Quote from: cidman on Today at 09:20:41 PM

Quote from: Ortadan on January 06, 2021, 11:27:24 PM

In RaveOS the latest version does not work, when starting the miner the platform it restarts using version 1.18 and 1.18A (worker_rebooting_by_flag_auto) this happens in my two 4GB platforms, instead version 1.17 works perfectly in both.

Am I doing something wrong or is it a problem with the version included in RaveOs?

im having the same symptom
i havent checked the other rigs cuz they all work with 4080 ZM 2
but the rig im having the same problems as u and cannot run at ZM 0
ZM 1 was low hashrate at first
but rebooting a few times then making sure the settings were right in logs got the desired outcome
im up 1-3 Mh per card across multiple rigs now

edit: im going to go make sure the pcie are all running gen 2
that rig is also running a pcie splitter which could be the problem

edit 2: yea maybe it doesnt work so much
the 4 "problem" cards (prolly the ones on the pcie splitter) dropped from 21Mh to 16Mh well under 1.17

If you don't have Gen2 the --zombie-tune 2 will not work and you should change 2 for 1 or 0.

The idea is to run miner-show, to see why it crash. Without that information is not possible to know where is the problem.

Best parameters with Gen2 are: --4g-alloc-size 4080 --zombie-tune 2

It could happen that GPU0, GPU2 can not run 4080, so it can be write it like:

--4g-alloc-size 4078,4080 ,4076,4080,... --zombie-tune 2

Choose each size for each card. And also same for --zombie-tune:

--4g-alloc-size 4078,4080 ,4076,4080,... --zombie-tune 0,2,1,0,2...

I hope that helps to have best optimisation.

2 is max optimisation
1 is medium optimisation
0 is like 1.17 with Watts improve