A simulation that is too large or slow for your workstation can bring an important project to a halt. Cloud computing can move heavier GROMACS stages off your local machine, but a remote job introduces its own questions: Is the setup ready? What will it cost? Where will the results go, and how do you bring them back into your workflow?
A little preparation can make the difference between a useful cloud run and an avoidable round trip. Here is a practical way to think through the process in SAMSON’s GROMACS Wizard.
Validate before spending credits
Cloud jobs are available for NVT equilibration, NPT equilibration, and production molecular dynamics. Before submitting one, define the input structure and simulation parameters using the relevant stage’s local workflow. If your computer can handle it, run a short local test—just a few tens or hundreds of steps—to catch obvious setup problems.
If the system is too large to run locally, use Generate inputs instead. It can validate part of the setup and reveal common issues before you launch a paid computation. This is a useful middle ground: you can check more than the settings on screen without requiring your machine to run the full system.
Know what the credit check means
Check that your SAMSON Connect account has at least one computing credit before starting. You can view your balance in SAMSON through Interface > Job manager, then Show credits, or by signing in to SAMSON Connect and opening My account.
One credit is the minimum balance required to start a cloud job; it is not a promise that the job will cost exactly one credit. Review the expected price in the machine-selection and confirmation dialogs. The machine and storage are charged, while file transfer is free. You can also open the machine details to compare available resources and pricing.
For testing, the tutorial recommends a machine with four vCPUs and no GPU: it is the slowest, least expensive option and is suitable for testing. That is a sensible starting point when your goal is to check the submission and results workflow, rather than to maximize performance. Choose resources with the simulation’s purpose in mind, and confirm the price before submitting.
Follow the job from submission to completion
After choosing the machine and storage, review the security dialog. It summarizes the job and expected pricing, shows your remaining credits, and lets you edit the job name, description, and notes. Clear names and notes can help you distinguish runs later, especially when several stages or parameter sets are involved.
Once submitted, the Job manager tracks initialization and execution. Enable the Running filter to see jobs being initialized or run. When initialization finishes, confirm the prompt to start the job; the status should progress from Starting to Running. The Events view provides messages, including an estimated finish time. Treat that time as an estimate: it can be updated.
You do not need to keep SAMSON open while the computation runs. The job remains accessible through the Job manager the next time you launch SAMSON. You can pause or cancel a running job from its context menu, and restart a paused job later.
Bring back the files you need
When the job is complete, check its status and review the event messages. To retrieve the output, open the job’s result window from the Job manager. Choose a download folder, then download individual files, a selection, or all files. Downloading all results is recommended because multiple files may be needed to process a trajectory or import the results into SAMSON.
After downloading, use the folder path in the appropriate GROMACS Wizard stage’s Import (*) results action. Select the results folder and confirm the import; depending on the stage, SAMSON may also ask about plotting or trajectory-import options. A resulting GRO file can also be used as the input structure for a subsequent step.
Close the loop on storage
Downloaded files remain available in cloud storage until you delete the job. Storage costs are low, but they can begin to be deducted from your credits after some time. Once you have confirmed that the results are downloaded and no longer need the cloud copy, review the storage cost in the job details and delete the job if appropriate. In other words: verify your local results first, then clean up the remote copy.
This gives you a simple routine: validate inputs, check the estimate, monitor events, download the full result set, import it into the next workflow step, and remove cloud files when they are no longer needed. For detailed instructions and screenshots, read the original GROMACS Wizard cloud-computing documentation.
Note: SAMSON and all SAMSON Extensions are free for non-commercial use. Get SAMSON.