I also have some Gigabytes with high wattage, like 1.5x the wattage of other Vega’s; must be something Gigabyte specific 
And on my 13x Vega rig, --eth_config=B, as opposed to what it should do, gives me 20W more out of the wall than without it.
Off topic; your 1080ti should be hashing much higher though; or you should get lesser wattage out of it for the same hashrate. Your memory should be able to go +1200 up to +1400 and set core maybe to +200. And have you applied the pill?
dude thanks alot even looking to photo and try to make same survived me 
hello [0Lucifer0 can you help me
Is there anybody using this tutorial with a hiveos version over 0.6-195 ?
I can not get it to work with any of the version > 0.6-195 if I downgrade to 0.6-195 it works fine but if I upgrade it just stuck the gpu at miner startup.
shared with hiveos support but the only result I got is : check your risers (those are fine it’s a hiveos version issue) and “that’s weird”
I’m also on 0.6-195 as all versions after that version don’t seem to load xinit.user.sh so straps aren’t applied.
I can thus confirm it is a hiveos issue and no risers.
If there was a clear bug reporting system I would have already reported it and supported to solve it with logs or whatever could help, but I haven’t found out how to contact support.
Guys - The straps work, but the directions originally created are wrong.
In the remote access screen, type the following
nano /home/user/xinit.user.sh
Add the straps, save it and you will be good to go!
that the file I modified and still it doesn’t apply the straps
straps there works fine /etc/xdg/openbox/autostart
It doesn’t seems like I can edit my first post : @HaloGenius if you can
Would be nice to add some things :
if you are on version 0.6-195+:
add the straps to /etc/xdg/openbox/autostart
else
add the straps to /home/user/xinit.user.sh
to apply the straps you need GUI enabled
another straps that works even better for me : --CL 20 --RAS 20 --RCDRD 12 --RCDWR 12 --RC 37 --RP 11 --RRDS 3 --RRDL 6 --RTP 5 --FAW 12 --CWL 8 --WTRS 4 --WTRL 9 --WR 14 --REF 65535 --RFC 248
also the first strap is having a – instead of – before CL
I upgraded and now I lost the /home/user/xinit.user.sh straps. I tried adding the straps to the script above with no luck. Did I script it wrong?
@kennysum1 I too am interested to know if all this will work on the Liquid Cooled version. I got one cheaply on ebay and it produces a rock steady 47.86 MH/s average on Windows with TeamRedeMiner playing with WattMan. But then I put it into my MixedGPU HiveOS rig and only does 38.75
I am new at al this, can I treat this Liquid Cooled version like any normal Vega64 and which 56 BIOS would I flash for it as there was no Vega56 Liquid Cooled?
Should I upgrade my HiveOS to something higher as I see many here are using newer versions?
Can I run TeamRedMiner for the AMD cards and PhoenicMiner for the NVIDIA’s somehow - by creating differernt workers and specifying cards for example?
As for the GPU in question, in HiveOS (0.6-190@210106) it shows:
Radeon RX Vega 64 8176 MB · AMD/ATI
Samsung HBM2 · 113-D0500500-102
It looks like a Sapphire RX Vega64 from images.
I have mix of 7 NVIDIA and AMD cards, so am using PhoenixMiner as it seems to work for both brands.
I would love to get at least the 47/8 I was getting in Windows. Any advice much appreciated.
Yes you can flash to liquid cooled Vega 64 down to the 56 and treat it like a normal vega. Only exception is that I have found is it needs a bit more power most likely for the pump.
You can run 2 different miners in Hive in 1 flight sheet. Run trex for Nivida and Teamredminer for AMD. Hive has some easy videos on how to set that up.
I have flashed my liquid cooled 64 to a 56 and run it at about 52 mhs. Every card is different and some can handle higher memory speeds.
@kennysum1 thank you for your feedback. Can I ask which version of 56 Bios I should choose, as the advice in this thread is to “Choose the same v56” card and use that BIOS?
Also can I safely do the BIOS flashing from with HiveOS or do you recommend I pull it out and use my Windows machine to do it?
With regards to HiveOS I am using stable 0.6-190. Is it safe to, and should one, hit that update option from within HiveOS and go to the latest 06-203-stable?
@DesertSweeper i think can chose any bios check only ram complatible manufacturer
eg i flash my VEGA64 asus strix oc with VEGA56 asus strix non oc stable than oc version
anyone can under clock lower than 1000 Mhz and get stable?
i can’t get mem higher than this so sad
So for those new to this, like me…I took the plunge and upgraded my HiveOS to the latest “stable” version from within the worker. So now I am on 0.6-203@210403
Then I identified my Vega64Liquid Cooled as an XFX model so I went to techpowerup and located the BIOS versions there for the XFX Vega56 card. There I found 4 variants with different numbers. I downloaded the first one ending with 8766 and in HiveOS, after stopping the mining (top of screen “Miner Actions”) and saving my BIOS there (in case I need to put it back later), flashed the BIOS to the card and rebooted. Without any further tweaks my Hash rate went from 38.75 to 40.24 and the power draw dropped from 144W to 138W. Even the temp is a little lower after running for an hour.
So now I need to go back up and read from the start to figure out how to do the tweaks, and move it to a new TeamRedMiner worker as I am currently using a single worker with mixed GPU’s on PhoenixMiner.
Updates forthcoming!
@kennysum1 thank you for the dual-miner tip - this has improved my overall rig numbers and saved some power, and my Vega64 jumped to 45.57Mh/s so the journey is a good one so far from the 37 I was getting when I started. But now I want to apply the amdmemtweak and overclock discussed in this thread but have run into a problem in HiveOS. In the examples I have found, on how to do this final step, they show the GPU’s listed in the miner numerically. So GPU 0, 1, 2 etc. But in my rig I get cards repeated…so I have the first two cards listed as GPU 0:
So how do I tell amdmemtweak which card to apply the straps to?
Any advice would be appreciated.
Greetings @0Lucifer0
You have mentioned above that —i Index of card can be used to specify the card you want to apply the settings to, but how do you identify that index when you have two miners showing two GPUs both as “GPU 0” for example?
Also I am seeing videos on YouTube where they use --GPU to identify - is there a different version of amdmemtweak?
This strap is $$$ Thanks for sharing, updated all my Samsungs to this and it got rid of invalids and let me push mem higher. and way more stable
Is there a simple straight forward guide to adding straps?
Remove those DPM settings.
Thanks to this guide and thread in general I have gone from 7Mh/s on my V65LC to 36 to 45 to 50 and now a rock-solid 56MH/s running absolutely stable now for 24+ hours. Gratitude to @0Lucifer0 and the others who shared their knowledge.
hello, what are your overclocks for 56mh?
