• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Does anyone consider PSU efficiency when looking at hash/watt?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

I.M.O.G.

Glorious Leader
Joined
Nov 12, 2002
Location
Rootstown, OH
I don't know that I've ever seen this mentioned when people are talking about GPU settings and hashing efficiency (hash/watt).

If you run a PSU rated for 85% efficiency, and your rig needs 1000W of power, you'll be drawing ~1185W from the wall. 185W of power disappears due to inefficiency.

If you run a PSU rated for 92% efficiency, and your rig needs 1000W of power, you'll be drawing ~1090W from the wall. Only 90W of power disappears due to inefficiency.

Have I missed it and PSU efficiency is often mentioned when people look at hashing efficiency, or is that just left out all the time?
 
I don't know that I've ever seen this mentioned when people are talking about GPU settings and hashing efficiency (hash/watt).

If you run a PSU rated for 85% efficiency, and your rig needs 1000W of power, you'll be drawing ~1185W from the wall. 185W of power disappears due to inefficiency.

If you run a PSU rated for 92% efficiency, and your rig needs 1000W of power, you'll be drawing ~1090W from the wall. Only 90W of power disappears due to inefficiency.

Have I missed it and PSU efficiency is often mentioned when people look at hashing efficiency, or is that just left out all the time?


I never sat down to understand the whole efficiency thing...

My PSU is 80+ Bronze, and I pull 220 Watts from the rig (measured using ammeter clamped on the the power cord of the PSU)
220 watts = ~1.8 Amps.

It's the PSU in the sig, Coolermaster i600 (600 watt PSU)
 
When I try and figure out my efficiency, I just go by the number that the the unit is pulling from the wall, not estimates of what is being pulled from the psu
 
When I try and figure out my efficiency, I just go by the number that the the unit is pulling from the wall, not estimates of what is being pulled from the psu

Cool. I was wondering because this makes efficiency comparison against others problematic - settings can be the same and all, but depending on total draw, where you are in the efficiency range for your PSU, and the PSU's overall efficiency rating you'll get different results between people. Those with more expensive PSUs, or those pulling wattage from the PSU in a more ideal efficiency range, may show better efficiency not due to better settings but because of the balance of their power needs to the PSU's efficiency.
 
I bought Miahs platinum 1000w partly because of the effieciency but most because of the price. It would take a while to make up the cost difference.
 
I never sat down to understand the whole efficiency thing...

My PSU is 80+ Bronze, and I pull 220 Watts from the rig (measured using ammeter clamped on the the power cord of the PSU)
220 watts = ~1.8 Amps.

It's the PSU in the sig, Coolermaster i600 (600 watt PSU)

That sounds really low, but I don't know how close of a reading that actually gives or how much it depends on the quality of the ammeter. 80+ bronze doesn't mean much, other than it's not supposed to waste more than 20% of power at various load ranges (but it might anyway, because 80plus and all).

To know more about the actual efficiency, especially since it doesn't look like your PSU has been reviewed by a reputable place and they don't provide any specific efficiency information in the product literature... You'd have to apply a known load on the PSU, and then compare that to the actual reading you get on the ammeter. If you know you put a 300W load on it, your reading should be 380W if its running at exactly 80% efficiency, or less than 380W if its a bit more than 80% efficient. You still wouldn't know your exact efficiency though, unless you knew the tolerance or could verify the calibration on the ammeter. I got no idea how accurate those things are.

But basically, if you are getting accurate readings, whatever you are reading being pulled from the wall is generally up to 20% higher than what your components actually are running on, but for a gold or higher rated PSU the reading could be as little as 10% higher than actual usage by components.
 
I was thinking about this the other day, but in a slightly different manner.. See I play guitar, and a common practice is to use a powerconditioner before the power supply to your pedals, and to your amp. Just to make sure your getting the cleanest power possible due to signal degradation (Improves sound quality) Could a PC PSU benefit from the same kind of system? Not just for mining but in general.

Something like http://www.sweetwater.com/store/detail/M8x2/ Just don't know if it would support the wattage (their are more and they can get quite expensive)
 
Good PSUs have stuff in there to deal with dirty input and feed quality power to components, but I don't know if its job might still be easier with a conditioner. I haven't seen any testing or discussion before of if that would have benefits. Would be interested to know if it does.
 
Good PSUs have stuff in there to deal with dirty input and feed quality power to components, but I don't know if its job might still be easier with a conditioner. I haven't seen any testing or discussion before of if that would have benefits. Would be interested to know if it does.

I don't have the equipment to test it out myself. But i know this website with a neat forum that does.. :chair: hint hint :grouphug:

Anyway on topic.

The PSU in my sig (I know, it is what it is) is rated for 73% efficiency. With a GTX550ti @ 1000/1250 1.087v and the MSI R5850 I picked I couldnt mine with both overclocked, and the 5850 OC's stable to BIOS max 775/1125 undervolted But I'm almost positive a decent 600w psu from a reputable manufacturer could power the rig. But when I put this together I wasn't thinking mining. And was trying to get mid-low level performance with mid-low level prices.
 
I use a platinum PSU. There is an efficiency thread where a few people have posted numbers. I have another rig running now that I can add in too.
 
Could a PC PSU benefit from the same kind of system? Not just for mining but in general.

Nope. APFC already takes care of that. You really want to get the most possible efficiency out of a power supply with active power factor correction? Run it at 230V.
 
I actually ran a 220 line for my mining rig, but the kilawatt meter doesn't work with it, so it is hard to know if there is any improvement.

OW is pretty much the end all be all when it comes to PSU's. If he said a rock would power my rig I'd hook the rock up. Seriously. Been reading his reviews for YEARS
 
Well, that takes the fun out of it! :rain: Thanks OW.

Well, if you really want to spend a big pile of cash on unnecessary hardware for mining, be my guest ;)

so it is hard to know if there is any improvement.

Should be about a small percentage improvement across the board. Power supplies are designed primarily to work with the highest voltage on the market they'll be subjected to, which is not usually found powering computers in North America ;)
 
That sounds really low, but I don't know how close of a reading that actually gives or how much it depends on the quality of the ammeter. 80+ bronze doesn't mean much, other than it's not supposed to waste more than 20% of power at various load ranges (but it might anyway, because 80plus and all).

To know more about the actual efficiency, especially since it doesn't look like your PSU has been reviewed by a reputable place and they don't provide any specific efficiency information in the product literature... You'd have to apply a known load on the PSU, and then compare that to the actual reading you get on the ammeter. If you know you put a 300W load on it, your reading should be 380W if its running at exactly 80% efficiency, or less than 380W if its a bit more than 80% efficient. You still wouldn't know your exact efficiency though, unless you knew the tolerance or could verify the calibration on the ammeter. I got no idea how accurate those things are.

But basically, if you are getting accurate readings, whatever you are reading being pulled from the wall is generally up to 20% higher than what your components actually are running on, but for a gold or higher rated PSU the reading could be as little as 10% higher than actual usage by components.


My 6850 @ stock full load eats 127 watts, and that's also what the ammeter measured. It's pretty accurate I'd say :D
 
Card pulls from board and psu, so that isn't a good source for calibration, though I would take that to indicate its relatively accurate. Just not sure how big or small "relatively" is.

Could be a fairly trivial amount of power the card pulls from the board, I don't know the technical details - I just know you can burn out mobos with subzero OC and multi sli/crossfire setups if you don't take extra measures to help out with power going to the pcie slots. So that indicates to me that whatever is pulled from the board may not be much under single card conditions, but it probably isn't trivial either.
 
Card pulls from board and psu, so that isn't a good source for calibration, though I would take that to indicate its relatively accurate. Just not sure how big or small "relatively" is.

Could be a fairly trivial amount of power the card pulls from the board, I don't know the technical details - I just know you can burn out mobos with subzero OC and multi sli/crossfire setups if you don't take extra measures to help out with power going to the pcie slots. So that indicates to me that whatever is pulled from the board may not be much under single card conditions, but it probably isn't trivial either.

Well 600 watts = Max 5 amps from the wall, and I'm only using 1.83 Amps...

I'm also :confused: on what you mean "Could be a fairly trivial amount of power..."?
 
i considered the PSU efficiency. but i didnt go all out with it.

yes, a 80+ platinum with like 94+% efficiency is great, and will cut down on overall power use, but they are also much more expensive than the bronze/gold PSUs.

personally, i'll look to 2x 80+ bronze/gold in the 700-800W range. much much much cheaper than say a single 80+ plat 1500w

my corsair 750w was like 75 bucks from best buy, and the gigabyte odin 800w was like 65 bucks used
 
so combined, the 750w corsair and the 800w gigabyte, cost me $140, with a power capacity of 1550w.

for comparison, a decent 80+ gold 1500-1600w PSU would be like $400+
 
Back