Pages:
Author

Topic: [GUIDE] BitFury Miner Support/Tuning - page 40. (Read 148085 times)

legendary
Activity: 1513
Merit: 1040
September 09, 2013, 03:38:25 AM
#46
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, 12:35:34 AM
#45
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, 09:32:23 PM
#44
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, 05:45:45 PM
#43
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, 01:25:44 PM
#42
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, 12:14:53 PM
#41
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.
sr. member
Activity: 434
Merit: 265
September 08, 2013, 08:53:28 AM
#40
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
hero member
Activity: 574
Merit: 501
September 08, 2013, 06:23:59 AM
#39
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: 1379
Merit: 1003
nec sine labore
September 08, 2013, 03:22:51 AM
#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 )


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, 04:25:18 PM
#37
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, 04:14:45 PM
#36
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, 03:50:34 PM
#35
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, 03:33:33 PM
#34
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, 01:17:34 PM
#33
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, 01:11:25 PM
#32
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 ...
hero member
Activity: 574
Merit: 501
September 07, 2013, 12:45:12 PM
#31
Dear Lord that thing is really small on a version-1 M-board (not the mod being discussed - the pencil mod is done on each H-board).  

My eyes aren't what they used to be, but are R01F and R02F REALLY size 0603 on the m-board??  If they are, how is it even possible to manually place 0402 parts Huh
legendary
Activity: 1946
Merit: 1006
Bitcoin / Crypto mining Hardware.
September 07, 2013, 12:33:54 PM
#30
we need a pictured guide for dummies ... so we can break our boards trying ...
courtesy c-scape
sr. member
Activity: 540
Merit: 250
BestRate.org
September 07, 2013, 11:05:34 AM
#29
is everyone here from Germany - Hamburg that can make the resistor mod for the H Card, i have two of them ?
sr. member
Activity: 404
Merit: 250
September 07, 2013, 04:22:05 AM
#28
has someone tried to use a wlan usb stick? how can i do this?
sr. member
Activity: 434
Merit: 265
September 07, 2013, 04:20:22 AM
#27
we need a pictured guide for dummies ... so we can break our boards trying ...
Pages:
Jump to: