When I was a kid, we had a rule of thumb (which was kind of a joke), that "If you don't fall down, you're not skiing hard enough. The point being you weren't pushing your limits if you didn't occasionally push them too far. I can't help but use the same rule of thumb for my GPUs. I'm lucky if my workers (of 6-7 GPUs) make it over 48 hours w/o a failure or reboot. Ideally they are soft-boot failures and don't require a reset. However, that does happen, and with some machines more than others.
I was wondering if anyone would like to share how hard you push your GPUs and worker machines? How often you you get a GPU crash, whether soft or hard? Does your "time running" go multiple days or are you like me and usually have about one reboot per day, on average?
When you are rebooting or recovering from a failure, you're not mining, so it's costly. Yet, if you aren't 'right at the limit' of your GPUs, you're leaving money on the table. It's a balance. Everyone has an opinion on what that balance is. I'd like to hear it.![:smile:]( ":smile:")
I was wondering if anyone would like to share how hard you push your GPUs and worker machines? How often you you get a GPU crash, whether soft or hard? Does your "time running" go multiple days or are you like me and usually have about one reboot per day, on average?
When you are rebooting or recovering from a failure, you're not mining, so it's costly. Yet, if you aren't 'right at the limit' of your GPUs, you're leaving money on the table. It's a balance. Everyone has an opinion on what that balance is. I'd like to hear it.
