6* RX 580 issues - Low Hash

[Bjorntore99]

Hi,

I'm quite new with gpu mining. I have set up a 6* rx 580 rig, consisting of 5* Sapphire Nitro+ 4gb and 1 Gigabyte 8gb. The other hardware are:
 Asus Prime z270-p,
Intel Pentium
4GB RAM
1400W PSU

The sapphires has Elpida Memory, and the Gigabyte has Samsung memory. I use EthOS 1.2.9

I have 2 issues:

1 - For now, les focus on the Sapphires. I can only get around 25 mhs per card on average. The memory overclock varies from 1960 to 2180. This because for some reason some of the gpus crashes if I overclock them more, while some can obviously handle more. They're all bought at the same time, and has the Elpida memory, as said before. But still, the hashing power varies from 24 mhs to 26,6. Quite low, right? What am I doing wrong?

2 - The other problem is that the gigabyte card, while it is recognised by ethos, crashes and goes up to 511* C the second it starts hashing. I've read that the 511*C is some sort of error code, but not sure what to do with it. So at the moment I have disconnected this card.

And yes I have tried changing all the risers.

Here is the relevant info from my config.file:

mem (worker name) 2070 2100 1960 2180 2120
globalcore 1411
stratumproxy enabled
maxgputemp 85
And my wallet and pool, and flags...

All help is greatly appreciated!

[MATHReX]

1. For reaching 30+ MH/s with your GPUs, you have to BIOS Mod your cards according to your memory vendor. After that finding the sweet spot of overclocking for each GPU. Then undervolting to reduce power consumption.
2. The error is probably some Driver issue which reads the GPUs temps incorrectly.
Anyway, I'll suggest going with Windows as it will be easier to perform all the activities I state above.