If I launch all 8 GPUs in one instance, the power usage curve jumps from 30% to 100% each 3 seconds, making miner not PSU-friendly.
I had same issue with other skunk miners, so my own recipe was to run 2 miner instances.
But when I launch 2 instances of miner to make separate process for 3x1080ti and 5x1070 GPUs,
1080tis are working great and very stable, but 1070s doing something weird: after a couple of minutes two last GPUs stop working at all.
Also running 2 instances does not change stratum difficulty, so my hashrate is lost between blocks if they found too quickly.
Since 3-4 SIGT blocks can be found in a row with small steps like 3-5 seconds, sharerate for SIGT should be at 15-30 shares/minute. I understand that it affects pool stability, but miners efficiency is on other side.
Miners will get reward only for 1 block in that cascade with low sharerate.
So actual mining efficiency drops by 75% in those cascades of blocks.
I had same issue with other skunk miners, so my own recipe was to run 2 miner instances.
But when I launch 2 instances of miner to make separate process for 3x1080ti and 5x1070 GPUs,
1080tis are working great and very stable, but 1070s doing something weird: after a couple of minutes two last GPUs stop working at all.
Also running 2 instances does not change stratum difficulty, so my hashrate is lost between blocks if they found too quickly.
Since 3-4 SIGT blocks can be found in a row with small steps like 3-5 seconds, sharerate for SIGT should be at 15-30 shares/minute. I understand that it affects pool stability, but miners efficiency is on other side.
Miners will get reward only for 1 block in that cascade with low sharerate.
So actual mining efficiency drops by 75% in those cascades of blocks.
first - have you set the --lodiff option in the parameter list? ... the miner will automatically force the hidiff port if you have more than 4cards in a single machine ...
second - CWIgm not meant to be psu friendly ... its meant to be stable and have continuous hash for the life of the mining session ... if your psu is not of high quality and cannot support the cards - you need either a bigger one or two to support the power consumption your cards will be working with ... you can of course use your oc settings to lower power and oc - but the investment into high end high quality psu units will make the mining a LOT more stable than what you have ...
i have the same issues when i first started mining - and quickly learned that the psu quality - and number of psu - is as important as the cards themselves ...
now if you last two 1070 stop after a few minutes - then you definitely have an issue of some sort ... check their heat / throttling - check the power they are consuming - watch closely to the display CWIgm displays ( you only need ONE session of CWIgm ) - and make sure the cards themselves are not the issue ...
let us know if any of this helps - or any other advice the community may have towards this ...
#crysx
