Running Concurrent Workloads

You can run multiple applications concurrently using the CLI in two ways:

Running Applications in the Same Pool

When you have a pool with sufficient GPU resources, you can run multiple applications using the same pool:

juice run [pool-id] application1
juice run [pool-id] application2

For example:

juice run pool-1 python train.py
juice run pool-1 jupyter lab

Running Applications Across Different Pools

When you need to distribute your workload across multiple pools (e.g., when no single pool has enough GPUs):

juice run [pool-id-1] application1
juice run [pool-id-2] application2

For example:

juice run pool-1 python train.py
juice run pool-2 jupyter lab

tip

Use juice pool list to see available pools and their IDs.