I would also suggest you get a bigger drive if you can.  If you plan on running full nodes for whatever crypto you mine and I guarantee you'll probably try a couple of different ones in a few months, you'll run out of space real quick.

Thanks, that's pretty much what I have, except my powerlimit is at +15% (and I've been playing with maxing it out to +50% with only a 10-20 hash rate/s increase).  Doesn't seem to be worth it.  Running right now at +5% and getting solid 2k hashrate (48 hours stable now) on Gigabyte cards.  Sapphire, however, freezes up when set to 950MHz, had to scale it down to 910Mhz, and it's hashing 1920 h/s stable.  All cards are running at 65C though, not 50s when ambient temp is about 20-22C.  Fan speeds are at 2k min, current at 3.3k all of them.

On stock vega 56s (stock bios), what are the best wattman settings without the reg edits?  I've been messing around with some of my cards and while some handle 950Mhz well, one of the cards can barely stay stable at 920.  Most of the other settings are all dynamic, but I'd like to change that.

Thanks.

Default 56 bios with cranked power (-30%/950 +50% power)

Is it strange that I'm getting those results (2k hash rate, 70C temp) and my clocks are 1112/950 +50% power?

How are you guys getting sub 50C temps hashing 2k on Vega 56s?  Are you blowing freezing air on them or something?  I'm at 68-70C doing 2k at room temperature and with a 120mm 5 fan array cooling.

Thank you very much.  increasing VM fixed the error message, now running consistent 1975 h/s per card without tweeks using default cooling on the blockhain driver.  Still connected to onboard video (Intel).  Wondering what I'll get with HBCC enabled.  For some reason it crashes my AMD Settings prog as soon as I try to change it to anything.

I have an issue with my setup.  Can someone point out what I'm doing wrong (my first setup).

I'm getting very low speeds on my 2x vega 56s regardless how I set it up.
I've tried the following with the results:

- index 0 only: 1 thread per GPU 2016 results in 600H/s
- index 0 only: 2 threads per GPU 2016+1600 results in 236H/s
- index 0+1: 1 thread per GPU 2016 results in 1175H/s
- index 0+1: 2 threads per GPU 2016+1600 results in weird rates ID 0 = 132, 1 = na, 2 = 221, 3 = 853 and 1 and and I get CL_MEM_OBJECT_ALLOCATION_FAILURE

"platform_index" is set to 1 (doesn't work when set to 0)
Sometimes xmr-stak-amd freezes when setting up 2nd "Device 0 work size 8 / 256" or sometimes on Device 1

My setup:
ASRock H81 BTC R2.0 mobo with G3260 CPU and 4GB of RAM running Win 10 64bit
Both cards are connected via risers (PCI-E #2 (long) and #3 (short) are in use out of 6 total) and powered by 1000w EVGA gold
I am connected to the onboard HDMI for monitoring
There are no HDMI dummy dongles in the Vegas
Running blockchain driver v17.7

Anything obvious pops up?  I'll try disabling onboard video and connecting the monitor to one of the Vegas, but can you see any other issues with this?  Pointers perhaps?  

All of this is running on default 1580MHz/800MHz speeds (haven't tweeked any Wattman settings and HBCC is off)

Thanks a lot in advance.