Pages:
Author

Topic: [GUIDE] BitFury Miner Support/Tuning - page 40. (Read 148018 times)

legendary
Activity: 1974
Merit: 1003
September 09, 2013, 05:36:52 AM
#53
can anyone help me trough team viewer ?
i just got the starter kit and i got very bad results ... 7 GH with both H cards ... Huh
legendary
Activity: 1513
Merit: 1040
September 09, 2013, 05:32:32 AM
#52
yep, information (chip-wise) like in cgminer would be nice.
sr. member
Activity: 434
Merit: 265
September 09, 2013, 05:25:49 AM
#51
we need to get some more data stuff into chainminer ...

.. but we will have to wait until bfsb commits there new chainminer version to github ...
.. one thing that is really important is the "putstat log" which shows the communication with the pools, we need it to be displayed in the web gui ... and also some long time data collection ...
legendary
Activity: 1513
Merit: 1040
September 09, 2013, 04:38:25 AM
#50
Does overvolting/clocking make sense due to total error rate (is it rising a lot)?
I'm thinking about changing R01F to 2,2k.
legendary
Activity: 1946
Merit: 1006
Bitcoin / Crypto mining Hardware.
September 09, 2013, 01:35:34 AM
#49
Is there any guide/manual to make chainminer display more values?? I was hoping it displayed a full set of values: accepted, rejected, difficulty, etc, etc, just like cgminer or bfgminer but chainminer is extremely shy and doesn't provide nearly the same feedback when running manually.

Similary, is there any guide on how to set up/optimize stratum? I heard someone changed the real target parameter or something like that to achieve better results.


you can check difficulty by
>sudo su
>screen -ls
>screen -r
It should show the difficulty reported by stratum proxy
sr. member
Activity: 408
Merit: 250
September 08, 2013, 10:32:23 PM
#48
Is there any guide/manual to make chainminer display more values?? I was hoping it displayed a full set of values: accepted, rejected, difficulty, etc, etc, just like cgminer or bfgminer but chainminer is extremely shy and doesn't provide nearly the same feedback when running manually.

Similary, is there any guide on how to set up/optimize stratum? I heard someone changed the real target parameter or something like that to achieve better results.
full member
Activity: 141
Merit: 100
September 08, 2013, 06:45:45 PM
#47
When I refresh or close and open the Bitfury Admin (in Chrome), it doesn't fill the fields of Pool Configuration. Mining still seems to continue on the pools I configured earlier. On the first start it did show the pre-programmed mining adresses (donating towards the Bitcoin Foundation)

Who else does(n't) experience this minor inconvenience? I like to know if it is my browser (setting) or the admin page itself and if someone has a fix for this.
hero member
Activity: 910
Merit: 1000
Items flashing here available at btctrinkets.com
September 08, 2013, 02:25:44 PM
#46
Heres how you use the make-shift stat collector with the tools currently available, please note that this is a work in progress. But asfar as I can see it works:

ssh in to your bitfury-pi
make a folder /home/pi/logs
in the folder: nano logger.sh
paste the following code in, modify the board count and desired data collection time and save the file.
Code:
#!/bin/bash
# Bitfury chainminer logfile consolidation script
# Jlsminingcorp and Isokivi, September 2013
# Version 1.3

# User configurable variables
# $logfile is the path to the bitfury chainminer log file
# $output is the path to the board-data output file that you would like to write to
# $outputdir is the directory to store output in
# $logtime is the time (in minutes) to collect data for
# $numboards is the number of H-boards your the miner
logfile="/run/shm/.stat.log"
output="./boards.log"
outputdir="."
logtime="20"
numboards="2"

# Timestamps
datestamp=$(ls --full-time "$logfile" | awk '{print $6}')
timestamp=$(ls --full-time "$logfile" | awk '{print $7}' | awk -F"." '{print $1}')

# If log file or output files don't exist then take appropriate action
if [ ! -e "$logfile" ]; then
    echo "$(date)"" : ""Logfile does not exist in the specified location"
    echo "$(date)"" : ""Logfile does not exist in the specified location" >> "$output"
    exit 1
fi

if [ ! -e "$output" ]; then
    echo "Date,Time,Board Position,Speed,Noncerate [GH/s],Hashrate [GH/s],Good,Errors,SPI-Errors,Miso-Errors" > "$output"
fi

for (( i=1; i<="$numboards"; i++)); do
    chipout="$outputdir""/chips_board_""$i"
    if [ ! -e "$chipout" ]; then
        echo "Chip stats for board: ""$i" > "$chipout"
        echo "Date,Time,Chip,ProgParams,Speed,Noncerate,Hashrate,Nonces/round,False nonce,SPIerr,MISOerr,Jobs/5min (hash rate),ChipID,CoresOK" >> "$chipout"
    fi
done

echo "Starting to log data"
echo "Time to collect data for: ""$logtime"" minutes"

# During the data collection period (set by $logtime) parse data from the logfile to the output files
let countdown="$logtime"*"60"
while [ "$countdown" -ge "0" ]; do

    # If timestamp in the log file is the same as the timestamp on the last entry in the output file then sleep for a while
    # Should make sure that we're somewhere in the middle of the 5 minute chainminer logging period
    # Could use "while" here, but risk getting stuck in a never-ending loop if log file is not being updated
    prevtimestamp=$(tail -n 1 "$output" | awk -F"," '{print $2}')
    if [ "$timestamp" == "$prevtimestamp" ]; then
        echo "Chainminer log file not yet updated. Will now sleep for a short while."
        echo "Chainminer log file not yet updated. Will now sleep for a short while." >> "$output"
        sleep 60
        timestamp=$(ls --full-time "$logfile" | awk '{print $7}' | awk -F"." '{print $1}')
    fi

    # Strip board data out of the chainminer log file and copy to the output file
    IFS=$'\r\n' datalines=($(grep -A "$numboards" record "$logfile" | tail -n "$numboards" ))
    for i in "${datalines[@]}"; do
        echo -ne "$datestamp","$timestamp", >> "$output"
        echo "$i" | tr ":" " " | awk '{$1=$1}1' OFS="," >> "$output"
    done

    # Strip chip data out of the chainminer log file and copy to chip output files (one for each H-board)
    for (( i=1; i<="$numboards"; i++)); do
        chipout="$outputdir""/chips_board_""$i"
        let startline="$i"*"16"-"15"
        let endline="$i"*"16"
        while read line; do
            echo -ne "$datestamp","$timestamp", >> "$chipout"
            echo "$line" | awk '{for (i=1; i<=12; i++) printf("%s%s", $(i), i<12 ? OFS="," : "\n"); }'  >> "$chipout"
        done < <(awk 'NR==v1,NR==v2' v1="${startline}" v2="${endline}" "$logfile")
    done

    echo "Time remaining: ""$countdown"" seconds"
    if [ "$countdown" -gt "0" ]; then
        sleep 300
    fi
    let countdown="$countdown"-"300"
    timestamp=$(ls --full-time "$logfile" | awk '{print $7}' | awk -F"." '{print $1}')

done

echo "Finished logging data"
exit 0
chmod +x logger.sh
./logger.sh

Once it has finished: less chips_board_1
Copy the contents and paste it to http://anduck.net/bfsb/#
Press go.
Repeat above three for the next board.
Enjoy!
If you feel like tipping the people who contributed jlsminingcorp (the script) 1JNeDQsT6Jh9XGqhcQPHZkpKzA9YASvNTT and Anduck (the web-app) 1Anduck6bsXBXH7fPHzePJSXdC9AEsRmt4 are the proper recipients.
hero member
Activity: 574
Merit: 501
September 08, 2013, 01:14:53 PM
#45
Look at the U numbers printed on the board - U40 is the first chip.  They're numbered from bottom left to top left, down the next row, up the next, etc.
legendary
Activity: 2702
Merit: 1468
September 08, 2013, 12:01:19 PM
#44
Anybody knows what "speed->0" at the end of the .stat.log line actually mean?

2       AIfDSo  55      0.243   2.019   17      126     0       0       191     [0:1]   726     0 1 2 2 1 0 0 0 2 0 3 0 4 1 1 0         9 8 7 7 8 9 9 9 7 9 6 8 5 8 8 9    speed->0

10      AIfDSo  55      0.272   1.924   19      110     0       0       182     [0:9]   722     0 2 1 2 2 2 0 2 0 1 1 1 2 1 1 1         8 6 7 6 6 6 8 7 8 7 7 7 6 7 7 7    speed->0

#2 and #10 chips are not doing what they suppose to.  I'm not sure if it is assembly or my configuration.

EDIT:

After a while the rates drop completely:

2       AiFDso  55      0.014   0.550   1       35      0       0       52      [0:1]   723     0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0         2 2 2 2 3 3 3 3 2 2 1 2 2 2 2 2
10      AiFDso  55      0.072   0.528   5       34      0       0       50      [0:9]   715     1 1 0 0 0 1 0 1 0 1 0 0 0 0 0 0         2 2 3 2 2 1 2 1 2 1 2 2 3 3 3 3




thats the tuning mod that was used for this cycle ...

each cycle it will use a autotuning mod depending on the errors u got ....

as long you got no errors it will speed up the clock +1 ...

Code:
13. Autoclock
    - shut down = Chip will be shutdown and unavailable (Aifdso)
    - speed->0 = If still more errors than correct nonces -> shut down
    - speed up = If you have 0 errors
    - speed down = If you have >5 errors clock down

Thanks, that makes sense now.

I turned off auto (aIfDSo) at 52, but the nonce rates are 0.35 GH on those chips.  Do you know how are chips numbered on the board, i.e which are #2 and #10 (for visual inspection).
sr. member
Activity: 434
Merit: 265
September 08, 2013, 09:53:28 AM
#43
Anybody knows what "speed->0" at the end of the .stat.log line actually mean?

2       AIfDSo  55      0.243   2.019   17      126     0       0       191     [0:1]   726     0 1 2 2 1 0 0 0 2 0 3 0 4 1 1 0         9 8 7 7 8 9 9 9 7 9 6 8 5 8 8 9    speed->0

10      AIfDSo  55      0.272   1.924   19      110     0       0       182     [0:9]   722     0 2 1 2 2 2 0 2 0 1 1 1 2 1 1 1         8 6 7 6 6 6 8 7 8 7 7 7 6 7 7 7    speed->0

#2 and #10 chips are not doing what they suppose to.  I'm not sure if it is assembly or my configuration.

EDIT:

After a while the rates drop completely:

2       AiFDso  55      0.014   0.550   1       35      0       0       52      [0:1]   723     0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0         2 2 2 2 3 3 3 3 2 2 1 2 2 2 2 2
10      AiFDso  55      0.072   0.528   5       34      0       0       50      [0:9]   715     1 1 0 0 0 1 0 1 0 1 0 0 0 0 0 0         2 2 3 2 2 1 2 1 2 1 2 2 3 3 3 3




thats the tuning mod that was used for this cycle ...

each cycle it will use a autotuning mod depending on the errors u got ....

as long you got no errors it will speed up the clock +1 ...

Code:
13. Autoclock
    - shut down = Chip will be shutdown and unavailable (Aifdso)
    - speed->0 = If still more errors than correct nonces -> shut down
    - speed up = If you have 0 errors
    - speed down = If you have >5 errors clock down
legendary
Activity: 2702
Merit: 1468
September 08, 2013, 09:45:20 AM
#42
Anybody knows what "speed->0" at the end of the .stat.log line actually mean?

2       AIfDSo  55      0.243   2.019   17      126     0       0       191     [0:1]   726     0 1 2 2 1 0 0 0 2 0 3 0 4 1 1 0         9 8 7 7 8 9 9 9 7 9 6 8 5 8 8 9    speed->0

10      AIfDSo  55      0.272   1.924   19      110     0       0       182     [0:9]   722     0 2 1 2 2 2 0 2 0 1 1 1 2 1 1 1         8 6 7 6 6 6 8 7 8 7 7 7 6 7 7 7    speed->0

#2 and #10 chips are not doing what they suppose to.  I'm not sure if it is assembly or my configuration.

EDIT:

After a while the rates drop completely:

2       AiFDso  55      0.014   0.550   1       35      0       0       52      [0:1]   723     0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0         2 2 2 2 3 3 3 3 2 2 1 2 2 2 2 2
10      AiFDso  55      0.072   0.528   5       34      0       0       50      [0:9]   715     1 1 0 0 0 1 0 1 0 1 0 0 0 0 0 0         2 2 3 2 2 1 2 1 2 1 2 2 3 3 3 3


hero member
Activity: 574
Merit: 501
September 08, 2013, 07:23:59 AM
#41
Code:
watch -n 305 tail -3 /run/shm/.stat.log

Mine should run on any version of Linux or UNIX released since 1975, to include V2R5, Solaris, BSD, HPUX, OS/390, z/OS, Mac OS, and more.  
Yours, not so much....   Grin
legendary
Activity: 1378
Merit: 1003
nec sine labore
September 08, 2013, 04:22:51 AM
#40
or leave an ssh window open and  run this:

while (true) do tail -3 /run/shm/.stat.log; sleep 305; done


( not run/smh )


or

Code:
watch -n 305 tail -3 /run/shm/.stat.log

Smiley

spiccioli
hero member
Activity: 910
Merit: 1000
Items flashing here available at btctrinkets.com
September 07, 2013, 05:25:18 PM
#39
or leave an ssh window open and  run this:

while (true) do tail -3 /run/shm/.stat.log; sleep 305; done


( not run/smh )
corrected my post, thanks ...I keep making that same typo over and over again Sad
hero member
Activity: 574
Merit: 501
September 07, 2013, 05:14:45 PM
#38
or leave an ssh window open and  run this:

while (true) do tail -3 /run/shm/.stat.log; sleep 305; done


( not run/smh )
hero member
Activity: 910
Merit: 1000
Items flashing here available at btctrinkets.com
September 07, 2013, 04:50:34 PM
#37
How did you look for the error rates? can it be calculated from .stat.log?
open bitfury.ip/stat.json for constantly updating counts per board.
or ssh in and
Code:
less run/shm/.stat.log
This is a 5 min log that is rewritten over and over again
legendary
Activity: 1946
Merit: 1006
Bitcoin / Crypto mining Hardware.
September 07, 2013, 04:33:33 PM
#36
Any word on the error rate, I managed to get to 30-33Ghs for short periods and huge error rates, with 0 (zero) actual gains.

How did you look for the error rates? can it be calculated from .stat.log?
hero member
Activity: 910
Merit: 1000
Items flashing here available at btctrinkets.com
September 07, 2013, 02:17:34 PM
#35
Any word on the error rate, I managed to get to 30-33Ghs for short periods and huge error rates, with 0 (zero) actual gains.
sr. member
Activity: 434
Merit: 265
September 07, 2013, 02:11:25 PM
#34
i got those pictures sent from a user ...

to guide through the process of the pencil mod ....

***be aware that you loose warranty of the h-board in doing this ... and you can brick your board!!!***

details:

- temp +10-15°
- 1.6 times higher electricity needed
- up to 50% increased hashrate

Step 1: unplug your h-board ... and locate the R02F Resistor



Step 2: use ur multimeter on the resistor and verify that it shows 1.3k Ohm



Step 3: use a HB (medium grade) pencil (http://en.wikipedia.org/wiki/Pencil) and rub 2-3 times



Step 4: use ur multimeter on the resistor and check how much it was lowered in resistance ...as long as your higher then 1.2k restart step 3 & step 4 ...
goal resistance is 1.1k - 1.2k, if you get under 1.1k just swip a little away with your finger ...

Step 5: plug in again and start




have fun ...
Pages:
Jump to: