Two identical computers have a 150 watt difference, how could this be?

[Brian DeLoach]

I have two computers that are completely identical to each other. They each run 4x 5830s. One runs at 645 watts, while the other runs at 790 watts. I can't for the life of me figure out why there is such a huge disparity between the two. All clocks are identical, as well as the OS.

Anyone have a clue to this mystery? 

Alphy

[Xephan]

I have two computers that are completely identical to each other. They each run 4x 5830s. One runs at 645 watts, while the other runs at 790 watts. I can't for the life of me figure out why there is such a huge disparity between the two. All clocks are identical, as well as the OS.

Anyone have a clue to this mystery? 

Alphy

Using same PSU model, measured using same meter and running at the same voltage and memory clocks?
Are they putting out identical hashrates as well?

[TiagoTiago]

There is a dead rat inside one of them?

[Swishercutter]

I have two computers that are completely identical to each other. They each run 4x 5830s. One runs at 645 watts, while the other runs at 790 watts. I can't for the life of me figure out why there is such a huge disparity between the two. All clocks are identical, as well as the OS.

Anyone have a clue to this mystery? 

Alphy

Are they wired exactly the same...same rails?  One rail loaded more on one could decrease efficiency.

Are the hashrates the same? One may not be mining as hard.

How about affinity?  Does one computer have it on and the other not?

[n4l3hp]

Probably different CPU and GPU utilization. Or one has power management disabled or something.

[Swishercutter]

Probably different CPU and GPU utilization. Or one has power management disabled or something.

Ya...forgot about bios stuff.  One might have all the power saving settings on.

[haploid23]

every video card and CPU is different, even if they come from the exact same batch. some are lucky chips in that they overclock really high, some are "bad" chips in that they're crappy overclockers or require a lot of voltage to achieve the same clocks when overclocking. now your problem could be that in the rig with 790w, you have some high leakage chips in there. a high leakage chip is one that will draw massive amounts of power with just a tiny bump in voltage or clocks.

try doing this: leave everything the same, and switch the 4x 5830 from one rig to the other, and do the same with the rig, then measure the power. if the power consumption numbers switched as well, then you know it's from the GPUs. but if those power consumption numbers don't switch, then you know it's from the CPU/motherboard/PSU/etc.

[sveetsnelda]

Meh.  I have 32 of the same card, 8 same motherboards, 8 same PSU's, and they all pull a similar amount of power.  Chips being binned differently aren't going to cause 150 watt swings in power.  They'd never be able to put an accurate TDP number on the thing...

[Axez D. Nyde]

I overheated a PSU once, it's still working, but the efficiency has gone down alot. maybe something similar happened to one of your PSUs?

[Kermee]

They each run 4x 5830s. One runs at 645 watts, while the other runs at 790 watts.

Sounds like one, you downclocked the memory on one rig (645W) and not the other (790W).

I save about 150W downclocking the memory of my 4x5830 rigs... They run between 620W-640W vs ~800W.

Cheers,
Kermee

[teflone]

They each run 4x 5830s. One runs at 645 watts, while the other runs at 790 watts.

Sounds like one, you downclocked the memory on one rig (645W) and not the other (790W).

I save about 150W downclocking the memory of my 4x5830 rigs... They run between 620W-640W vs ~800W.

Cheers,
Kermee

Your numbers just answered a question ive been trying to answer for a long time! thanks..

I can fit one more 5830 on my 750 w   lol

Thanks, Useful numbers

[BkkCoins]

If you have similar power rated PSUs but different brands/efficiencies then that can account for quite a big power difference at the wall. Mining is one app where higher efficiency PSUs really pay for themselves.