ANTMINER S5: 1155GH(+OverClock Potential), In Stock $0.319/GH & 0.51W/GH

[toptekk]

Can you load balance on these again?  My brother wants to buy half a miner with me, and we want to make sure we can split the hashrate 50/50 to both of our pools, also any other variation split 40/60 and/or 30/70?

Thanks,

DebitMe

you can but you have to edit some files on the  miner with SSH or WinSCP i haven't figured out which ones hold the setting it's not like the S3 or lower you could set the /usr/bin/ CGMIner file to --balance on a line  Polly still can but haven't had a need to balance farm yet these things are nice just farming straight up bit coins.


This use to be the old way


cd /etc/init.d
 vi cgminer 


which is no longer there my bad above was wrong for the S3 and below.

[dogie]

Can you load balance on these again?  My brother wants to buy half a miner with me, and we want to make sure we can split the hashrate 50/50 to both of our pools, also any other variation split 40/60 and/or 30/70?

Thanks,

DebitMe

you can but you have to edit some files on the  miner with SHH or WinSCP i haven't figured out which ones hold the setting it's not like the S3 or lower you could set the /usr/bin/ CGMIner file to --balance on a line  Polly still can but haven't had a need to balance farm yet these things are nice just farming straight up bit coins.

If you encounter the problem that changes are wiped on a reset [once you find the config file via SSH], we should be able to get around that. If you flash a 4GB+ SD card with the this image [old firmware, will need updating], you should be able to run the S5 off it. SD card takes priority over the flash memory (when you /etc/init.d/pgnand.sh) and so acts as primary OS. Changes to that should be permanent. Can also be used if you network brick / get corruption on the flash OS which reset doesn't fix.

[I say 'should' a lot as I've not tested it, but its from an internal Bitmain report.]

Edit as of 17th Feb: Don't do this, it'll only cause you trouble / likely to brick.

[toptekk]

Call me worried but, I think I have found a setting that gets me a decent hash rate with temps below 60c. 300 with the current set up that I have will yield 59 and 57 degrees on the blades but within the first hour or so I get 1 hardware error and over the course of a day and a half (36hours) I have 12 hardware errors. Now, I am powering this miner with a high quality 80 gold 650watt power supply.

first of all, are 12 hardware errors over the course of 36 hours ok for the miner? does that seem a bit high? I suppose the cooler I get the less likely the hardware errors will be produced. I am getting two new noctua nf-f12 fans, one at 3000rpm (push) and one at 2000rpm (pull) to replace my current setup as the wife has finally said the miner it too loud and producing too much heat.

I am asking about hardware errors again because I noticed that when I first set the miner to 300 it performed at just over 1Th/s and now it is working at just under 1Th/s. Do hardware errors effect hash rate?

The number of HW errors is not important. It's the percentage. HW errors are a GOOD thing for ASICs. It means your ASIC is being pushed to its limits for speed. Generally there is a tradeoff early on. Going from 0 to 1% HW errors results in a 5-15% increase in speed. and that takes about a hour to even out or become stable.

Future increases generally have smaller gains. Good ASIC firmware will automatically push the clocks up on ASICs to maximize the effective hash rate. but if it goes over a certain %  that is bad, for the S5 i think the safe % is 0.388 or some thing like that . mine never go over 0.02 that's  for days not mins or hours and i OC with a 750 gold psu or higher on all my S5, I'm not pushing OCing just saying even at 300 i would use a 750 or higher .


I would buy a 750 Watt psu gold . that gives you a lot of head room, but that's me. That  may even be why the high temps it's  not supplying enough power to keep it cool, when it needs it. I may be wrong. i  can't say I all ways over kill on PSUs  not to much but just enough that it's not a waste .

How did you calculate the number? 0.388% of what?

Thanks!

I would do what they are asking you to do above there might be some thing wrong.  


I would think for the S5 it would be about the same, for any ASIC miner that advice some one gave me a while back i couldn't get it in my head why there was any to start with till i reed that , i only give it out when i see some one is having a hard time trying to understand just how HW errors work .it all depends on the miner and speed . some speeds may give better HW errors while other speeds may give worse HW errors and others will tell you that to. and 0.388 is the danger sigh to watch for any thing below that you more then likely are fine and make sure there is no xxxx  were there should be all 00000 across form each blade . I would show a image but have yet to figure out how to on the forums, I'm  not one to ask , I'm hard headed on stuff like that and only ask if I need it.

[toptekk]

Can you load balance on these again?  My brother wants to buy half a miner with me, and we want to make sure we can split the hashrate 50/50 to both of our pools, also any other variation split 40/60 and/or 30/70?

Thanks,

DebitMe

you can but you have to edit some files on the  miner with SHH or WinSCP i haven't figured out which ones hold the setting it's not like the S3 or lower you could set the /usr/bin/ CGMIner file to --balance on a line  Polly still can but haven't had a need to balance farm yet these things are nice just farming straight up bit coins.

If you encounter the problem that changes are wiped on a reset [once you find the config file via SSH], we should be able to get around that. If you flash a 4GB+ SD card with the this image [old firmware, will need updating], you should be able to run the S5 off it. SD card takes priority over the flash memory (when you /etc/init.d/pgnand.sh) and so acts as primary OS. Changes to that should be permanent. Can also be used if you network brick / get corruption on the flash OS which reset doesn't fix.

[I say 'should' a lot as I've not tested it, but its from an internal Bitmain report.]

Thanks man I was wondering how to use the SD part of the S5 or how it worked, i was about to look it up .

[toptekk]

Can you load balance on these again?  My brother wants to buy half a miner with me, and we want to make sure we can split the hashrate 50/50 to both of our pools, also any other variation split 40/60 and/or 30/70?

Thanks,

DebitMe

found it if this helps any

cd /etc/init.d
 vi cgminer.sh


 PARAMS="--bitmain-dev /dev/bitmain-asic --bitmain-options 115200:32:8:$timeout:$chip_value:$freq_value:0725 --bitmain-checkn2diff --bitmain-hwerror --version-file /usr/bin/compile_time --queue $queue_value --load-balance"


I added load balance as a example .


Cya

[dogie]

Can you load balance on these again?  My brother wants to buy half a miner with me, and we want to make sure we can split the hashrate 50/50 to both of our pools, also any other variation split 40/60 and/or 30/70?

Thanks,

DebitMe

you can but you have to edit some files on the  miner with SHH or WinSCP i haven't figured out which ones hold the setting it's not like the S3 or lower you could set the /usr/bin/ CGMIner file to --balance on a line  Polly still can but haven't had a need to balance farm yet these things are nice just farming straight up bit coins.

If you encounter the problem that changes are wiped on a reset [once you find the config file via SSH], we should be able to get around that. If you flash a 4GB+ SD card with the this image [old firmware, will need updating], you should be able to run the S5 off it. SD card takes priority over the flash memory (when you /etc/init.d/pgnand.sh) and so acts as primary OS. Changes to that should be permanent. Can also be used if you network brick / get corruption on the flash OS which reset doesn't fix.

[I say 'should' a lot as I've not tested it, but its from an internal Bitmain report.]

Thanks man I was wondering how to use the SD part of the S5 or how it worked, i was about to look it up .

I'll write it up properly for if others need it again once I've tested it myself.

[netcoingeek]

Call me worried but, I think I have found a setting that gets me a decent hash rate with temps below 60c. 300 with the current set up that I have will yield 59 and 57 degrees on the blades but within the first hour or so I get 1 hardware error and over the course of a day and a half (36hours) I have 12 hardware errors. Now, I am powering this miner with a high quality 80 gold 650watt power supply.

first of all, are 12 hardware errors over the course of 36 hours ok for the miner? does that seem a bit high? I suppose the cooler I get the less likely the hardware errors will be produced. I am getting two new noctua nf-f12 fans, one at 3000rpm (push) and one at 2000rpm (pull) to replace my current setup as the wife has finally said the miner it too loud and producing too much heat.

I am asking about hardware errors again because I noticed that when I first set the miner to 300 it performed at just over 1Th/s and now it is working at just under 1Th/s. Do hardware errors effect hash rate?

The number of HW errors is not important. It's the percentage. HW errors are a GOOD thing for ASICs. It means your ASIC is being pushed to its limits for speed. Generally there is a tradeoff early on. Going from 0 to 1% HW errors results in a 5-15% increase in speed. and that takes about a hour to even out or become stable.

Future increases generally have smaller gains. Good ASIC firmware will automatically push the clocks up on ASICs to maximize the effective hash rate. but if it goes over a certain %  that is bad, for the S5 i think the safe % is 0.388 or some thing like that . mine never go over 0.02 that's  for days not mins or hours and i OC with a 750 gold psu or higher on all my S5, I'm not pushing OCing just saying even at 300 i would use a 750 or higher .


I would buy a 750 Watt psu gold . that gives you a lot of head room, but that's me. That  may even be why the high temps it's  not supplying enough power to keep it cool, when it needs it. I may be wrong . i  can't say I all ways over kill on  PSUs  not to much but just enough that it's not a waste .

I thought about a higher PSU but I couldn't afford the higher psu and well, i didn't plan on overclocking the miner either so I just went with a supply that meet the miners needs, 4 VGA and at least 600 watts. So how to you calculate hardware errors in percentage? so I can make sure my miner does not go over a certain percentage.

[toptekk]

Call me worried but, I think I have found a setting that gets me a decent hash rate with temps below 60c. 300 with the current set up that I have will yield 59 and 57 degrees on the blades but within the first hour or so I get 1 hardware error and over the course of a day and a half (36hours) I have 12 hardware errors. Now, I am powering this miner with a high quality 80 gold 650watt power supply.

first of all, are 12 hardware errors over the course of 36 hours ok for the miner? does that seem a bit high? I suppose the cooler I get the less likely the hardware errors will be produced. I am getting two new noctua nf-f12 fans, one at 3000rpm (push) and one at 2000rpm (pull) to replace my current setup as the wife has finally said the miner it too loud and producing too much heat.

I am asking about hardware errors again because I noticed that when I first set the miner to 300 it performed at just over 1Th/s and now it is working at just under 1Th/s. Do hardware errors effect hash rate?

The number of HW errors is not important. It's the percentage. HW errors are a GOOD thing for ASICs. It means your ASIC is being pushed to its limits for speed. Generally there is a tradeoff early on. Going from 0 to 1% HW errors results in a 5-15% increase in speed. and that takes about a hour to even out or become stable.

Future increases generally have smaller gains. Good ASIC firmware will automatically push the clocks up on ASICs to maximize the effective hash rate. but if it goes over a certain %  that is bad, for the S5 i think the safe % is 0.388 or some thing like that . mine never go over 0.02 that's  for days not mins or hours and i OC with a 750 gold psu or higher on all my S5, I'm not pushing OCing just saying even at 300 i would use a 750 or higher .


I would buy a 750 Watt psu gold . that gives you a lot of head room, but that's me. That  may even be why the high temps it's  not supplying enough power to keep it cool, when it needs it. I may be wrong . i  can't say I all ways over kill on  PSUs  not to much but just enough that it's not a waste .

I thought about a higher PSU but I couldn't afford the higher psu and well, i didn't plan on overclocking the miner either so I just went with a supply that meet the miners needs, 4 VGA and at least 600 watts. So how to you calculate hardware errors in percentage? so I can make sure my miner does not go over a certain percentage.

you really can't because diff speeds do diff HW errors or can but i don't know how to other then watch it, why not set it to default 350 and not 300 your not  OCing then actually getting better hash rates and that's the bitmain allowed safe setting . it should do no higher then 590 at the wall . it seems your about right in temps for those two fans from what Ive seen from others using the same fans. you want at least 141 CFM per fan to sty around 55 to 57 and lower with those fans because of the low CFM it might go up to 60 or better, I have noticed from other posts . btw if that's all its doing your fine you can't stop the HW errors but can control how cool it stays by making sure it stays in a cool room all year long. either use the very noicely default fan or replace it with two id say 151 CFM or the ones you have but the temps will go up close to 60 or over with them, 80 C the miner will shut down and 78 or some thing the miner may start beep as a warning ..  stay at 350 defaults then watch the HW error after a hour it should settle down and  be ok  unless something is wrong with the miner .


I would send a PST to the  players who asked you to if you haven't, to make sure nothing is wrong with the miner, if there is they will replace it. After 36 hours imo  all you get is 12 errors  not bad. I get after two days any were from 2000 to 4000 errors but my % stays low in a very safe range and that's all i worry about . . I just explained that the HW errors are good as long as they don't go over the limit and won't if you stay at default speed or lower . as for what % to what speed before they go up I would stay at default  setting it should be fine . Ive notice the only time ever HW errors get bad is if it's OC to much , it's not cool enough or some thing is wrong .


I hope this helps and sorry i don't know the  exact math to calculate hardware errors per hash  or can  tell you how much hash you get before it happen, i think your asking other then 0 to 1 Hw errors incenses speed by 5-15%. maybe that should be rewritten some how in CGminer so others don't feel alarmed to much, to a diff name .



Cya and GL

[vabchgent]

Anyone Know if the miners can be SSHed into and be rebooted or Shutdown command issued. If so how going in under root at the command line is not allowing me to issue a shutdown.

[dunand]

Anyone Know if the miners can be SSHed into and be rebooted or Shutdown command issued. If so how going in under root at the command line is not allowing me to issue a shutdown.

did you tried?
sudo reboot

[toptekk]

Anyone Know if the miners can be SSHed into and be rebooted or Shutdown command issued. If so how going in under root at the command line is not allowing me to issue a shutdown.

cd /etc/init.d
/etc/init.d/cgminer.sh start
/etc/init.d/cgminer.sh stop



log into the S5 with ssh type reboot at root works every time for me .


make sure to use


User root
password admin

[MyRig]

To reboot you should be able to just issue the command.

Code:

reboot

If for some reason that doesn't work you can try a force reboot.

Code:

echo b > /proc/sysrq-trigger

To restart cgminer you can issue the command.

Code:

/etc/init.d/cgminer.sh restart

[Rabinovitch]

If you didn't know, there is no need to send your bricked S5 (if it simply became unbootable while the HW is OK...) to Bitmain for reflashing. Put a firmware for S4 on SD card, insert it into empty slot on controller board and power up your s5. Update FW to latest version for S5 then and enjoy.

Thanx to smit1237 for figuring it out.  


p.s. S5 could be cooler and quieter in radiators inside it would be like those in Avalon4: with more thick radiator fins and more space between them (since Bitmain decided to use aluminium radiators).

[coinits]

This setup is returning about 0.014 - 0.015 BTC per day on Nicehash.

Thoughts? Should I be concerned about the HW Errors?

[dogie]

This setup is returning about 0.014 - 0.015 BTC per day on Nicehash.

Thoughts? Should I be concerned about the HW Errors?

No, still statistically insignificant.

[coinits]

This setup is returning about 0.014 - 0.015 BTC per day on Nicehash.

Thoughts? Should I be concerned about the HW Errors?

No, still statistically insignificant.

Thanks for that but could you elaborate please.

[toptekk]

This setup is returning about 0.014 - 0.015 BTC per day on Nicehash.

Thoughts? Should I be concerned about the HW Errors?

four days up time if it's showing right at 400 Frequency is awesome for HW errors. for westhash or nicehash.



don't go by the HW errors as a number they add up over time, %'s under Diff1# DiffA# is what to worry about i was told any thing above 0.388 is bad or can be, your fine if it's been run in for 4 days.


ASIC status
oooooooo oooooooo oooooooo oooooo
oooooooo oooooooo oooooooo oooooo

after fours days still at 0 .



to bad those blocks are not BTC blocks but alt coins .

[dogie]

This setup is returning about 0.014 - 0.015 BTC per day on Nicehash.

Thoughts? Should I be concerned about the HW Errors?

No, still statistically insignificant.

Thanks for that but could you elaborate please.

0.06% hardware errors is the equivalent of your 1239GH miner hashing at 1238.3GH. It makes no difference what so ever and is perfectly normal.

[FlensGold]

Does anyone have an idea why I am only getting a bit more than half of the hashing power?
Frequency is set to 200 (standard), PSU is Corsair CX750M which is working fine for all other S5 miners I have.

[Rabinovitch]

At frequency 400 I have about 1.3 Th. Linear dependence.