Fabiano Corsetti, April 2014
f.corsetti {at} nanogune.eu
This web app provides a quick estimate of the walltime needed to run SIESTA on a modern HPC system. To use it, simply fill in the details of your simulation in the boxes below, select the architecture which is most similar to the one you are planning to run on, and click the Submit button to see the results.
A few important things to note:
- The results are for Γ-point simulations using SIESTA's default diagonalization solver. They are based on timing tests performed using the system's native compilers and optimized libraries (most importantly, BLAS).
- The estimates provided are very rough. This has to be the case, since in reality there are very many factors (both predictable and unpredictable) that affect performance.
- The y-axis of the output plot is fixed to a sensible range. If your simulation is estimated to take less than 6 minutes or more than ∼1 year, you will not see any trendline, and a warning message will be printed.
- Memory requirement estimates per core between 1 and 16 GB are also shown (when in range) by labelled horizontal lines; these are calculated for SIESTA's default options, and should be taken as lower bound estimates.
For more information on how these estimates are obtained, see:
F. Corsetti, Performance analysis of electronic structure codes on HPC systems: A case study of SIESTA, PLOS ONE 9, e95390 (2014). DOI: 10.1371/journal.pone.0095390 [LINK]
Please cite this paper if you have used the app for your research!