Viewing Run (simulation) Results
Throughout AMP, "Simulations" that result in models for stars are referred to predominantly as Runs.
Basic details such as the subject star's name, coordinates, and identifiers, as well as the submitting scientist's name and time of submission, are displayed for each run.Each run progresses through several processing stages:
- Queued: The run has been submitted to the system but has not been processed. There is usually a short delay between run submission and the back-end system's discovery of the new run. In addition, optimization runs are queued for remote supercomputer capacity control. Because AMP runs are computationally intensive, the AMP website may only submit a limited number of runs to each remote supercomputer at any time. In this stage, the run is queued with other AMP jobs submitted on the AMP website.
- Pending: The run has been submitted to the remote supercomputer, but the computational task is waiting in the supercomputer's batch scheduling system queue. In this stage, the run is queued with all of the jobs from other users on the target supercomputer system.
- Running: The run is being executed on the remote resource.
- Done: The run has completed and results are now available.
- Failed: Something went wrong with the run! There is no need to panic -- your job probably just triggered an internal check that prevents nasty computer system error messages from being passed on to unsuspecting astronomers. A system staff member is notified when a job fails. They will investigate and correct the issue, and then resubmit the job for continued processing.
The input data for each run is displayed. For Direct Model runs, this includes the star's general model parameters and a target stellar age. For Optimization runs, the seismic and non-seismic observables are listed.
Output Data and Files
The output data for the simulation is displayed. For Direct Model runs, the large frequency separation is shown. For Optimization runs, the final model's parameters and the solution fitness are displayed. For all runs, an automatically generated Hertzsprung-Russell Diagram and Echelle Diagram are displayed.
Output files from the run are available for download. The file "results.tar.gz" package includes all of the data from the remote supercomputer execution. Note that the downloadable package contains binary files in either big-endian or little-endian format (denoted with .BE or .LE file name suffixes). Simulation result text files are listed separately for download.
Finally, the computational resource used to process the run is displayed. For Optimization Runs, the progress and solution of each of the genetic algorithm instances are shown separately, and all of the individual supercomputer jobs and their duration and computational cost are listed.