Supports SLURM, Openstack, k8s

CEEMS was designed to be resource manager agnostic. Although thoeritically it can support many resource managers, we are focusing to support SLURM, Openstack and Kubernetes deployed on baremetal.

Uses eBPF for perf metrics

Besides energy and carbon footprint, CEEMS monitors and reports performance, IO and network metrics for user workloads using eBPF.

Supports Continuous Profiling based on Grafana Pyroscope

CEEMS is capable of eBPF based continuous profiling of compute workloads.

ML/AI workloads

CEEMS supports energy monitoring of both NVIDIA and AMD GPUs allowing to quantify the energy consumption of your AI workloads.

Realtime emissions

CEEMS uses realtime emission factors to estimate the emissions of your workloads. It is also possible to include emboided emissions data, if and when available.