I have been running a 6xRX480 rig (Sapphire, 4xNitro, 2xNitro+) for a couple of weeks now, and every now and then (1-2 days) it would lock up.
There is no BSOD, nothing in the Windows event logs or miner's logs. Just a black screen when i attach the monitor, no connection on VNC and no activity on the pool. Fans spin down to idle.
A screenshot tool that stores the last 30 seconds in a loop does not show anything unusual - a second before the lockup looks exactly like any other second, all GPUs report normal speeds and are in the 60*C range, no errors or anything unusual in the console.
Windows 10 with all the update stuff disabled, AsRock H81 Pro BTC R2 board, all GPUs are modded to some extent, each was tested on their own for a few hours under full load, Claymore miner, 16.9.1 driver.
Cards are 1625->2000 modded, downclocked to 1000..1140 core and overclocked to 2050..2100 by memory (manually found peak performance for each). Each is undervolted to it's stability limit +50mV, which is between 950mV and 1100mV.
What could be causing this?
If one of the GPUs is marginal, then how can i identify it without having to go through them one by one days at a time (RMA windows starts to end around May 13th)?
There is no BSOD, nothing in the Windows event logs or miner's logs. Just a black screen when i attach the monitor, no connection on VNC and no activity on the pool. Fans spin down to idle.
A screenshot tool that stores the last 30 seconds in a loop does not show anything unusual - a second before the lockup looks exactly like any other second, all GPUs report normal speeds and are in the 60*C range, no errors or anything unusual in the console.
Windows 10 with all the update stuff disabled, AsRock H81 Pro BTC R2 board, all GPUs are modded to some extent, each was tested on their own for a few hours under full load, Claymore miner, 16.9.1 driver.
Cards are 1625->2000 modded, downclocked to 1000..1140 core and overclocked to 2050..2100 by memory (manually found peak performance for each). Each is undervolted to it's stability limit +50mV, which is between 950mV and 1100mV.
What could be causing this?
If one of the GPUs is marginal, then how can i identify it without having to go through them one by one days at a time (RMA windows starts to end around May 13th)?