What are Core-Hours? How are they estimated?

A core-hour is a measurement of computational time. In OnScale, if you run one CPU for one hour, that's one core-hour. If you run 1000 CPUs for 1 hour, then that's 1000 core-hours.




Users will be spending core hours to run their simulations and design sweeps so we feel it is necessary to provide users with a estimation of how long this may take. The number of core-hours spent per job however can be varied due to a number of different factors e.g. model size (number of elements), model inputs/outputs and simulation type.

An estimate based on the model file you submit to the cloud is used as the baseline core-hour usage a single simulation. When a user runs hundreds of simulations, the estimation is extrapolated linearly based on the initial estimate value. Note that parametric variable settings can have an effect on model size which the estimate will not take into account.

Tip: Default model variables to run for the largest simulation which will allow the estimator to give an idea of 'worst case' core-hour cost.