Braiins OS & Braiins OS+ custom ASIC firmware: optimize performance & efficiency

[Bitcguru77]

Yes, you can just change the MAC address in uEnv.txt with notepad

Thank you for the info.

I'll be back with results later.

Plan is to replace Nov 17th Autotune FW with most recent Braiins-OS and find the sweet spot. In other words max hashing rate with Asic boost enabled and by setting frequency and voltage without increasing electric consumtion (too much any ways). At the moment averaging 13,6 TH with S9 13,5 TH january batch and will measure watts from the wall before and after.

BR.
Steff

[1714928384]

1714928384

[1714928384]

1714928384

[1714928384]

1714928384

[1714928384]

1714928384

[1714928384]

1714928384

[Apprentice]

Can it be installed on a T9?
Has anyone tried it?

[Bitcguru77]

So no luck this evening.  

I had my sd card with me and when I got to the miner it didn't fit. Seems like I need a Micro SD card that's usually found in mobile phones. If I got the time I'll go back tomorrow evening.

Off topic, but insted I put the Bitmain 23 oct LPM software in the miner for now and can confirm a drop from 1395W to 1220W with no noticable drop in hash rate. Temperatures also seems to be the same and fan speeds are lower than before (noticeably quieter). Maybe even a small increase in hash rate, but by tomorron night I'll have a more credible average. If I don't get the time to fix the Micro SD tomorrow at least I know if the decline in hash rate after time (reported in other threads) also applies to my miner as I probably won't get the time to try Braiins OS until next weekend.

BR.
Steff

[rockminer1]

Second try.

My case. I have usual Antminer S9 and I'd like to install braiins-os. It is working only from SD card now. I cannot to install it onto memory and cannot upgrade firmware via upgrade command

rockminer1: ^ if you are seeing that the firmware should be loaded, have you tried a web browser to http://192.168.0.33 ??

You are ssh'd to your miner with braiins OS .. you can see if the services are loaded by typing

Code:

ps | grep bin

tnx, I did

Code:

root@MINER:~# ps | grep bin
    1 root      1332 S    /sbin/procd
 1000 root       968 S    /sbin/ubusd
 1001 root       672 S    /sbin/askfirst /usr/libexec/login.sh
 1196 root      1012 S    /sbin/logd -S 64
 1205 root      1244 S    /sbin/rpcd
 1244 root      1520 S    /sbin/netifd
 1339 root      1332 S    /usr/sbin/uhttpd -f -h /www -r MINER -x /cgi-bin -u /ubus -t 60 -T 30 -k 20 -A 1 -n 3 -N 100 -R -p 0.0.0.0:80 -p [::]:80
 1400 root      1040 S <  /usr/sbin/ntpd -n -N -S /usr/sbin/ntpd-hotplug -p 0.lede.pool.ntp.org -p 1.lede.pool.ntp.org -p 2.lede.pool.ntp.org -p 3.lede.pool.ntp.org
 1415 root     25988 S <  /usr/bin/cgminer --default-config /etc/cgminer.conf
 1430 root      2196 S    {cgminer_monitor} /usr/bin/lua /usr/sbin/cgminer_monitor.lua
 1626 root       816 S    /usr/sbin/dropbear -F -P /var/run/dropbear.1.pid -p 22 -K 300
 1647 root       840 S    /usr/sbin/announce
14604 root       884 S    /usr/sbin/dropbear -F -P /var/run/dropbear.1.pid -p 22 -K 300
28084 root      1036 S    grep bin
root@MINER:~# 

It was unlucky installation too

Code:

root@MINER:~# opkg update
Downloading http://downloads.lede-project.org/releases/17.01.4/targets/zynq/dm1-g19/packages/Packages.gz
*** Failed to download the package list from http://downloads.lede-project.org/releases/17.01.4/targets/zynq/dm1-g19/packages/Packages.gz

Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/base/Packages.gz
Updated list of available packages in /var/opkg-lists/reboot_base
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/base/Packages.sig
Signature check passed.
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/luci/Packages.gz
Updated list of available packages in /var/opkg-lists/reboot_luci
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/luci/Packages.sig
Signature check passed.
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/miner/Packages.gz
*** Failed to download the package list from http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/miner/Packages.gz

Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/packages/Packages.gz
Updated list of available packages in /var/opkg-lists/reboot_packages
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/packages/Packages.sig
Signature check passed.
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/routing/Packages.gz
Updated list of available packages in /var/opkg-lists/reboot_routing
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/routing/Packages.sig
Signature check passed.
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/telephony/Packages.gz
Updated list of available packages in /var/opkg-lists/reboot_telephony
Downloading http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/telephony/Packages.sig
Signature check passed.
Collected errors:
 * opkg_download: Failed to download http://downloads.lede-project.org/releases/17.01.4/targets/zynq/dm1-g19/packages/Packages.gz, wget returned 8.
 * opkg_download: Failed to download http://downloads.lede-project.org/releases/17.01.4/packages/arm_cortex-a9_neon/miner/Packages.gz, wget returned 8.
root@MINER:~# opkg install firmware
Unknown package 'firmware'.
Collected errors:
 * opkg_install_cmd: Cannot install package firmware.

Please any ideas. I want it )))

[Breeze]

Would be nice to see HW% error count on the next release.
Keep up the nice work! 

[Shad0wSmurf]

Second try.

My case. I have usual Antminer S9 and I'd like to install braiins-os. It is working only from SD card now. I cannot to install it onto memory and cannot upgrade firmware via upgrade command

[...]

There's a command to remove and clean the os on their wiki. Do that. Maybe hard reset the s9 with the reset in, and powering it on. Then wait and see if it goes through the normal preheat test.

[evilwar]

cant get windows to format the image for booting it properly. Using Rufus 3.3 or Win32 Disk Imager 1.0 both load the sd card with the image file but it partitions the drive for 2 seperate drives. After changing the jumpers as per the guide and changing the mac address to match the one I was using nothing shows up on Angry IP scanner. The miner turns on with fans blazing and both led statuses stay lit up (Red & Amber)Windows 10 x64 Home Antminer S9 13.5TH model. Kinda stuck at this point.

[Bitcguru77]

cant get windows to format the image for booting it properly. Using Rufus 3.3 or Win32 Disk Imager 1.0 both load the sd card with the image file but it partitions the drive for 2 seperate drives. After changing the jumpers as per the guide and changing the mac address to match the one I was using nothing shows up on Angry IP scanner. The miner turns on with fans blazing and both led statuses stay lit up (Red & Amber)Windows 10 x64 Home Antminer S9 13.5TH model. Kinda stuck at this point.

Hi evilwar.

Don't know if I can help you, but as I also had som problems I'll try (discribed in earlier post).
- Download SD Memory Card Formatter from San Disk and format the SD card with the program.
- Extract the Braiins OS file package ones more on your harddrive.
- Start Win32DiskImager.
- Choose the SD.img from the folder you just extracted and the disk drive that shows you freshly formatted SD card.
- Press write.
- Change the mac addres in uEnv.txt on the on the SD card with notepad++ or similar.

Don't know if this does anything for you, but worked for me.

Regards.
Steff

So no luck this evening.  

I had my sd card with me and when I got to the miner it didn't fit. Seems like I need a Micro SD card that's usually found in mobile phones. If I got the time I'll go back tomorrow evening.

Off topic, but insted I put the Bitmain 23 oct LPM software in the miner for now and can confirm a drop from 1395W to 1220W with no noticable drop in hash rate. Temperatures also seems to be the same and fan speeds are lower than before (noticeably quieter). Maybe even a small increase in hash rate, but by tomorron night I'll have a more credible average. If I don't get the time to fix the Micro SD tomorrow at least I know if the decline in hash rate after time (reported in other threads) also applies to my miner as I probably won't get the time to try Braiins OS until next weekend.

BR.
Steff

Today I gave Braiins a try as hash rate in the LPM seemed to decline pretty much for me in 24 hours and the hash rate was also unstable in my opinion.

As there is more or less no profit for me mining with the S9 at the moment I don't really mind bricking it, so I played around a little with frequensies and voltages without really knowing what I did. I got awful results so i returned to square one.

As my chains seems to be pretty different (third one lagging really bad) I set the frequensies as Bitmains autotune did before, meaning 646, 642 and 606. Surprisingly the hash rate on each chain started to look normal.

Braiins suggested 9 V for chain one and two and 9.1 V for chain three. After this I rised each chain by 25 MHz and then once more, but the hash rate kept dropping.

So back to 646, 642 and 606. I then locked the voltages at 9, 9 and 9.1 and then rised the frequencies again on each chain first by 25 and then 25 once more. It started to look better.

I ended up with 9, 9.1 and 9.2 and frequensies 696, 692 and 656. Now chain one and two are almost the same, and three still about 400 GHs behind.

Before AsicBoost with Bitmain Autotune the S9 did avg. 13.6 THs at 1395 W. Now with Braiins, AsicBoost activated and with settings above it did 14.3 THs at 1350 W after running about 15 min. Probably needs a little more time to settle, but I had to go home by then.

First question about temperature. Before I had a maximum of 67 degrees celsius (chain 1). Now 83. However the fans seems to turn much slower than with Autotune, but I can't find any fan speed reading nor setting in Braiins. Are the fans running slower in Braiins and will they speed up automatically if the temperature goes higher? Can they be set manually by configuring some file on the SD card? 83 Is not too bad, but I think the temperatures would be pretty close to before if the fans also had the same speed as before.

Second question. I thought that higher frequency needs higher voltage, but Braiins recommends higher voltage the lower the frequency. How can this be?

Is 9, 9.1 and 9.2 too high to keep the voltage at and then still try to increase the frequensy by steps of 25 and see what happens? Do I dare to rise the voltage even higher?

So what to try to find the sweet spot (best THs/electric cost)? I need to sit down with the miner, with a six-pack, with lots of time and change a little at a time to see what happens. However if you have any thoughts about which way to go with voltages and frequesies they would be appreciated. If the S9 gets bricked it's all my own responsibility.

//Steff

[Enmadaio]

So my normal temps are between 75-80. I flashed this and everything shot up to 103-105. I tried taking off all boosts and click speeds and no luck. I just flashed back to stock and it's all fine. Any ideas? This was an s9i, and it also was only hashing like 13 while overheating.

[Ameador1]

Will this OS with LPM and/or the newest 1023 Bitmain firmware with LPM work to mine on NiceHash? Could one of you who have changed over to the LPM firmwares/OS's test this to see?

Thanks!

Just got a partial answer. NiceHash just made an announcement that they now support AsicBoost mining! Yeah, but, I'm still interested if the Braiins OS will mine with NiceHash and other pools - or this only for SlushPool?

[NotFuzzyWarm]

... or this only for SlushPool?

Does it say 'only Slushpool'? No.
Ergo, works with ANY pool that supports AB such as Kano.is, -ck's solo and his regular pool, and others.

[pipcrypdoh]

Has anyone had an issue with Braiins-OS on 14TH S9i's where they reboot every 30 minutes or so consistently.  I have units with both sd cards and flashed nand, no change.  Not sure if this is isolated to the S9i batch I have, but would like to hear if anyone has had this issue.

*EDIT* Figured this out, you have to create a password within the braiins web interface.  I guess that's okay, but.

[Bitcguru77]

So I continued trying to figure it out today and ended up with the following settings that I will keep for a while:

Frequency: 725 on all three chains.
Voltage: Chain 1: 8.9 V, chain 2: 9.1 V and chain 3: 9.2 V.
AsicBoost: enabled

Result:
All three chains hashing within 100 GH/s at around 5000 GH/s. So a total of just over 15 TH/s pulling 1440 W out of the wall. Yet to see if it's stable or not.

So after doing my calculations Bitmains LPM would be more efficient IF it would stick at 13,6 TH/s pulling 1220 W out of the wall.

Efficiency
Bitmain autotune: 13,6 THs at 1395 W = 9,75 GHs/W
Bitmain LPM: 13,6 THs at 1220 W = 11,06 GHs/W but drops after a few hours
Braiins OS: 15 THs at 1440 W = 10,42 GHs/W (Obviously I probably could get at least the same results as the LPM with Braiins, but where's the fun and the balls in that? )

Please correct me if I got something wrong.

This would have been impossible with the other dev free solutions in this forum as my hashing chains are so different. With the possibility to set the frequency and voltage separately for each chain I think I got a pretty good result with a miner not that good.

So a big thanks to the Braiins-team.  

BR.
Steff

[ruber1]

So testing on 1 machine for about 6 days now. Good start . On day 2, overclocked. Some Trial and error, got a steady 14.2  at 80 - 82 degrees (13.5 machine. usually ran about 13.6) .
Ran great and steady at 14.2 or so for several days, until last night. Then dropped to 13.4 -13.5. Machine re-starts daily. Tried resetting stock speeds and now machine is running at 12.9 or so. Will continue to play with....

[Apprentice]

on a 13.5 machine am getting 14.6T at 695 with old firmware and 14.2 with the new firmware with AB enabled. can pump it to 15T but one of the chains is not going beyond 4.7T

chains names are not consistent. on pages it's 0-5 other 1-6

[pipcrypdoh]

I've restored to factory on a couple with no issues.  Having one with this problem now though.  

Code:

Rebooting.................................................Traceback (most recent call last):
  File "restore2factory.py", line 177, in <module>
    main(args)
  File "restore2factory.py", line 149, in main
    restore_from_dir(args, args.backup)
  File "restore2factory.py", line 123, in restore_from_dir
    wait_for_reboot(args.hostname, REBOOT_DELAY)
  File "restore2factory.py", line 106, in wait_for_reboot
    while not wait_net_service(hostname, 22, 1):
  File "restore2factory.py", line 83, in wait_net_service
    s.connect((server, port))
OSError: [Errno 113] No route to host

This kills the web interface but SSH is still fine.  I try to run the script again and get this.

Code:

Connecting to remote host...
/home/pi/braiins-os_am1-s9_2018-10-24-0-9e5687a2/factory_transition/.env/lib/python3.5/site-packages/paramiko/client.py:822: UserWarning: Unknown ssh-rsa host key for 192.168.65.27: b'd63e5c5cef9b62df2dcfdaaef4f28b8f'
  key.get_name(), hostname, hexlify(key.get_fingerprint())
Traceback (most recent call last):
  File "restore2factory.py", line 177, in <module>
    main(args)
  File "restore2factory.py", line 149, in main
    restore_from_dir(args, args.backup)
  File "restore2factory.py", line 121, in restore_from_dir
    ssh.run('fw_setenv', RECOVERY_MTDPARTS[:-1], '"{}"'.format(mtdparts_params))
  File "/home/pi/braiins-os_am1-s9_2018-10-24-0-9e5687a2/factory_transition/ssh.py", line 228, in run
    self._check_exit_status(cmd, stdout, stderr)
  File "/home/pi/braiins-os_am1-s9_2018-10-24-0-9e5687a2/factory_transition/ssh.py", line 159, in _check_exit_status
    raise CalledProcessError(returncode, cmd, stdout, stderr)
subprocess.CalledProcessError: Command 'fw_setenv recovery_mtdparts "mtdparts=pl35x-nand:32m(BOOT.bin-env-dts-kernel),144m(angstram-rootfs),80m(upgrade-rootfs)"' returned non-zero exit status 1

I can reboot and everything is as it was for Braiins, but can't get back to original firmware.
Not sure where to put bug reports for this but ayyyy.  Here ya go.

[Shad0wSmurf]

I've restored to factory on a couple with no issues.  Having one with this problem now though.  

[...]

This kills the web interface but SSH is still fine.  I try to run the script again and get this.

[...]

I can reboot and everything is as it was for Braiins, but can't get back to original firmware.
Not sure where to put bug reports for this but ayyyy.  Here ya go.

Set the Mac address on your router to a static address for that miner. And try again

[pipcrypdoh]

Set the Mac address on your router to a static address for that miner. And try again

Static leases are maintained by a windows DHCP server in my environment.  Thank you for the suggestion, I'll run with it in a couple ways and see if it alleviates the issue.

[AnonyMustard]

I tried reflashing an S9 using upgrade2bos.py. Seemed to work flawlessly - It completed by printing "Content of U-Boot configuration:", states that "Upgrade was successful!", and "Rebooting..." After quite some time I am still unable to access it via browser or SSH. It shows up on my router and is definitely drawing power.

What am I missing here? Is there a specific port?

Thanks.

[Apprentice]

is the firmware setting the "Recommended voltage" automatically?

or the voltage should be manually changed as recommended?