Post Your Mining Rig Efficiency

[Mousepotato]

I thought there was another thread about this, but I can't find it to save my life.  SMF Search sux  

Anyway, let's see your mining rig efficiency in MH/j (or MH/w, I'm not sure what the diff is exactly).  I added another 5970 to my system and finally figured out how to tune down the CPU cores.

CPU: AMD Phenom II x4 960T, throttled down to 1-core
GPU: 2x 5970 @ 625/150MHz @ 899mv
MH/s: 1128
At the plug: 394W
2.86 MH/w

Whoa, I think I finally beat D&T!!!  Except now he has a bunch of BFL Singles

Edit: New numbers after adding another 5970:
GPU: 3x 5970 @ 625/150MHz @ 899mv
MH/s: 1690
At the plug: 554W
3.05 MH/w

[1714791638]

1714791638

[1714791638]

1714791638

[1714791638]

1714791638

[1714791638]

1714791638

[rograz]

best rig i've got
5xSapphire 5850 xtreme 690-720Mhz 0.95V
1x XFX 5870 ref 0,95V 710MHz
some MSI X48 775 mobo with a c2d celeron at 800 mhz
Corsair HX850

Total: 1760MH/S, 562W  at the wall giving 3,13MH/W

[legolouman]

I thought there was another thread about this, but I can't find it to save my life.  SMF Search sux  

Anyway, let's see your mining rig efficiency in MH/j (or MH/w, I'm not sure what the diff is exactly).  I added another 5970 to my system and finally figured out how to tune down the CPU cores.

CPU: AMD Phenom II x4 960T, throttled down to 1-core
GPU: 2x 5970 @ 625/150MHz @ 899mv
MH/s: 1128
At the plug: 394W
2.86 MH/w

Whoa, I think I finally beat D&T!!!  Except now he has a bunch of BFL Singles

Wouldn't a sempron get you better efficiency?

[Mousepotato]

Wouldn't a sempron get you better efficiency?

Probably. This is my gaming rig though   I could probably knock another 50W off by going with a low power CPU, which would put me in the 3.1X MH/w range.  But that would mean I'd have to swap CPUs to play games, and that's a big hassle.

[legolouman]

Wouldn't a sempron get you better efficiency?

Probably. This is my gaming rig though   I could probably knock another 50W off by going with a low power CPU, which would put me in the 3.1X MH/w range.  But that would mean I'd have to swap CPUs to play games, and that's a big hassle.

I'm sure your GPU's love having the clocks and voltages change that much, so often.

Just saying.

[DeathAndTaxes]

Whoa, I think I finally beat D&T!!!  Except now he has a bunch of BFL Singles

I don't have any singles but you didn't beat me either.   Still 2.86 is pretty impressive.

My new rig configuration:
4x5970
3.2 GH/s @ 1080W
3.02 MH/W

That is on 240V circuit, 80-Plus Gold PSU, usb drive running linux, optimized BIOS, underclocked & undervolted sempron, and >3 GH/s per rig.

I am pretty sure it isn't possible to get higher efficiency from a GPU rig without undervolting the GPUs.  A 4x7990 (or possibly 3x7990) rig will likely beat that though.

[JinTu]

I'm still doing some tuning and will soon make the move to a passively cooled water cooled setup but my current measurements show it ranges from 4.52-2.78 MH/W for a dual 6990 rig with the memory underclocked to 150MHz. I run the fans at 50% during daylight hours to regulate the WAF


cgminer: 2.3.1
CentOS: 6.2 x68_64
AMD App: v2.5
AMD Catalyst: v11.11 (8.911)
Volts: 1.00V
Mem Clock: 150MHz (OC BIOS reflashed)
MH/s (night time): 926.33 968
MH/s (day time): 795.63 843
Rig Power (Idle): 188.48W
Rig Power (night time): 332.61W 567.49W
Rig Power (day time): 175.95W 419.92W
Rig MH/W (night time): 2.79 MH/W 1.71 MH/W
Rig MH/W (day time): 4.52 MH/W 2.01 MH/W


Day time GPU clock averages (using --auto-gpu so it varies):
GPU0: 434.62MHz
GPU1: 470.32MHz
GPU2: 501.34MHz
GPU3: 500.32MHz



The wattage rating is computed from Amps via the current probe of the UPS that provides power to the rig and I am assuming a PF of 1 as the PSU is power factor corrected. I'm still a bit skeptical of these numbers though as the efficiency seems to be implausibly high for the day time...

Edit: Revised numbers based on iMeter Solo load measurements at the wall with more accurate cgminer stats

[Syke]

Rig Power (night time): 332.61W
Rig Power (day time): 175.95W

Your night time power looks really good, but still plausible. But I can't imagine how you pulled off that day time power draw. What are the rest of your settings (volts, clocks, etc.)?

[Mousepotato]

I'm jealous of you guys getting 3.XX MH/w and more!  Dammit, you're going to make me go buy one of those power-sipping CPUs this weekend

[JinTu]

Rig Power (night time): 332.61W
Rig Power (day time): 175.95W

Your night time power looks really good, but still plausible. But I can't imagine how you pulled off that day time power draw. What are the rest of your settings (volts, clocks, etc.)?

Day time GPU clock averages (using --auto-gpu so it varies):
GPU0: 434.62MHz
GPU1: 470.32MHz
GPU2: 501.34MHz
GPU3: 500.32MHz

Volts: 1.00V
Mem Clock: 150MHz (OC BIOS reflashed)

AMD App: v2.5
AMD Catalyst: v11.11 (8.911)

Updated original post with this info.

[gyverlb]

Day time GPU clock averages (using --auto-gpu so it varies):
GPU0: 434.62MHz
GPU1: 470.32MHz
GPU2: 501.34MHz
GPU3: 500.32MHz

Volts: 1.00V
Mem Clock: 150MHz (OC BIOS reflashed)

AMD App: v2.4
AMD Catalyst:11.11

Updated original post with this info.

Even being optimistic I don't see how the rig could be bellow 300W with these frequencies and voltage. Something is definitely wrong here.

[DeathAndTaxes]

Even being optimistic I don't see how the rig could be bellow 300W with these frequencies and voltage. Something is definitely wrong here.

+1.  Something is wrong somewhere.  Even if AC/DC loss was 0 and MB/CPU was 50W.  We are talking <130W per 6990? 

[JinTu]

Even being optimistic I don't see how the rig could be bellow 300W with these frequencies and voltage. Something is definitely wrong here.

+1.  Something is wrong somewhere.  Even if AC/DC loss was 0 and MB/CPU was 50W.  We are talking <130W per 6990? 

Agreed. Even for me this didn't pass the smell test.

I'm looking to add an appliance power meter like this one to get real-time power stats (in Watts so no conversion required) for just the mining rig. Currently the UPS power stats are for the entire rack and I subtract out the current draw for the other components in the rack so the error margin is larger than need be.

[ewhenn]

Atom board with 1 16xPCIE, and 1 1xPCIE

PSU: Corsair CX 430 (~81% efficient)

GPU0: 5770 @ 850/1.000V
GPU1: 5770 @ 896/1.000V

Total power draw when mining: 156W
MHash: ~396

All parts were purchased used except fo the cx430 which was  a16.99 newegg special, full rig price: $151

MHash/watt: 2.53
MHash/$: 2.62

[P_Shep]

Linksys (cisco) e3000
Rosewill 7 port USB hub
Rosewill 650W PSU
2x BFL singles

155W
1650MH/s

10.6MH/W

Am I not welcome here anymore?

[Mousepotato]

Atom board with 1 16xPCIE, and 1 1xPCIE

PSU: Corsair CX 430 (~81% efficient)

GPU0: 5770 @ 850/1.000V
GPU1: 5770 @ 896/1.000V

Total power draw when mining: 156W
MHash: ~396

All parts were purchased used except fo the cx430 which was  a16.99 newegg special, full rig price: $151

MHash/watt: 2.53
MHash/$: 2.62

That's pretty damn good for a 5770 setup.

[TheHarbinger]

PSU : SeaSonic X-1250
MOBO : MSI 890FXA-GD70
CPU : AMD Sempron 145 Sargas

Running BAMT off a USB key...

5 x 5870s
1.000v 825/150 clocks
1875 Mh/s
615W average at the wall... (Varies day/night)

3.05 Mh/W   

But I do have a box fan pulling air across the cards and a wireless bridge, so there's another 30w maybe.  :/

[Gomeler]

Depends on the box fan. Mine pulls 70w on the low setting and 110w on the high setting. Best used to move air across 4-6 rigs to reduce the hit on efficiency.

[BinaryMage]

PSU: OCZ 1250W 80 Plus Gold
MOBO: Gigabyte GA-EP45-UD3P
GPU: 5 x 5870 (ref) @ 950/300, stock V
CPU: C2D e8500
Fans: 3

Power @ Wall: ~990W
MH/s: ~2150
MH/s per Watt:: ~2.2

Taking clocks down doesn't drop power consumption significantly...

Any idea how much undervolting would do?

[Isokivi]

PSU: OCZ 1250W 80 Plus Gold
MOBO: Gigabyte GA-EP45-UD3P
GPU: 5 x 5870 (ref) @ 950/300, stock V
CPU: C2D e8500
Fans: 3

Power @ Wall: ~990W
MH/s: ~2150
MH/s per Watt:: ~2.2

Taking clocks down doesn't drop power consumption significantly...

Any idea how much undervolting would do?

5870's react very well to undervolting, however finding the otimal settings can and will take long, each card is an individual. I for one am running two with the following settings (I trust these to be spot on with my miner/kernel and other mining conditions, including electricity price, difficulty, exchange rate...):
 #1 core 872Mhz, memory 200Mhz, core voltage 1.000 for 376Mhs
 #2 core 927Mhz, memory 200Mhz, core voltage 1.000 for 395Mhs

You can also go as low as 0.95v on 5870's but I havent started tweaking at that voltage yet, propably will when the block reward gets halved.. depending on the exchange rate and difficulty ofc.