PBS Professional

Migrate from Grid Engine to PBS Professional


PBS Professional is a cost effective, commercial-grade HPC workload and resource management solution, which scales to support hundreds of thousands of processors – from clusters to the largest HPC systems – ensuring the maximum value from computing investments is received.



Benefits for migrating to PBS Professional:




  • Petaflops Scalability + Reliability

  • Topology-aware Scheduling

  • Supports all MPI Integrations

  • License Management Support

  • GPU Scheduling

  • Green Provisioning

  • OS Provisioning

  • EAL3+ Security Certified

  • Backfill TopN Scheduling




  • Estimated Job Start Times

  • Tunable Scheduling Formula

  • Standing + Advance Reservations

  • Keeps Job History

  • Python Hooks

  • Cost Effective

  • Global Support

  • Partner Alliance Program












































































































Capability
Grid Engine
PBS Professional
Business Comparison    
Security Certifications No

Yes; globally recognized
EAL3+ certification

Scalability

Max: 64k cores

Unlimited cores

Company Split: Oracle vs. Univa

Global with local presence; 25+ yrs. profitable

Support

U.S. & Germany

Local experts in 18+ countries & global 2nd level escalation support

Business Model

Middleware only;

traditional licensing

Middleware, business intelligence apps, & optimization apps all enabled by HWUs tokens (including partner apps!)

Technical Comparison    

Scheduling policies

All All

Managed resources for hardware, GPUs, licenses, Green Provisioning

No; hardware & green provisioning only

Yes

Single, powerful formula for scheduling complexity

No; requires mixing shares, functional,

& overrides

Yes

Backfill optimization

Yes

Yes; plus backfill depth

Topology scheduling (inside, network, site defined)

No; inside node only

Yes

Estimated start times

No Yes

Individual & recurring reservations

No; individual only

Yes

Job filtering (submit, modify, move, etc.)

No; submit only

Yes

All MPI integrations,

including full accounting

No

Yes

NVIDIA CUDA SDK support

No

Yes

OS Provisioning

No

Yes; built-in with per-job flexibility

Workload Analytics

Yes

Yes


Bookmark and Share

Subscribe to join our Newsletter
Learn about product training, news, events and more.