[GUIDE] BitFury Miner Support/Tuning

[darkfriend77]

So he's going to yank power connectors off - and risk damaging - multiple, multi-thousand dollar pieces of hardware just to try and save himself $90 on a PSU?  

He deserves whatever happens to his rigs...

Hmmm ... I don't know why you think I would deserve something ... but anyway ...

I'm working with 2 PSU right now until stability isn't an issue anymore and then move onto one PSU.

Thx for all the constructive comments especially yours Redacted.

Peace and happy mining.

[-Redacted-]

Just doesn't seem to be a good idea skimping on fairly inexpensive PSU's, or pulling out power cables to turn off an expensive rig - for the same reason you wouldn't expect a motherboard to last for very long if that was how you shut your computer down - by pulling power cables out of the connectors.

PSU's have bleeder circuitry to prevent big amp surges from taking place when you shut them off.  If you yank a connector out of the socket, or just pull the plug out of the wall, there are lots of electrons that suddenly find themselves needing to go home -RIGHT NOW- and they aren't too fussy about the path they take getting there.  They're perfectly willing to make themselves a new path of least resistance, and will happily travel through - and burn up - a part that wasn't designed to handle that kind of surge, if they can.

Just sayin'...

I only mention this having learned a great deal about electron return, flyback currents, and paths-of-least resistance working in an environment that sometimes required us to switch 100Kv @ 10K Amps on and off.  (Think LHC accelerator-beam superconducting magnet type power.)  People are better conductors than air-gaps, and electrons were more than happy to oblige if you did something incredibly stupid and offered them a shorter return path than the one you expected them to take through air gaps and enormous relays and pools of mercury. 

[Jumpsu]

I made step by step, but i cant acces, 6. How to start tuning chips manually?

What is wrong...? I get:
pi@bitfury ~ $ nano /run/shm/.stat.log
g++ -g -Wall -pthread -O2 -march=armv6 -mfloat-abi=hard -mfpu=vfp -Wall -I . -c spidevc.cpp -o spidevc.o
g++ -g -Wall -pthread -O2 -march=armv6 -mfloat-abi=hard -mfpu=vfp -Wall -I . -c miner.cpp -o miner.o
make -C json
make[1]: Entering directory `/opt/bitfury/chainminer/json'
g++ -pthread -I .. -O2 -g -c json_reader.cpp -o json_reader.o
g++ -pthread -I .. -O2 -g -c json_value.cpp -o json_value.o

Also:

pi@bitfury ~ $ nano /run/shm/.stat.log
g++ -g -Wall -pthread -O2 -march=armv6 -mfloat-abi=hard -mfpu=vfp -Wall -I . -c handylib.cpp -o handylib.o
In file included from /usr/include/c++/4.6/backward/strstream:52:0,
                 from handylib.cpp:5:
/usr/include/c++/4.6/backward/backward_warning.h:33:2: warning: #warning This file includes at least one deprecated or antiquated header which may be removed without further notice at a future date. Please use a non-deprecated interface with equivalent functionality instead. For a listing of replacement headers and interfaces, consult the file backward_warning.h. To disable this warning use -Wno-deprecated. [-Wcpp]
handylib.cpp: In member function ‘void handylib::HTTPServer::loadMultipart(std::istream&)’:
handylib.cpp:653:18: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
handylib.cpp: In member function ‘void handylib::Time::scache() const’:
handylib.cpp:2850:25: warning: comparison between signed and unsigned integer expressions [-Wsign-compare]
handylib.cpp: In function ‘void handylib::tm_loadtzi()’:
handylib.cpp:3048:13: warning: variable ‘tm’ set but not used [-Wunused-but-set-variable]

Also

pi@bitfury ~ $ nano /run/shm/.stat.log

g++ -g -Wall -pthread -O2 -march=armv6 -mfloat-abi=hard -mfpu=vfp -Wall -I . -c spidevc.cpp -o spidevc.o
g++ -g -Wall -pthread -O2 -march=armv6 -mfloat-abi=hard -mfpu=vfp -Wall -I . -c miner.cpp -o miner.o
make -C json
make[1]: Entering directory `/opt/bitfury/chainminer/json'
g++ -pthread -I .. -O2 -g -c json_reader.cpp -o json_reader.o

/rung++ -pthread -I .. -O2 -g -c json_value.cpp -o json_value.o
g++ -pthread -I .. -O2 -g -c json_writer.cpp -o json_writer.o
rm -f libjson.a
ar cruv libjson.a json_reader.o json_value.o json_writer.o
a - json_reader.o
a - json_value.o
a - json_writer.o
make[1]: Leaving directory `/opt/bitfury/chainminer/json'
g++ -g -Wall -pthread -O2 -march=armv6 -mfloat-abi=hard -mfpu=vfp -Wall -I . jobconnect.o handylib.o spidevc.o miner.o json/libjson.a -o miner
pi@bitfury /opt/bitfury/chainminer $ sudo killall screen
pi@bitfury /opt/bitfury/chainminer $ sudo /etc/rc.local
miner: no process found
pi@bitfury /opt/bitfury/chainminer $ pi@bitfury ~ $ nano /run/shm/.stat.log
-bash: pi@bitfury: command not found
pi@bitfury /opt/bitfury/chainminer $ pi@bitfury ~ $ nano /run/shm/.stat.log
-bash: pi@bitfury: command not found
pi@bitfury /opt/bitfury/chainminer $
pi@bitfury /opt/bitfury/chainminer $
pi@bitfury /opt/bitfury/chainminer $ /run

[Keefe]

Why are you compiling chainminer? If all you're trying to do is change clock speeds, just edit /opt/bitfury/best.cnf and reboot. If that file doesn't exist yet, copy /run/shm/.stat.log to /opt/bitfury/best.cnf first.

[Jumpsu]

Why are you compiling chainminer? If all you're trying to do is change clock speeds, just edit /opt/bitfury/best.cnf and reboot. If that file doesn't exist yet, copy /run/shm/.stat.log to /opt/bitfury/best.cnf first.

1. How i copy /run/shm/.stat.log to /opt/bitfury/best.cnf ?
2. How i get /run/shm/.stat.log ?

[tom99]

Why are you compiling chainminer? If all you're trying to do is change clock speeds, just edit /opt/bitfury/best.cnf and reboot. If that file doesn't exist yet, copy /run/shm/.stat.log to /opt/bitfury/best.cnf first.

1. How i copy /run/shm/.stat.log to /opt/bitfury/best.cnf ?
2. How i get /run/shm/.stat.log ?

use cp  /run/shm/.stat.log  /opt/bitfury/best.cnf

you can read more about cp command http://www.computerhope.com/unix/ucp.htm

[Keefe]

Why are you compiling chainminer? If all you're trying to do is change clock speeds, just edit /opt/bitfury/best.cnf and reboot. If that file doesn't exist yet, copy /run/shm/.stat.log to /opt/bitfury/best.cnf first.

1. How i copy /run/shm/.stat.log to /opt/bitfury/best.cnf ?
2. How i get /run/shm/.stat.log ?

1. cp /run/shm/.stat.log /opt/bitfury/best.cnf
2. If /run/shm/.stat.log doesn't exist, let the miner run for at least 5 minutes and it should be there.

[Meizirkki]

Even without overclocking, a full rig pulls something like 25A from the 12V input. You'll need a much heftier switch than that.

10A 250V switch should do fine. Low voltage DC is much easier on the switch.

I've used a 10A 250V switch on my electric kick scooter for years. Peak current >100A and it's still working fine after probably 2000 on/off cycles.

[goxed]

I've some questions ... maybee some one has a good solution ...

I'm using a multi rail atx psu ... 1200 Watt ... at which 2 miners are connected.

So if I want to do some overclocking at the moment I've to shut down both miners ...

What solution do I have to only shut down the miner Im modding ....

1) just pull out the pcie ... one by one on the miner? dangerous? also is reinserting to have it on again ... a problem when connecting one pcie after the other when psu is on?
2) is there a switch for a pcie cable? ...
3) a second psu ... yep i got one but i would be happy to only use ... one ...

thanx or your help

It may trigger the TVS diode on your raspberryPi and destroy it. I would suggest turning off the PSU instead of yanking out PCIE power cables.

[CMaxx22]

Another question:
What means if one chip shows aifdso 240?
Like the first one
http://s020.radikal.ru/i716/1311/ea/46685811bb64.jpg

And it seems that autotune is not working, on the screenshot I manually worked with 2,5,12 chips. The first one is setted up to AIfDSo 55 but automatically falls down to aifdso 240. The second chip was turned off by me because of very low hashrate (200Mhs).
Can you tell something about the first chip? And maybe you know what should I do for 2,5,12 chips to increase their hashrate?
Thank you!

ps: overclocking to 0.95v and oil cooling change nothing...

[-Redacted-]

A lowercase 'a' starting the string means you have turned off autotune for that chips and manually set the values...

[CMaxx22]

But why the other chips with capital "A" is not changed their frequency from 55? Day after day frequency is not changing...

[Keefe]

The column after the [4:x] shows how many of the 756 cores have not produced good results. It looks like chips 5 and 12 are half (or more) dead, so you'll need to leave autotune off and just accept that the noncerate on those will be less than half their hashrate. You can try different speeds to see what produces the best noncerate. You'll probably want to clock them as high as the rest, and just expect 50%+ error rate. Chainminer autotune can't handle chips with alot of dead cores and assumes they need to be downclocked or shut off. After running for long enough to get most of the bad-core counts near 0, look at /tmp/.core.log and see what 5 and 12 look like. It'll probably show that one side or the other is dead. Here's one of my half-dead chips:

Code:

pi@bitfury ~ $ grep ^41 /tmp/.core.log
41        35  36  30  22  31  31  35  35  33  39  21  31  34   0   0   0   0   0   0   0   0
41        31  35  36  38  29  18  42  27  28  26  25  31  25   0   0   0   0   0   0   0   0
41        34  34  36  34  25  28  27  29  26  31  28  26  33   0   0   0   0   0   0   0   0
41        32  31  29  40  33  29  28  26  35  40  31  35  42   0   0   0   0   0   0   0   0
41        19  23  27  29  40  31  40  18  32  22  35  30  38   0   0   0   0   0   0   0   0
41        31  18  26  34  21  25  31  38  39  24  25  34  39   0   0   0   0   0   0   0   0
41        30  24  35  34  32  19  33  26  36  38  30  34  34   0   0   0   0   0   0   0   0
41        43  28  26  36  24  33  36  26  33  44  23  29  40   0   0   0   0   0   0   0   0
41        31  37  32  24  28  31  39  31  22  42  28  37  30   0   0   0   0   0   0   0   0
41        36  20  25  35  28  29  30  25  27  25  35  29  33   0   0   0   0   0   0   0   0
41        28  25  36  38  24  33  32  27  36  25  28  34  34   0   0   0   0   0   0   0   0
41        37  41  36  40  37  24  27  20  39  32  30  34  26   0   0   0   0   0   0   0   0
41        32  26  34  28  32  32  26  33  31  36  35  31  23   0   0   0   0   0   0   0   0
41        27  31  33  26  24  37  30  32  29  28  26  32  31   0   0   0   0   0   0   0   0
41        30  23  25  30  25  23  30  24  30  19  26  39  30   0   0   0   0   0   0   0   0
41        29  38  39  33  30  32  32  37  27  25  41  29  32   0   0   0   0   0   0   0   0
41        29  31  33  30  30  32  39  37  25  35  19  30  51   0   0   0   0   0   0   0   0
41        25  20  22  33  29  39  31  29  30  36  36  34  38   0   0   0   0   0   0   0   0
41        19  30  33  37  23  26  29  37  34  28  35  27  24   0   0   0   0   0   0   0   0
41        41  23  40  27  32  33  28  36  26  28  45  32  30   0   0   0   0   0   0   0   0
41        33  36  29  34  23  36  31  25  27  25  30  37  20   0   0   0   0   0   0   0   0
41        24  33  26  38  23  29  23  30  28  41  36  35  23   0   0   0   0   0   0   0   0
41        35  31  34  34  28  29  38  18  27  33  31  37  37   0   0   0   0   0   0   0   0
41        36  41  30  20  30  35  45  29  37  38  25  26  44   0   0   0   0   0   0   0   0
41        43  33  33  27  41  19  27  33  25  30  23  36  31   0   0   0   0   0   0   0   0
41        26  26  32  30  30  25  39  33  32  32  35  29  37   0   0   0   0   0   0   0   0
41        45  29  34  38  36  25  29  37  34  37  36  26  38   0   0   0   0   0   0   0   0
41        30  43  34  25  35  29  28  30  33  27  27  39  28   0   0   0   0   0   0   0   0
41        32  27  39  36  38  28  34  44  31  32  37  30  35   0   0   0   0   0   0   0   0
41        28  30  37  35  39  16  39  30  25  27  30  46  28   0   0   0   0   0   0   0   0
41        35  27  28  30  39  35  19  23  33  39  35  43  30   0   0   0   0   0   0   0   0
41        30  31  29  34  28  37  33  22  34  43  37  32  45   0   0   0   0   0   0   0   0
41        30  27  39  23  25  24  39  38  28  28  28  29  30   0   0   0   0   0   0   0   0
41        27  28  26  29  37   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
41        27  25  21  32  31   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
41        35  31  40  35  22   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
41      312

[Keefe]

But why the other chips with capital "A" is not changed their frequency from 55? Day after day frequency is not changing...

Are you still using the old version of chainminer that shipped with the v1 M-boards? It either doesn't change the clocks, or it doesn't display it right. I upgraded to the latest version on github, and now autotune actually works (except for half-dead chips). But make sure to edit the version number before compiling if you have a v1 M-board!

Update for v1 M board owners. If you want to use the latest build of chainminer with your version 1 M-board please do the following

a) Download chainminer master on your raspberry PI from github and unzip it

Code:

https://github.com/bfsb/chainminer/archive/master.zip

https://github.com/bfsb/chainminer/archive/master.zip

b) change line 6 in miner.h

Code:

  1 #ifndef MINER_H
  2 #define MINER_H
  3 #include "handylib.h"
  4 
  5 #define NDEBUG
  6 #define VERSION 1

c) make

Code:

cd chainminer-master
make
cd ..

d) move /opt/bitfury/chainminer

Code:

mv /opt/bitfury/chainminer /opt/bitfury/chainminer_v1_original

e) copy chainminer-master directory to /opt/bitfury/chainminer

Code:

cp -r chainminer-master /opt/bitfury/chainminer
 

f) mine with latest chainminer goodness  

[Jumpsu]

I have problem, one my h-card doesn't work.

Its not hot, no hashing, 20W, it worked 12h ago very well (28Gh/s), it seems m-board not even get information from h-board.
I tried change new ssd card also.

What is wronng, how i can fix that problem?

http://img62.imageshack.us/img62/4086/v80m.jpg

Uploaded with ImageShack.us

[GandalfG]

I have problem, one my h-card doesn't work.

Its not hot, no hashing, 20W, it worked 12h ago very well (28Gh/s), it seems m-board not even get information from h-board.
I tried change new ssd card also.

What is wronng, how i can fix that problem?

Can you post log from miner page ?

[Meizirkki]

I have problem, one my h-card doesn't work.

Its not hot, no hashing, 20W, it worked 12h ago very well (28Gh/s), it seems m-board not even get information from h-board.
I tried change new ssd card also.

What is wronng, how i can fix that problem?



Uploaded with ImageShack.us

heatsinks should be on the other side of the board. Those components you circled should be 0ohm so that's not the problem.

[Jumpsu]

yeah, i know it... But i measured other card and third, capacitor was like 0.7M ohm...?

[Keefe]

A better test would be to power the card and measure voltage at the top of the Pulse inductor. If you get ~0.75v there (or whatever it's adjusted to), the power circuit (including that capacitor) is probably fine. I only say "probably" because I just fixed a board that was missing a cap on the back side. Before replacing that tiny back-side cap, the voltage at the Pulse top was correct, but the card wouldn't hash stably.

[klondike_bar]

Another question:
What means if one chip shows aifdso 240?
Like the first one


And it seems that autotune is not working, on the screenshot I manually worked with 2,5,12 chips. The first one is setted up to AIfDSo 55 but automatically falls down to aifdso 240. The second chip was turned off by me because of very low hashrate (200Mhs).
Can you tell something about the first chip? And maybe you know what should I do for 2,5,12 chips to increase their hashrate?
Thank you!

ps: overclocking to 0.95v and oil cooling change nothing...

change the first to either autotune 'A' and/or change the value to 53. it shouldnt be 240 and is probably caused by a mis-match the last time the file was modified (mine showed 545 once but worked once returned to a usable number). check your spacing

turn chip 2 on. use autotune or start at 53.

tune chips 5 and 12 to either 53 or autotune. 50 is too low and may be your issue.

finally, im surprised you could get 0.95v even with liquid cooling - pretty impressive. scale back to 0.90-0.92v as its a safer level and watch your autotune. ideally your chips should be 53-55 (most of mine run 54 when my system is stable near its cooling capacity)