Image Artist: What is the difference between the batch analysis job queues?

ImA is configured with 3 batch analysis job queues: ImA quick, ImA default and ImA debug. Each queue is configured with a "per task" resource reservation of CPU and RAM, which allows users to select the appropriate resource for the assay type. The table below details the "per task" resource allocations of each queue.

Each well within a plate is processed as an individual "task" during batch analysis execution. Given a 384 well plate, 384 image processing tasks will be executed. The number of tasks/wells which ImA is able to process in parallel is dependent on the ImA servers CPU and RAM reserves, as well as the job queue which is used to submit the job.

For example, given a ImA server with 10 CPUs and 50 GB of RAM:

• If the analysis job was submitted using the ImA default job queue (2 CPUs / 10GB per task), ImA could process 5 tasks/wells in parallel. Once the first 5 wells have been analysed, the next 5 wells will be processed, and so on. 
• If the analysis job was submitted using the ImA quick job queue (1 CPUs / 5GB per task), ImA could process 10 tasks/wells in parallel. Once the first 10 wells have been analysed, the next 10 wells will be processed, and so on. 

The ImA quick job queue requests less compute power than the ImA default queue, and because of this, allows for a greater number of wells to be processed in parallel. This queue is ideal for quickly processing basic image analysis pipelines, but may be under resourced for complex analysis tasks.

The ImA default queue requests more compute power than the ImA quick queue, and as such is better suited to general and complex image analysis pipelines.

For basic 2D analysis work the ImA quick queue may provide the best performance. For more complex 3D analysis and/or cell painting assays, the ImA default queue may be optimal.