Topic: [Mining OS] SimpleMining.net - Manage Your GPU farm the easy way! (30 days free) - page 515. (Read 835461 times)

I am setting mine to 1100/2125/5, since all my ROMS are custom for low power already.

Will let the 3 rigs for testing in RX farm overnight with this settings to test stability.

I will monitor the counter for high reset/reboot/reloads.

test on one. also noye wattage before and after.

okay I set all my rigs to underplot 3

i use 16 amps x 240 volts

I thought nobody will ask
Its my idea ov undervolting.
It is DPM level which you want modyfy and set your clocks.
Each dpm has specified voltage so taking lower and lower will result in more less undervolting until it will be unstable.
You can compare wattage you have with the same clocks on dpm 7 and dpm 4.
DPM 3 shoudl also work, meaby 2 too, but 1 and 0 will propably hang/freeze rig
I myself have in most rigs values of 3 (but i also have offset underfolting in custom bioses)

What exacty stand the undervolting values for?

Yes you can do this but accually i am not sure if gpus will be initialized in the same orer after reboot.
I still didnt have time to specify this.
If you will figure out tell me
anyway you can

Core: 1010,1020,1030
Memory: 2000,2010

When you specify like that then
GPU0 1010/2000
GPU1 1020/2010
GPU2 1030/2010
GPU3 1030/2010
etc
So if there is more gpus than you specified in config then they will use last value.

this is GREAT man! thanks a lot... will try and tell how its working out for me

With the latest update can you now set specific hashrates with different rates ex 1050,1075,1150

It is very simple to get a pxe server up and going
I think I'm having a memory issue. I am only using 8gb of ram for a 7.7gb os lol

I never had a chance to do PXE boot so i really dont know how to do this

Theoretically Pxe for mass deployment wouldn't be wise at all
But for something small it would be worth having for that moment you forgot to get a hard drive or USB
I can't get the compressed Img of smos to boot.
This is more of less for me to do as a hobby/option for future involvement 

Update v1077:
- in RX OS you can now overclock to specific MHz values (like core: 1055 mem: 1988)
NOTE: you can write new specific MHz values and save it, but after that you need to do one reboot of your rig

- in RX OS added little undervolt option in which you can write values from 0 to 7.
NOTE: Default value is 5 and up to 7 its save. Lower values like 3-4 sould be also stable but 0-2 propably not.
When you will write there lower values your rig can freeze !!! so be advised and test this when you are near rigs !
Also REFRESH dahsboard WEBSITE


- total restarts count - means how much times does your rig started miner (so how many reboots or reloads)
Soon i will add also "reset counter for all rigs" feature. Now you can monitor how may times does your rigs is restarting and if too much then lower clocks or less undervolting.

This adds each time the miner starts so reload and reboot and SRR reboot - all will add +1 to this counter.
I will also make soon button "RESET ALL COUNTERS" - i think this will be usefull and no need to do detained history reboots.


I have plans for making all informations in table but you will be able to select WHICH COLUMNS you want to SHOW/HIDE.
I think its best solution
So if you will want to search ip addres then you will just need to show ip collumn and then ctrl+f.
And when no needed you can disable this column

Wattage of what ? In OS you cant really know how much does gu takes, those numbers are not right.
Best way is to buy analog watt meter and plug in lone rig to it and do the modifications and check which are best.

Can you ask again ? I dont understand this.

This is idea that i had few months agon.
One PC with PXE boot delivery and each rig with Ethernet boot.
It is more less ok. But...
I resigned this feature becasue lets say we have 1000 PC farm.
After power failure it would take ages to transfer 2GB of data to 1000 rigs
With 1 Gbit network it would take 20480seconds (5 hours)
with 100 rigs it would be 30 minutes to boot all rigs.
Any thoughts ?
Meaby for small deployments between 10-50 rigs it would make sense but not less becasue extra PC costs, and not more becasue it will eat 1 GBit.s right ?
Anyway i thought about this and i wanted to do this but calculating numbers just discouraged me to do this

Has anyone tried to load this into ram over the network?
I have tftp server running on my network and I can't seem to get this to boot fully into ram
Seems silly to do this with USB prices so low but this is just for funs really

Is there any way to fasten konky manager there ready-made modules displaying many necessary parameters?

Do you plan to incorporate a watt usage metric on the page?  If not can anyone give a suggestion on best tool to show wattage usage in linux?

Hi T,

Is it possible to show SRR info, without going to "MORE" menu?

For example, maybe under RIG NAME...

MyRigName
SRR: 1/2

1/2 = srr serial number/slot number

If no SRR, then maybe "SRR: NA" or no display or blank.

This way, at one glance from dashboard, you could see correct SRR info to trigger manual reset on SRR. This could be useful in troubleshooting or testing overclocking.

Thanks

Hi T,

The status - counter feature is very good and useful.

I could see immediately that I have to check 3 rigs that are still alive but has registered reboots or restarts.

Can you confirm, when the counter starts? During miner's WD process (restart) OR SRR's reboot (OS reboot)?

Thanks

Keep the work going man.
smOS best 

il solved with bios mod just change max freq of gpu an vram  , after from panel just  put down or up