/ getting started

cryoSPARC software highlights

Aside from algorithmic advancements, cryoSPARC offers several features that can help you get the most out of cryo-EM data processing.

cryoSPARC queued job example

Centralized access to data

cryoSPARC facilitates the ability of multiple users in a lab or institution to access cryo-EM data in a centralized manner. We recommend configuring your data input directories when installing cryoSPARC. This allows a GPU machine running cryoSPARC to directly read and save data to a cluster file system, network storage drive, or other centralized data storage solution, without users needing to worry about copying files or results over a lab network manually.

SSD caching

With increasing dataset sizes, loading particle stacks during processing from spinning hard drives can lead to unnecessary bottlenecks in processing time. For this reason, cryoSPARC offers an SSD caching feature to allow you to make the most of your time by locally storing data for future use.

The first time you load and Visualize a New Dataset, the built-in caching layer will automatically copy data onto your local SSD and remove old cached data when space is required. Files relating to active experiments and those that are accessed most frequently will remain in the SSD cache to ensure optimal performance. This is especially beneficial for users who want to quickly run multiple experiments on a single dataset, e.g., to identify heterogeneity or multiple conformations.

To enable SSD caching, you will need to know a path to a location that resides on your local SSD. For more information, please visit the installation instructions in our Guide.

Automatic job scheduling

CryoSPARC's built-in job scheduler works alongside the SSD caching feature to ensure optimal performance in labs with multiple experiments and/or multiple users. If no GPUs are available when you launch a New Experiment, cryoSPARC will automatically queue and commence the experiment when GPU(s) become available, eliminating the need for manually coordinating or following up on experiments. The queueing system is 'intelligent' and takes into account the type of experiment started, the amount of required CPU RAM, and available system resources when scheduling jobs.


Questions or feedback? Please drop us a line at feedback@structura.bio.