Hi there, i have a rig running with 6 AMD W5700 Pro cards on hiveOS 0.6-203. The cards have a modded mining bios and produce 55,x MH/s @ 86-92W. The whole rig consumes ~720W.
Last week i updated my rig to hiveOS 0.6-207 because it has a newer version of TeamRedMiner. The update was successful, but then all cards consumes 20W more like before. (~110W instead of ~90W) The whole rig then consumes 850W with hiveOS 0.6-207.
Therefore i downgraded hiveOS back to version 0.6-203 and the power consumption is low like before.
I dont know what caused the higher power consumption, maybe changes in hiveOS or in the newer version of TeamRedMiner.
Has anyone of you get the same experience or maybe even a solution? Thanks in advance…
Edit:
I just see TeamRedMiner is not the cause, because its remained version 0.8.4 after the downgrade to hiveOS 0.6-203.
Hello there. Are the measurements on the wall or in the GUI? With what memories are the cards micron or Samsung? Can you send me the moded bios? kosiosln@gmail.com
check the changelog of TRM, they made a huge kernel optimization some time ago that needs you to adjust core clocks. could be that you were on the old kernel and just updated to the new TRM kernel? might be the case…
Hi, thank you for your answer. The value is not only from the GUI, but also measured directly at the power socket.
I have a card with Micron memory, which consumed ~105W before, after the update 130W. The 5 other cards have Samsung memory, they needed 86-92W before, after the update each needed ~110W.
I tried 3 times updating to different newer hiveOS versions, the result was always the same. At some point I had no more desire and then just let it run with hiveOS 0.6-203.
You can download the modded bios here, in posting #70 on page 4.
maybe a new fresh install with actual 0.6-211 ?
did you check amd driver? should be 20.40 now, might be that you were on 20.30 with your previous version? maybe the driver is the issue.
I would give core clock a try…lower at around 100 and check mhash and power just to be sure. could be a kind of card detection that you were running at older core with newest TRM version. You also can force the kernel via command line.
I’ve had it yesterday after some changes to the rig. I was looking at my screenshot from earlier and I’ve noticed that each card was like 20w more than it was before. All I did is clear up the OC settings on the Overclock tab.
