Folding@Home (and other distributed computing things)
Could someone tell me if it's permitted to run Folding@Home (or similar services like Rosetta) on my linodes to help out with this coronavirus work they're working on? Or does this voilate some terms of service by using too much vCPU resource?
2 Replies
jyoo
Linode Staff
You're absolutely allowed to run services such as Folding@Home, but due to the resource-intensive usage that comes along with these services, we require that they be run on a Dedicated CPU Linode. Dedicated CPUs are designed for resource-intensive projects; you wouldn't need to worry about your amount of CPU use, since you wouldn't be sharing your processor with other Linodes.
In the project I'm developing I'll need to run several computation instances with predictable CPU load (for some heavy signal processing and linear algebra). Instances will be clustered, so I'll be able to distribute workload between instances in any manner. But in case of some problems (for example, at the time of workload migrations) the load will temporarily increase for several seconds (or minutes) up to the limit.
So, what sustained CPU load I should define as a "base" load, leading to migration of tasks to another instance if load is higher?
Am I allowed to use, f.e., 80% of vCPU, without causing problems to other users of the host? Or 70%? Or some other part?
Is it possible to be "good well-behaving customer" while using nanodes and lower "standard instances" for such a workload? Or I'll be strongly required to use only Dedicated CPU plans?