ethOS - Ethereum Mining Platform

[wolfen]

Claymore works great with Awesome Miner as a monitor.
Shows start hashing speed at max much faster than the ethos interface.
I use both. Ethos rocks. So does Awesome Miner Monitor.

[Duncanopolus]

Nvidia support provided. My congratulations! But there is a problem in Claimore-zcash:
.....AMD OpenCL platform not found
No AMD OPENCLGPUs found, exit....

optiminer-zcash: [error] Found no matching devices! Quitting!

Please, help Guys!

[Jonesy]

Nvidia support provided. My congratulations! But there is a problem in Claimore-zcash:
.....AMD OpenCL platform not found
No AMD OPENCLGPUs found, exit....

optiminer-zcash: [error] Found no matching devices! Quitting!

Please, help Guys!

Even though I havent tried EthOS yet, I think u will find some mining software on it only supports AMD cards, try use EWBF for nvidia cards.

[alien2108]

NVIDIA support was added so I switches to ethOS. I now have 6 ASUS 1070GTX cards working on Z270 without issues, but NVIDIA support is buggy and I hope someone from ethos is reading this:

Scripts for setting mem clock and gpu clock are written in a way where baseline Mhz is read and difference is calculated. But there is a problem with gpu clock!

My cards without overclock run on ~1560Mhz as reported by "nvidia-smi -q -d CLOCK", but when system starts somehow this clock gets bumped to MAX of ~1920Mhz for a brief moment, but long enough that ethOS writes this WRONG values to /var/run/ethos/defaultcore.file. Now the problem is, that you can't set clocks anymore. For example if I write "cor 1650" this will result in scripts setting totally wrong clocks as they calculate the difference between defaultcore and config core, which would wrongly result in 1650-1920 = -270 which is out of range and wrong anyhow. The correct result would be "1650-1560 = 90". This is all to the fact that nvidia-seettings use OFFSET for setting clocks and not raw value....

So my suggestion is to get the default clock somewhere else....ethos-oc-nvidia get's it with "/opt/ethos/sbin/ethos-readdata core" which now correctly reads the values:

1582 1582 1556 1594 1569 1544

but somehow during bootup this values are written in  /var/run/ethos/defaultcore.file:

1911 1911 1949 1936 1936 1949

My suspicion is that you should have longer delay before reading default clocks, since some other process influences them in between....Or as I said, read theme somewhere else....

I can help you out with testing as I am many decades Linux user if you need any help....

PS: Also to notice that memclock overclocking is also not straightforward....as +100 does not mean +100Mhz...rougly +50....so you have to use approximately double value...but this is nvidia-settings issue not ethOS scripts....

[Rostik]

Hello. I have some questions.
1) Are you able to add new miner? CCminer?

2) And what about some notifications when miner falls or rig not online? For example to telegram.

3) What about option to automatic reboot when miner or card fall? 

[buffy007]

3) set autoreboot true in conf

[Rostik]

3) set autoreboot true in conf

just autoreboot true without any numbers?

[bumblebee24]

# autoreboot:           set to a number (autoreboot 3), allow rigs to autoreboot if they soft crash (max of 3 reboots, run clear-thermals to reset counter)

Looks like you can set the autoreboot at either 1,2,3 or false; having it set to false it same as not having it in the config.

ex;

autoreboot false
autoreboot true
autoreboot 3 

if set at true I'm assuming it will just keep rebooting everytime

[kr3st]

Now that 1.2.4 is out and it uses the new AMD drivers does that mean I can use my VEGA 64 cards to mine with on ethos?

[opelgtman]

I wish there was a trial of this OS.  I'd hate to spend $39 and not be able to get it work.

[bumblebee24]

Now that 1.2.4 is out and it uses the new AMD drivers does that mean I can use my VEGA 64 cards to mine with on ethos?

The update was mainly for RX cards by looking at the change log. It did fix the DAG issue at least on my rigs.

Not sure if anyone is running VEGA 64 on it, but, when rx560's first came out I ran them on ethos before official update and they worked. Ethos just didn't recognize their names. Kept saying unknown for card names but they hashed correctly.

[alred911]

hi everyone,

i just bought ETHOS 1.2.4

I have a bunch of 1070 but they are not recongnize by Ethos,
they are detected (show in boot)  that's it.

They clearly say they support NVDIA now but maybe im missing somehing ?

[chanberg]

1.2.4 was a RX dag fix, but for some reason it also made my 1070's faster... was hashing at 30.2 now its going at 32.23... they probably fixed the tables which dictate the overclock speed.

[duniek]

I wish there was a trial of this OS.  I'd hate to spend $39 and not be able to get it work.

if you make a little digging you can download it free,
I did that and after week of testing i bought it

1.2.4 was a RX dag fix, but for some reason it also made my 1070's faster... was hashing at 30.2 now its going at 32.23... they probably fixed the tables which dictate the overclock speed.

config?

globalcore ??
globalmem ??
powertune ??



btw i just checked gtx1070

windows7 mem +725 (4500) 32mhash @ claymore
ethOS mem (4800) 31mhash @ethermine/claymore - same speeds

core 1550/1600 (-400) or 1700(-200) doesnt change anything

[chilliefly]

I see the following has been added to latest release. Has anyone had any success with RX voltage. Ive tried but it seems the powerstage/powertune trumps it?!?!? Can anyone tell me how the voltage works with the RX Cards  (i use 570,580 cards). 

-Added ability to set voltage for RX series GPUs.

[chanberg]

config?

globalcore ??
globalmem ??
powertune ??



btw i just checked gtx1070

windows7 mem +725 (4500) 32mhash @ claymore
ethOS mem (4800) 31mhash @ethermine/claymore - same speeds

core 1550/1600 (-400) or 1700(-200) doesnt change anything

cor 1675
mem 5275
ptune 2

its stable. been running like that for 52 hours

[duniek]

thanks

do you know power usage ? (i mean real - from socket)

i tried ethOS and i just get very high power consumption in comparasion to windows mining

[bileq]

thanks

do you know power usage ? (i mean real - from socket)

i tried ethOS and i just get very high power consumption in comparasion to windows mining

you can use globalpowertune option or set each rig with pwr switch

globalpowertune 90 (means performance level state  max 90 watt for nvidia gpu series)
globalpowertune 2 (means dpmstate for amd gpu series)
or
for a1b2c3 rig name
you can use

pwr a1b2c3 90 90 90 90 90 90 (6 nvidia gpu max 90 power consumption)
pwr a1b2c3 2 2 2 2 2 2 2 (6 amd gpu dpmstate)

[Porticus]

I have updated and clean ethos, but I can't get it to mine with 1050ti graphics, supposedly the minefield and the reading of the drivers should be automatic, but it tells me that it doesn't detect available drivers..

[Porticus]

[miner started] /home/ethos # show miner                                                                                                 
tailing /var/run/miner.output (use ctrl+c to exit)                                                                                                                           
terminate called after throwing an instance of 'std::runtime_error'                                                                                                         
  what():  No CUDA driver found                                                                                                                                             
terminate called after throwing an instance of 'std::runtime_error'                                                                                                         
  what():  No CUDA driver found                                                                                                                                             
terminate called after throwing an instance of 'std::runtime_error'                                                                                                         
  what():  No CUDA driver found                                                                                                                                             
terminate called after throwing an instance of 'std::runtime_error'                                                                                                         
  what():  No CUDA driver found                                                                                                                                             
terminate called after throwing an instance of 'std::runtime_error'                                                                                                         
  what():  No CUDA driver found                                                                                                                                             
terminate called after throwing an instance of 'std::runtime_error'                                                                                                         
  what():  No CUDA driver found                                                                                                                                             


any idea?