Scheduled Downtime
On Tuesday 24 October 2023 @ 5pm MT the forums will be in read only mode in preparation for the downtime. On Wednesday 25 October 2023 @ 5am MT, this website will be down for maintenance and expected to return online later in the morning.
Normal Operations
The forums are back online with normal operations. If you notice any issues or errors related to the forums, please reach out to help@ucar.edu

Performance of multi-instance runs

Hi,I did a series of single-instance and multi-instance runs on Titan at Oak Ridge. For the same kind of simulations, the multi-instance runs cost significantly more core hours. Here are some numbers from the cesm_timing.xxxxx files: -------------------------------------------------------------------------------                                   Single-instance        Multi-instance                                   (1x2000 cores)      (32x1888 cores)                                       Wallclock                WallclockDRIVER_INIT                   61.501438           344.224579 (Called = 2)DRIVER_RUN_LOOP      1219.016846         1695.854736DRIVER_CPL_RUN            24.72876            134.189667DRIVER_CPL_COMM         27.324247          529.453613--------------------------------------------------------------------------------Compset: FC5PE layout: same number of processes for all active components;                same number of instances (1 or 32) for atm, lnd, rof, ice, ocn;                one instance for cpl,                one instance for glc and wav which are stub components.Code: ACME code which branched from CESM at 1_3_beta10. I'd like to ask whether similar behavior has been seen with CESM, and what I could do to improve the performance. Thanks in advance.Hui
 
Top