extreme IV - Z
CPU: Intel Core i7-2600K
GPU: 3 x Sapphire Radeon 7970's
RAM: 16GB of Corsair Vengeance 1866MHz
SSD: 2x OCZ Vertex 3 Sata 6 SSD
HDD 2x 1TB Western Digital Velociraptors
PSU: XFX Pro 1250w Black gamer Edition
Optical: Pioneer 12x Blu Ray Combo Drive
Pump: Alphacool VPP655 - D5 Pump
Fans 120mm: 3 x Scythe Gentle Typhoon 1850RPM - AP15s on the radiator.
fans 140mm: 3x Corsair 140mm fans around the chassis.
Lighting: LED's and a UV inverter so far.

Just thought it may help people in future who are wondering the same thing.

Eyefinity testing at 5760x1200 in battlefield three with all settings at ultra will be the next test... if we ever get a driver that lets me run trifire with BF3. :(

**Edit**
Sorry, to add, the CPU is running at 5.2ghz and the GPUs all at 1125 with ram @ 1575
 
Last edited:
Remember to account for the PSU efficiency (Gold for that one) so its roughly 90% of that figure.

Whilst I totally understand what you are saying, N-Power would disagree with you come bill time. In the real world, the "PC" as a whole, is consuming 991W peak.

However, if you get a more efficient PSU, say 90%, yes, you could tweak that figure down somewhat to maybe 900w.
 
Try 3DMark11 on P preset plz, that is where i recorded my peak usage of around 900w if i remember correctly with sli 580's, the P preset ran higher wattage than X preset.
 
That is a really impressive rig,sick with envy.
If you can afford a build like that Im sure the electriciy bill is of no concern lol.
Great to see a member posting real results of psu use.
Again,smashing setup.
 
However, if you get a more efficient PSU, say 90%, yes, you could tweak that figure down somewhat to maybe 900w.

Your PSU is Gold rated so should operate at approaching 90% efficiency at high loads.

The PSU is only outputting around 90% of what you see at the mains so only around 900W of its 1250W at the highest reading you've recorded.

But it's still drawing close to 1000W from the mains at your highest reading and that's what you'll be paying for.
 
That is a really impressive rig,sick with envy.
If you can afford a build like that Im sure the electriciy bill is of no concern lol.Great to see a member posting real results of psu use.
Again,smashing setup.

Thank you. :)

Try Crysis 2 and see if you get over 1000!

Trying exactly that tonight. :)

Your PSU is Gold rated so should operate at approaching 90% efficiency at high loads.

The PSU is only outputting around 90% of what you see at the mains so only around 900W of its 1250W at the highest reading you've recorded.

But it's still drawing close to 1000W from the mains at your highest reading and that's what you'll be paying for.

Yeah, I do totally understand you mate, For the purpose of this topic I am trying to report the electricity used by the PC as opposed to the energy actually used by the PC components themselves after the PSU.
 
Remember to account for the PSU efficiency (Gold for that one) so its roughly 90% of that figure that the parts are sucking from the PSU.

A quick question about this- in terms of making sure a PSU is up to the job, do you need to look at the 'at the wall' measure? I.E. in this case 991W, meaning a 1kw PSU would be cutting it very fine, or would a 1kw 90% efficiency PSU technically be ok up to 1100w (at the wall measurement)?
 
Whilst I totally understand what you are saying, N-Power would disagree with you come bill time. In the real world, the "PC" as a whole, is consuming 991W peak.

However, if you get a more efficient PSU, say 90%, yes, you could tweak that figure down somewhat to maybe 900w.

Point he was making was that 991w wasn't the power the PSU needs to source. Calculating that gives you a figure you can use to spec PSUs independent of the PSU efficiency.
 
Point he was making was that 991w wasn't the power the PSU needs to source. Calculating that gives you a figure you can use to spec PSUs independent of the PSU efficiency.

That's not strictly true though surely?

You want to know what power you need to Input to support your load.

If your components draw a true 500Watts your PSU needs to source the 500Watts + the overhead lost outside of it's efficiency rating.

No point speccing a 500Watt PSU for a true load of 500Watts when you need 550Watts to account for a 90% efficiency.

You still need to factor in the losses, they are not 'free'.
 
When i ran 3x 6970's with a Antec 1000w all games would run ok but BF3 would max the power usage out and reboot my pc.. got the xfx1250 and all was ok.
 
That's not strictly true though surely?

You want to know what power you need to Input to support your load.

If your components draw a true 500Watts your PSU needs to source the 500Watts + the overhead lost outside of it's efficiency rating.

No point speccing a 500Watt PSU for a true load of 500Watts when you need 550Watts to account for a 90% efficiency.

You still need to factor in the losses, they are not 'free'.

What matters is the DC power your components draw. Not the AC power drawn from the wall, as this can vary for two reasons: 1) power efficiency in terms of thermal losses, and 2) Reactive power vs real power.

I would adjust power drawn from the wall for at least the operational power efficiency before deciding what the components are drawing. That way, say if 991w is on a PSU that operates at 90% efficiency then you use that when speccing for a different PSU which may, for example, have 80% efficiency. In this instance as long as the power the PSU sources is within the range, it's good.
 
Back
Top Bottom