Overview
Running your first batch job on Rescale requires understanding key platform fundamentals and following specific best practices to ensure success. This guide provides essential considerations for transitioning from local simulations to cloud-based batch processing.
Prerequisites
Before submitting your first batch job, ensure you understand the basic workflow by reviewing our comprehensive documentation: How to Run a Basic Job.
Since batch jobs primarily run on Linux environments (unless otherwise specified), familiarize yourself with Linux file system conventions and limitations.
Essential First-Run Guidelines
Understand Your Simulation
Before running any simulation on Rescale, ensure you:
Fully understand the type of simulation you're running
Have successfully executed the simulation locally (on-premise) at least once
Are familiar with the expected inputs, outputs, and runtime behavior
Start with Familiar Configurations
For your initial runs:
Use the same number of cores as your successful local runs, or as close as possible
Once the job runs successfully, then experiment with scaling to optimize performance
Run jobs in configurations as familiar as possible to minimize potential sources of error
Software Version Compatibility
Ensure that the software versions available on Rescale match those used locally to create your input files. The only exception is when using a higher version that supports backwards compatibility with your input files.
File Naming Requirements
Critical for Linux compatibility: Input files must not contain spaces in their names. While Windows applications commonly allow spaces in filenames, Linux systems (which power most batch jobs) do not handle spaced filenames by default.
Before uploading files:
Remove all spaces from filenames
Use underscores (_) or hyphens (-) instead of spaces
Verify all referenced files in your input scripts use the corrected names
Recommended Workflow
Prepare locally: Ensure your simulation runs successfully on your local system
Check compatibility: Verify software versions and file naming conventions
Start conservative: Use familiar core counts and configurations
Submit and monitor: Run your first job with close attention to logs and outputs
Optimize gradually: Once successful, experiment with performance improvements
Next Steps
After your first successful batch job:
Review job performance metrics to identify optimization opportunities
Experiment with different core types and counts for better performance
Explore advanced features like job templates and automation
Consider setting up monitoring and notification preferences
Following these guidelines will help ensure a smooth transition to cloud-based simulation and set the foundation for successful ongoing use of the Rescale platform.