Multiple GPUs – training crashing

[HydrohomieParticipant]

Hello everyone,

I am experiencing a problem with my water-cooled 3 GPU 3090 setup, and I could use some help to resolve it. Whenever I try to train with 3 GPUs, my computer shuts down immediately upon training. I initially thought it was a power supply issue, so I installed a second power supply dedicated solely to powering the GPUs, but the problem persists.

I’ve also tried running the system with only 2 GPUs, but it trains for about 5 minutes before the Windows blue screens. The only configuration that works is a single GPU.

I am unsure what else to try, and I could use some advice on how to resolve this issue. Has anyone experienced a similar problem before? Do you have any suggestions on what I could try to diagnose and fix the issue?

Any help or suggestions would be greatly appreciated. Thank you in advance.

[Author]

Posts