i've been trying to crack this problem with the different hashrate performance on 5 of my reference vega 64s.
they are all reference cards except that 1 is a limited edition and 1 of them has a MSI sticker, 1 has a xfx sticker while the rest are sapphire.  (as far as I can tell, they are not from the same brand)

I even went as far as to save the bios from the card that hashes at 2000h/s and force flashing into the cards that hash at 1600h/s

all 5 cards went through the same disable/enable scripts and the only difference i can find is in their P states.


out of the 5 cards, only 2 hashes at 2000h/s, the rest hash at 1500 to 1900h/s





anyway to force the cards that have 1200mhz at p5 to run at 1400mhz?

anyone knows why I can't seem to get shares recorded in nanopool (xmr) when I am using xmr stak?

Previously I was using xmr stak with supportxmr pool then some issues happened so i switched to cast xmr with nanopool. then I decided to give xmr stak with nanopool another go and it was hashing fine for a few mins until i added failover server addresses (bad idea, it doesn't seem to switch to other servers like claymore) so i undid those changes

but nanopool records it as offline but it regularly submits shares






as you can see, the wallet address is correct and it is hashing fine.

I've been trying to sync my monero gui wallet but it seems to be taking forever. As a result I can't see any monero sent to my wallet.

 I even tried entering "--block-sync-size 10" but it is stuck at 777** blocks remaining and the number is actually increasing!

The status is filled with messages like this

2018-01-02 11:04:47.946   3580   INFO    global   contrib/epee/src/mlog.cpp:145   New log categories: *:WARNING,net:FATAL,net.p2p:FATAL,net.cn:FATAL,global:INFO,verify:FATAL,stacktrace:INFO
Height: 1400002/1400002 (100.0%) on mainnet, not mining, net hash 258.11 MH/s, v5, update needed, 0(out)+0(in) connections, uptime 0d 0h 0m 1s

2018-01-02 11:10:25.140   8428   INFO    global   contrib/epee/src/mlog.cpp:145   New log categories: *:WARNING,net:FATAL,net.p2p:FATAL,net.cn:FATAL,global:INFO,verify:FATAL,stacktrace:INFO
Height: 1400002/1477991 (94.7%) on mainnet, not mining, net hash 258.11 MH/s, v5, update needed, 7(out)+3(in) connections, uptime 0d 0h 5m 37s

2018-01-02 11:11:08.425   8736   INFO    global   contrib/epee/src/mlog.cpp:145   New log categories: *:WARNING,net:FATAL,net.p2p:FATAL,net.cn:FATAL,global:INFO,verify:FATAL,stacktrace:INFO
Height: 1400002/1477992 (94.7%) on mainnet, not mining, net hash 258.11 MH/s, v5, update needed, 6(out)+3(in) connections, uptime 0d 0h 6m 20s





I'm using samsung 960 pro ssd and 1gbps fibre connection so I doubt it's a connection or ssd issue,.

asus mining expert with latest bios
windows 10
Intel g4400
8GB ram
120GB ssd
11x vega 64
2x 1600w PSU
min 60Gb of virtual memory, max 80Gb


this miner worked with minor issues when it just had 5 vega 64s and 1 of the 1600w PSU. but the moment I added another 6 vega 64s, it would take a long time to reach windows but everything becomes extremely sluggish and BSODs appears with amd drivers as the cause of it.

ok so I removed the new cards, raised the virtual mem to 60gb, max 80gb. but didn't work.

Installed 17.10.2 drivers, but it stops me from shutting down the miner with some memory errors






clicked ok and when it reaches desktop, it says no amd drivers installed.

also, I can't seem to right click to pin any program to the taskbar or make the windows start menu appear.

I am having problems after problems with my vega 64s and I have already spent days trying to fix it. the first gpu is a msi branded V64 (v64-1) and the 2nd is a xfx vega 64 (v64-2)

(PC Specs)
msi z370 gaming pro
ryzen 7 cpu
32GB gskill ddr4 ram
amd aug23 blockchain drivers

V64-1 has been working a little strange at first.

I started with Vega_64_Reg_GPU_0000-1100HBM but after a while it refused to work and hashed at 500mhz or 22mhs no matter what I did. reinstalled drivers, etc. eventually tried another reg file from overclock.net (MorePowerVega64_142_Modded) which worked nicely. 44mhs at decent power draw.



(miner PC setup)
asus mining expert
intel g4400

4gb kingston ram
xfx vega 64
msi vega 64

I bought another vega 64 (v64-2) and installed both on the asus mining expert board and that's when all the problems started.
Used the exact same registry mod file, overdriveNtool settings (in fact i copied the file over so I had the exact same settings). for some strange reason, the acoustic limit default is at 15mhz for the new setup instead of 2400mhz.

performed the exact same steps as per the guide but one card hashed at 44mhs, the other at 22mhs. applied the registry mods again and same thing happened. then a riser gave problems. replaced it and both cards hashed at 22mhs. it turns out the memory was running at 500mhz.

I couldn't apply the same overdriventool settings to the same v64-1 graphics card that was using the same settings on the ryzen 7 PC, 1 of the P states for CPU jumped back to some random number.

reinstalled the V64-1 back to the ryzen 7 pc and it worked fine. so it's not the card's fault.

I reinstalled windows, redid the exact same setup. but during the step where I had to reset settings in wattman and the p states would drop. that part didn't happen, as if the altered P states were already changed before I had to reset it.

The registry showed some very interesting results, normally. 0000 refers to the integrated graphics, 0001 and beyond is for any additional cards. But this time round. 0000 and 0002 are the 2 vegas and 0001 is the int gpu. so i modified the target folders in the reg file.

This is where I found another strange part. Somehow, somewhere during all the copying from PC to memory stick and back to the miner PC, the registry files contained an additional line about donations to a certain ether wallet.



I decided to try Vega_64_Reg_GPU_0000-1100HBM which had worked for 1 of the cards then mysteriously not work. both cards started hashing at 44mhs but produced twice as much heat and noise. the moment I used the morepowervega mod, it hashed at 22mhs again.

I've always wonder how do people power their mining pcs that have more than 8-9 GPUs if they require 8+6 pins and each psu only has 4x 8pins and 4x 6+2pins with just 2 PSUs?

The PSUs need to power the risers (either using sata power or 4 pin molex) and any connectors on the motherboard itself.