Quotas

Jetstream2 tries to apportion the memory, storage, and network available to researchers based on both allocation type and Jetstream2 resource node type.

Limits

Jetstream2 quotas

Please note that these quotas are currently under review as we determine how to make them best work with the new allocation types. We will update this page once we do so.

Quotas are a function of: allocation and resource

allocation type:

Startup/Research	Education	Jetstream Trial Allocations (JTA)
default allocations for collaborative users in a common research project.	special allocations for teaching and workshops. Quotas are set slightly higher because of typical larger user counts	Sample allocations with minimal capacity quotas designed for evaluation purposes

resource type:

CPU (compute)	GPU	LM (large memory compute)
normal bulk compute resources with higher available core counts	GPU accelerated resource with additional ceph storage	2X-memory compute resource with additiona ceph storage
75% of JS2	19% of JS2	6% of JS2
3.9 GB per RAM per core	7.8 GB RAM per core	3.9 GB RAM per core

STARTUP/RESEARCH	defaults *
CPU	GPU	LM
150 cores 585 GB RAM 10 floating IPs	64 cores 250 GB RAM 4 floating IPs	128 cores 1000 GB RAM 2 floating IPs

EDUCATION	defaults *
CPU	GPU	LM
200 cores 780 GB RAM 25 floating IPs	320 cores 1250 GB RAM 25 floating IPs	128 cores 1000 GB RAM 2 floating IPs

JTA	defaults
CPU	GPU	LM
2 cores 6 GB RAM 2 floating IPs	none	none

* modifications available with proper documentation

Other Limits

Jetstream2 policies may also impact quotas, particular regarding networking.