Topic: 896 mh/s firmware release - Butterfly Labs - page 8. (Read 18548 times)
120W on 1 single? That don't seem right.
It'd be closer to 5 BTC, and no, it's not worth it for me to pay for that myself.
http://www.newegg.com/Product/Product.aspx?Item=N82E16882715001
OH, a Kill A Watt (I also would have understood kill a watt, kill-a-watt, Kill-A-Watt, etc, killawatt just looked like a typo for kilowatt).http://www.newegg.com/Product/Product.aspx?Item=N82E16882715001
Not worth it I guess, there must be too many hardware variations for it to be useful. P_Shep's unit with the 872 bitstream gets 849 mhash/s at 77 watts, but mine with the same bitstream gets 866 at 120 watts.
Wow, that is quite a difference!For reference, I'm getting 850 MH/s on both of the miners that arrived using the 864 bitstream.
Just doing some math...
With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
You have a little miscalculationWith 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
813 / 72 = 11.29 MH/s/W
849 / 77 = 11.03 MH/s/W
Also interesting :
849MH/s - 813MH/s @ 77W - 72W = 36MH/s @ 5W = 7.2 MH/s/W
The 7.2 MH/s/W for the extra power is probably still profitable for you, if not, you better go back to 813 MH/s.
P.S.
I like math
These measurements should really be done at the board level, measuring DC current and voltage. Ideally it would be done on a large sample pool, too, but at least looking at the power dependence on hashrate for a single board under the same conditions (ambient temp, airflow, etc) should be enough to infer how much your own profit will increase or decrease for your slightly different board.
Just doing some math...
With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
You have a little miscalculationWith 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
813 / 72 = 11.29 MH/s/W
849 / 77 = 11.03 MH/s/W
Also interesting :
849MH/s - 813MH/s @ 77W - 72W = 36MH/s @ 5W = 7.2 MH/s/W
The 7.2 MH/s/W for the extra power is probably still profitable for you, if not, you better go back to 813 MH/s.
P.S.
I like math
It'd be closer to 5 BTC, and no, it's not worth it for me to pay for that myself.
http://www.newegg.com/Product/Product.aspx?Item=N82E16882715001
OH, a Kill A Watt (I also would have understood kill a watt, kill-a-watt, Kill-A-Watt, etc, killawatt just looked like a typo for kilowatt). http://www.newegg.com/Product/Product.aspx?Item=N82E16882715001
I'll do it if someone pays for a killawatt for me.
I don't have a single, so I don't really care, but are you suggesting that it's not worth the fraction of a BTC someone could send you to pay for a kilowatt to know what is most efficient for your own single?http://www.newegg.com/Product/Product.aspx?Item=N82E16882715001
I'll do it if someone pays for a killawatt for me.
I don't have a single, so I don't really care, but are you suggesting that it's not worth the fraction of a BTC someone could send you to pay for a kilowatt to know what is most efficient for your own single?
I'd do it... if I could be bothered.
Just doing some math...
With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
You have a little miscalculationWith 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
813 / 72 = 11.29 MH/s/W
849 / 77 = 11.03 MH/s/W
Also interesting :
849MH/s - 813MH/s @ 77W - 72W = 36MH/s @ 5W = 7.2 MH/s/W
The 7.2 MH/s/W for the extra power is probably still profitable for you, if not, you better go back to 813 MH/s.
P.S.
I like math
Just doing some math...
With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
You have a little miscalculationWith 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
813 / 72 = 11.29 MH/s/W
849 / 77 = 11.03 MH/s/W
Also interesting :
849MH/s - 813MH/s @ 77W - 72W = 36MH/s @ 5W = 7.2 MH/s/W
The 7.2 MH/s/W for the extra power is probably still profitable for you, if not, you better go back to 813 MH/s.
P.S.
I like math
Just doing some math...
With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
You have a little miscalculationWith 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
813 / 72 = 11.29 MH/s/W
849 / 77 = 11.03 MH/s/W
Also interesting :
849MH/s - 813MH/s @ 77W - 72W = 36MH/s @ 5W = 7.2 MH/s/W
The 7.2 MH/s/W for the extra power is probably still profitable for you, if not, you better go back to 813 MH/s.
P.S.
I like math
Just doing some math...
With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
Good to know... thanks for calculating that out. With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
Just doing some math...
With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
With 4 @ 832, I was drawing 300W at the wall including the router (12W), so (300-12)/4 = 72W per BFL.
Currently have 2 @ 872, and 2 @ 832, drawing 310W at the wall including the router (the other 2 won't run at 872).
So with 72W per 832 unit, the 872 units must be using ((310-12) - (2*72))/2 = 77W.
Comparing hash rates:
832: 813MH/s @72W = 11.26 MH/s/W
872: 849MH/s @77W = 11.02 MH/s/W
So SLIGHTLY less efficient.
...
Somewhat related: I've sent an email to BFL to ask for the details of all the commands and specifically also mentioned the ZAX command so I can see if I can write what should be reasonably simple code to do the bitstream update.
If they don't reply (or wont give that particular command details) then I guess we'll just have to look at the EasyMiner code since that is easily reversed according to someone I had a chat with who had already done it but had difficulty understanding the result (I also worked out that it's just .NET 2 also with VB in it so it shouldn't be hard to determine the actual code - but determining the protocol may or may not be easy)
... and the reply was "No" Somewhat related: I've sent an email to BFL to ask for the details of all the commands and specifically also mentioned the ZAX command so I can see if I can write what should be reasonably simple code to do the bitstream update.
If they don't reply (or wont give that particular command details) then I guess we'll just have to look at the EasyMiner code since that is easily reversed according to someone I had a chat with who had already done it but had difficulty understanding the result (I also worked out that it's just .NET 2 also with VB in it so it shouldn't be hard to determine the actual code - but determining the protocol may or may not be easy)
OK, now I see the topic has changed. Theres a 872 MH firmware out now too. Maybe that will fix my problem (hehe, I don't think so).
Is the order of the singles listed in cgminer the same order that I put them in my config file? com3, com4, com7,com13 etc?
I don't want to reflash the wrong ones.
Oh, forgot to mention, if you have the API enabled, the devdetails command will tell you which BLFn is which COMn device in windowsIs the order of the singles listed in cgminer the same order that I put them in my config file? com3, com4, com7,com13 etc?
I don't want to reflash the wrong ones.
(and of course which /dev/ttyUSBn device in linux)
Another curious thing I found this morning, One of my 872 units seems to have stalled.
The 5s hash rate was reporting as 857, the ave as 408 and U as 6.92 (this is after running for 12 hours). Disabling, re-enabling the unit did nothing. It appears that some point during the night it got stuck, I guess just reporting busy or something(?), rather than dieing. Re-starting CGminer got it working again.
Def think the cgminer code needs reviewing. I may have a look and try and get my head around it.
The 5s hash rate was reporting as 857, the ave as 408 and U as 6.92 (this is after running for 12 hours). Disabling, re-enabling the unit did nothing. It appears that some point during the night it got stuck, I guess just reporting busy or something(?), rather than dieing. Re-starting CGminer got it working again.
Def think the cgminer code needs reviewing. I may have a look and try and get my head around it.
So this is curious...
The first 2 BFLs I got had a HS on the bottom of the PCB, and had temps of 54 - 58. The 2nd 2 had a HS+fan on the bottom of the PCB and temps of 47 - 50.
I assumed that the lower temp units would allow the higher rate firmware to work, turned out not to be the case and they would throttle, even with the 864 firmware.
I uploaded the 872 firmware on the first 2 units with the higher temps, and they're working perfectly.
I'm thinking Kano is on to something and the temp reported is not related to the temp the unit throttles.
The first 2 BFLs I got had a HS on the bottom of the PCB, and had temps of 54 - 58. The 2nd 2 had a HS+fan on the bottom of the PCB and temps of 47 - 50.
I assumed that the lower temp units would allow the higher rate firmware to work, turned out not to be the case and they would throttle, even with the 864 firmware.
I uploaded the 872 firmware on the first 2 units with the higher temps, and they're working perfectly.
I'm thinking Kano is on to something and the temp reported is not related to the temp the unit throttles.
I have lowered the temp on mine by replacing those cheap spring loaded screws that hold down the heat singe with 1/4 inch (size 4) bolt and screws. I was having trouble with throttling even before the fireware release. Everything works fine now.
yeah, I think it's a cgminer issue.
One of my units, the 47 one, keeps bombing out. Just stops hashing. CGminer says it's disabled, re-enabling it doesn't work. Re-starting CGminer does (for a while).
Gonna try the 872, just for the hell of it.
Edit:
OK the other one I updated just bombed out too :/
Edit2:
After observing it for a while it would seem it bombs out when it throttles. Gonna put them back to 832.
Gonna try the 872, just for the hell of it.
Edit:
OK the other one I updated just bombed out too :/
Edit2:
After observing it for a while it would seem it bombs out when it throttles. Gonna put them back to 832.
I had exactly the same problem on one of my units (it was at the stock firmware). The symptom was that, only under linux, it would die when it throttled. I sent it back to BFL but they never found a root cause as far as I know. Under Windows, it was stable (it just slowed down when it throttled instead of dying).
Jump to: