On most publicly accessible clusters, only GPU-level power telemetry is available. This model takes a pragmatic approach to estimating total power draw by combining that telemetry with estimates for the remaining components and overheads.
Requires Python 3.12 or newer. From this directory:
python -m venv .venv
source .venv/bin/activate
python -m pip install -e ".[dev]"
python -m power_model --helppython -m power_model --gpu-level-power-per-gpu=400 --system=h100 \
--workload=agentic-cpu-offloading --scale-out-enabled --power-breakdown-per-chassis| Option | Meaning |
|---|---|
--gpu-level-power-per-gpu |
Required actual GPU electrical watts per GPU; underscore spelling also accepted |
--system |
Required system from the table below |
--model |
advanced (default) or basic-example |
--workload |
fixed-seq-len (default), agentic, or agentic-cpu-offloading |
--scale-out-enabled |
Activate NICs and external switches; alias --using-scale-out; default off |
--systems |
Advanced model quantity of the selected chassis or rack; default 1 |
--power-breakdown-per-chassis |
Advanced model nested BoM for one chassis or rack, followed by cluster totals |
--help |
List options, systems, and models |