Right-size resources
Set CPU, memory, GPU, environment, and scheduler parameters from reusable defaults, expressions, and runtime rules.
Galaxy resource intelligence
Total Perspective Vortex
TPV is a Galaxy-integrated Python library for right-sizing job resources and meta-scheduling work across heterogeneous compute destinations using transparent YAML rules.
What TPV does
Policy-driven job routing for Galaxy administrators.
TPV sits at the Galaxy job dispatch layer. It combines tool, user, role, and destination rules to choose suitable resources and an execution target before the backend scheduler takes over.
Route across destinations
Match jobs to Slurm, Pulsar, local runners, clouds, or other configured destinations based on capacity and policy.
Keep policies transparent
Express routing decisions through YAML entities, inheritance, tags, and auditable Python expressions when more control is needed.
How it is used
TPV is running across the usegalaxy federation.
Galaxy Australia, Europe, and the United States use TPV concepts in production to reduce over-allocation, route jobs through heterogeneous infrastructure, and make shared compute allocations go further.
- For researchersJobs can move through the system with fewer queue delays and fewer resource-related failures, without changing the Galaxy interface.
- For administratorsCommon tool requirements are inherited once, then adjusted locally only where a site has different hardware or policy constraints.
- For infrastructure providersTPV helps avoid over-allocation and supports more efficient use of shared HPC, cloud, and institutional resources.
Shared resource database
Community defaults reduce repeated local tuning.
The TPV shared resource database captures empirically derived defaults for common bioinformatics tools. Galaxy sites can import the database, inherit maintained recommendations, and layer local overrides for site-specific destinations, limits, and policies.
Curated centrally
Defaults are maintained in version control by the Galaxy community and informed by production use and targeted benchmarking.
Overridden locally
Sites can clamp resources to local hardware, add destination-specific parameters, or override any inherited rule.
Improved collaboratively
Better rules can move from one deployment back into the shared database, making production lessons reusable across Galaxy.
Learn more
This page is the overview. These are the canonical resources.
Read the DocsInstall TPV, configure Galaxy, and learn the full YAML model.GitHub sourceBrowse the Python package, tests, issues, and release history.Shared resource databaseUse community-curated defaults for common bioinformatics tools.Galaxy Training NetworkFollow the administrator tutorial for Galaxy job destinations.Published paperRead the TPV paper on resource allocation and production results.Galaxy ProjectLearn about the open platform TPV was built to support.