[Mining OS] SimpleMining.net - Manage Your GPU farm the easy way! (30 days free)

[Tummler]

Hi,
Is there any update from today because from 1 hour my rigs are super unstable restarting a milion times , and different GPUs are hanging non-stop.
They were super stable and i doubt its a hardware problem on both of them at the same time.
Its for sure the OS.

Same problem here. My rigs were rock solid before this morning, often going weeks without rebooting. Now, my rigs with RX 470 cards are very unstable. They will run for a bit, and then one card will experience a brief jump up in processing speed and then crash. It almost seems like the problem is occurring when a card switches over to mine for Claymore.

I have one rig with RX 580 cards, and it does not seem to be affected.

At least in my case, I think it was a problem on Nanopool's end. I switched back to Dwarfpool, and my problems with unstable rigs subsided.

From Nanopool's DAG epoch clock, I'm guessing it reset sometime last night. Perhaps it is a problem related to that.

EDIT: I spoke too soon. Immediately after I clicked on the post button, I went back and checked my rigs, and one just crashed after one of the cards dropped to 0 mh/s.

[JohnJamesRambo13]

I have a few questions.

1. Does anyone know why SmOS doesn't seem to be able to overclock the RX 460?  No matter what values I put in it won't overclock or underclock it.  For the RX 480 and RX 580 I am able to do it just fine.  It seems to be rig specific.  It will work in a rig with one card but in mobos with multiple cards it won't overclock and just says "Invalid OC values: core1850 memory: , skipping OC" and defaults to 1750.  I know it is stable up until 1900 or so in Windows.

2. Is it normal that the Powerstage setting doesn't seem to do anything?  I can put it at 0 and the rigs don't really seem to notice a difference.  Does it work?  How much power can you save using it?  When I measured it with a Kill-A-Watt I didn't really notice much difference.  When I lower powerstage with an rx480 rig it works down to 4 but all lower stages don't reduce power usage.

1130w at power state 6, 1062w at 4, 0-3 are all 1060 also.  Do you know what might be going on?  

It is stable and hashes at all levels of powerstage even 0.

[philipma1957]

smOS is changing some of my card's core clock speeds without any input.

The snapshot below is a of a rig running 7 RX 570s, all with stock bios, and no OC settings applied (ever) via the Dashboard.



As you can see, two of them have switched from the default 1128 to 1191.

This seemingly happens at random to different GPUs across all of my rigs.

The change to 1191 is not random, however. A couple of my rigs currently have some GPUs with a custom bios which has a default core clock of 1191. The rig in the photo has not ever had anything but cards with the stock bios connected. I have also manually input 1191 as the core clock on other rigs of mine, but I have never applied an OC to the rig in the picture.

Changing the core clock settings in smOS does not seem to have any effect. It will change the clock upon startup, but it may change again before mining even starts, without any notice in the terminal.

edit: 1128 is not even the stock clock according to MSI spec sheet, it is actually 1268. So why is smOS automatically downclocking my cards???

it is using linux drivers not windows  which may vary a bit

[viperman2]

Pretty sad that no answers on the issue with cards dropping to 0 mh/s from Tytanick.

[ddsilveira]

Ok something is obviously wrong with SMOS after 1098 update, my cards hangs randomly, they drop to 0MH/s and then rig restarts. This never happend to me before, Tyt can you please take a look at this issue? from what I can see I am not the only one struggling with SMOS stability issues after patch 1098. Thanks.

In my case, I have a theory ... I have 2 diferents memory values in my rig for diferent cards. What I believe is hapening, its if the system is rebooted for any cause, the gpu order is mixed, and the wrong memory clock are applied. That causes fails on different VGAs ... For final test, I get all the cards for the lower memory clock and now I'm waiting to see whats going to hapen.

Any sugestion?

[supersporttdi]

I have some RX570's that had no overclocking set at all that ran solid no crashes or reboots for a few days. I made a change and reloaded the miner with 1098 and now I'm getting GPU FAULT PROTECTION across the screen then eventually a crash and reboot. 1098 is the only change that made a difference. i switched back to claymore 9.4 and that made no difference.

[Sequoia93]

smOS is changing some of my card's core clock speeds without any input.

The snapshot below is a of a rig running 7 RX 570s, all with stock bios, and no OC settings applied (ever) via the Dashboard.



As you can see, two of them have switched from the default 1128 to 1191.

This seemingly happens at random to different GPUs across all of my rigs.

The change to 1191 is not random, however. A couple of my rigs currently have some GPUs with a custom bios which has a default core clock of 1191. The rig in the photo has not ever had anything but cards with the stock bios connected. I have also manually input 1191 as the core clock on other rigs of mine, but I have never applied an OC to the rig in the picture.

Changing the core clock settings in smOS does not seem to have any effect. It will change the clock upon startup, but it may change again before mining even starts, without any notice in the terminal.

edit: 1128 is not even the stock clock according to MSI spec sheet, it is actually 1268. So why is smOS automatically downclocking my cards???

it is using linux drivers not windows  which may vary a bit

Well that may explain why they seem to default to 1128. But I'm still unsure as to why the core clock would be changed, sometimes after booting/passing the "OC Phase" of startup, but before/just as mining starts.


On another note, one of my rigs spits this out before eventually booting:



Then mines at a severely reduced hashrate, sometimes with as little as 2 or as many as 6 cards, eventually crashing and rebooting.

[viperman2]

If this is not fixed by tomorrow morning i guess i'll make my own Ubuntu distro like i did for my Nvidia rigs.
This auto updating should be optional in my opinion.

[supersporttdi]

If this is not fixed by tomorrow morning i guess i'll make my own Ubuntu distro like i did for my Nvidia rigs.
This auto updating should be optional in my opinion.

For the life of me I can't find where to skip or disable the auto-update feature either. What's the point of offering older builds if it just updates to the newest as soon as it boots?

[philipma1957]

Pretty sad that no answers on the issue with cards dropping to 0 mh/s from Tytanick.

okay  show a screen shot of your rigs

with your clocks

[Tummler]

Ok something is obviously wrong with SMOS after 1098 update, my cards hangs randomly, they drop to 0MH/s and then rig restarts. This never happend to me before, Tyt can you please take a look at this issue? from what I can see I am not the only one struggling with SMOS stability issues after patch 1098. Thanks.

In my case, I have a theory ... I have 2 diferents memory values in my rig for diferent cards. What I believe is hapening, its if the system is rebooted for any cause, the gpu order is mixed, and the wrong memory clock are applied. That causes fails on different VGAs ... For final test, I get all the cards for the lower memory clock and now I'm waiting to see whats going to hapen.

Any sugestion?

I am experiencing the problem, and I have the same OC settings for all cards.

[Tummler]

If this is not fixed by tomorrow morning i guess i'll make my own Ubuntu distro like i did for my Nvidia rigs.
This auto updating should be optional in my opinion.

I use a customized Arch-based image - but if you don't wanna do that, IMO, give EthOS a try. Auto-update is not only fucking stupid from the POV of possible breakage, but if this OS is ever super popular, the mechanism will almost certainly be compromised and malicious updates pushed to all the poor bastards who use it.

I share your opinion on auto updates. I had perfectly stable rigs for over a month, and now that is out the window. I understand if Tytanick decides that he will only help troubleshoot rigs that are completely up to date, but for those of us that have a good thing going, let us be.

[xleejohnx]

If this is not fixed by tomorrow morning i guess i'll make my own Ubuntu distro like i did for my Nvidia rigs.
This auto updating should be optional in my opinion.

I use a customized Arch-based image - but if you don't wanna do that, IMO, give EthOS a try. Auto-update is not only fucking stupid from the POV of possible breakage, but if this OS is ever super popular, the mechanism will almost certainly be compromised and malicious updates pushed to all the poor bastards who use it.

I wish I had the ability to make something like that!

[philipma1957]

okay lots of testing of beta nvidia build.

I can get 95 watts on gigabyte 1070 itx cards

pretty good compared  to 150 stock.


the clocking does not work for my msi 1070 itx cards

[HolyHatred]

Hey im having an issue where when the miner starts up the miner says there are no gpus for the miner.
the os it self see's 4 on the rig. any suggestions?

[-droid-]

was going to switch to this on my NVidia rigs, but with no zec support ill hold off for now

[philipma1957]

so I will run 2 gigabyte 1070  itx cards on eth

hoping to get more support in a few days

ewbf_ 3.4   

 would be nice

better clocks

would be nice

I can only do 2 cards  on a mobo that did 4 amd cards

so being able to do 4 cards would be nice.

[carrionconfetti]

Hello, I'm having some strange problems whenever I try to get more than two cards working on any of my systems. I've managed to get 3 cards working with some bios fiddling, but I couldn't reproduce it.

I'm running

Intel 3900 (Skylake)
Asus PRIME z270-P
8gb Ram
A mixture of various 580 4gb and 8gb

I've enabled 4g and find that I'm most stable with everything (PCI, etc) set up to Gen 1. Are there any other settings that need to be changed?

I'm using a 16gb usb 2.0 pen drive. I've tried both the normal and z170 Simple Mining versions.

When I boot it, regadless of success, I get a few flashing screens of this weird rainbow static.

https://i.imgur.com/tl1W0bN.jpg

Then, I boot in Xubuntu, as normal, but then get this.

https://i.imgur.com/LUeBvEK.jpg

There's another similar error that just locks the system up until you press a button and it resets the entire system, but I haven't been able to reproduce it. My biggest issue is that I'll have my system be stable at 3 cards, but when I restart it it'll not boot up normally again.

Any one have any ideas/bios settings? I've had some success, but I really want to reproduce that 8 GPU z270 build I saw.

[Hostels]

Use Gen2 on PCI. This setting in 3 places in BIOS.
100% 6x580 worked.

[jadefalke]

what is the Username/password for Login to Linux? I want to assign static IP Addresses.