adamrhster@gmail_com
New Member
Hi all,I'm running cesm1_2_2 with an aquaplanet compset (2000_CAM5_SLND_SICE_AQUAP_SROF_SGLC_SWAV) using the spectral element dycore at ne60 (ne60np4_ne60np4). I realize that this is not a supported configuration, but I am hoping that someone would be willing to chime in, and help me trouble shoot this error. During runtime, but prior to writing out the first time-step to history (the simulations doesn't even get that far), I receieve the following error in the cesm.log (which is also attached): 413: singular matrix in gaussj 2 413: 1 1 NaN NaN 413: 1 2 NaN NaN 413: 2 1 NaN NaN 413: 2 2 NaN NaN 413:(shr_sys_abort) WARNING: calling shr_mpi_abort() and stoppingThis error is not related to any initial condition files I supplied. The error is also not related to unstable time-stepping (based on sensitivity runs). I have sucessfully run ne16, ne30 and ne120 configurations, and have never rec'd this error. This error occurs for the ne60 configuration only, and in three different versions of cam (cesm1_2_2, cesm1_2_1, developmental cam5_4_39). I've also included the atm_in and drv_in files.If anyone has any clue where to look to trouble-shoot this 'gaussj' error, I would appreciate it!Adam