Crashing / OC settings / fan test


#1

Specs -

Z370-a
Strix 1070 (8)
Corsair ax1200i
4 gb dd4
i3-8100
8 risers
1 m.2 adapter
16gb usb3.0

All 8 cards are hashing stable with no overclock and a 76w power draw. Any time i add any OC, or any more wattage, my machine reboots. It is making me crazy, because earlier today, i had all of my wiring loose on the table, with OC settings of +100, +1000, 87w and it was stable for a few hours with no reboots while i took a nap. I was getting 200± MHs. I woke up, turned off the machine, disconnected all power, and rerouted all the wiring to get clean wire management on my rig. Upon reboot, i was getting constant reboots everytime the cards started hashing. I cleared out all OC, and started at 75 watts (minimum for 1070), sloly increasing until i got to 77, and it crashed. I am stumped.

With 1200 watts, 80% is 960. If i give up 200 for other accessories, i am left with 760. This should easily leave me 95 watts per cards to hash with…and that is being conservative with the 80% max load i am choosing to utilize.

I dont understand what could have happened when i powered down and rerouted cables which would have given me such drastically lower performance.

I did the gpu-fan-find command, and found that it was spinning my card on the first 16x lane whenever it spun card 0, and card 7, which is the one using the m.2 slot. This means that the fans on the card plugged into the m.2 adapter (card 8) never was identified with spinning fans throughout this test. I dont know enough about all of this to know whether this could possibly be part of the problem, but it doesnt seem to be normal. Is there any way that it could be trying to send voltage to the same card for both lanes (i dont know if that makes sense) and this is what could be causing the reboot? I am convinced that the rig is rebooting due to power concerns since whenever I OC or increase wattage, it reboots.

This is the first post ive ever made on here, so any help or knowledge would be greatly appreciated.

Oh, one more thing. If i boot this rig with SMOS, I am able to OC and give the cards more wattage without the reboot, but my hash rate always tends to be higher with Hive, so I would prefer to be able to get this diagnosed and stay on Hive.

Thanks in advance for any help!!


#2

Update - I have another rig with 8 1070Ti’s.

I went into it and did the gpu-fans-find command, and as i suspected, each card test spun a seperate fan. This confirms, to some extent that it has something to do with the lane assignments in Hive.

So…next question. What could have happened when i was rewiring my rig that would have caused this to happen, and, more importantly, what can I do to resolve this problem? I have tried delting the rig from my Hive account, and reloading it, but it comes back with the same problem.


#3

Update 2 -

Back to the problem rig.

I changed my m.2 adapter to the m.2_1 slot to match my 1070Ti rig. I restarted, and ran the fan-find test again. All fans are spinning independantly now, as they should, but i am still unable to OC without a crash.

So, apparently, the fact that one card spun for two of the tests was not related to the issue. Or if it was, it didn’t solve the problem.

Back at it.