Managed large-scale parallel and high-performance computing (HPC) job execution.
Choose Azure Batch for large-scale parallel and high-performance computing (HPC) — rendering, simulations, Monte Carlo analysis, media transcoding, or any embarrassingly parallel workload. It schedules jobs across a pool of VMs that scale up for the work and back to zero when done.
Avoid Batch for always-on services, web hosting, or low-latency request/response workloads — it's a job scheduler, not an application host. For long-running services use App Service, AKS, or Container Apps; for event-driven short tasks use Functions.
| Hosting model | Managed (scheduled jobs) |
|---|---|
| Container support | Container-compatible |
| Minimum nodes | 1 (scales to zero after the job completes) |
| State management | Stateless |
| Web hosting | No |
| Autoscaling | Not applicable |
| Load balancer | Azure Load Balancer |
| Scale limit | 900 dedicated cores and 100 low-priority cores (default) |
| Multiregion | Single region only |
| Virtual network integration | Supported |
| Hybrid connectivity | Supported |
| GPU support | Supported |
| TLS | Supported |
| Architecture styles | Big compute (HPC) |
| Required skills | Job scheduling, parallel processing |
| Operational overhead | Medium — job and pool management |
| Best for teams | Teams that run HPC or batch workloads |
Batch targets large-scale parallel and HPC workloads — rendering, simulations, financial modeling, and media processing — split into many tasks that run concurrently across a VM pool.
Yes. A Batch pool can scale its compute nodes down to zero between jobs with autoscale formulas, so you pay for compute only while jobs run.
Yes. Batch can run tasks inside containers on its pool nodes, combining container packaging with Batch's job scheduling and scale.
This is one candidate from the Azure compute decision guide. The right choice depends on your full requirements — evaluate scaling, cost, and operational fit before committing.
Browse all tools →