noel
noel
Is it known what this error message means? Using GCP, the following test fails `SMS_D_Ld1.ne30pg2_EC30to60E2r2.WCYCLSSP370.gcp_gnu.allactive-wcprodssp` But the file in message seems OK ``` 0: PIO: FATAL ERROR: Aborting... FATAL ERROR:...
Using master of July 13th, I see a MEMLEAK with `ERP_Ld9.ne4_ne4.FIDEAL.pm-cpu_gnu`. ``` memory_write: model date = 00010101 0 memory = 825.01 MB (highwater) 83.84 MB (usage) (pe= 0 comps= cpl...
Using GCP, I've hit a similar error with a few tests. Trying same test on Cori (with GNU), I don't see the problem. Here are the test names that have...
On GCP (which is using gnu 8.5) I see a COMPARE fail with `ERP_Lm3.ne4_oQU240.F2010`. It also triggers a MEMLEAK fail. With DEBUG, it passes COMPARE, but does then fails with...
The machines I've tried this test with GNU have all failed compare. The DEBUG version passes. I've tested with master of May6th on cori-haswell (with GNU v10.3): `/global/cscratch1/sd/ndk/e3sm_scratch/cori-haswell/ERS.ne4_oQU240.F2010.cori-haswell_gnu.eam-hommexx.20220517_122134_hut7b6` Also same...
I'm trying to run with the following: ``` set compset=WCYCL1950 set mach=chrysalis set res=ne120pg2_r0125_oRRS18to6v3 ``` I'm trying to run on pm-cpu, which only has GNU. I get runtime errors, so...
Reviving older issue with `SMS_D.ne4pg2_ne4pg2.F2010-CICE.pm-cpu_nvidia`. After adding the `pm-cpu` machine and updating nvidia compiler version, this issue still fails with out of range, but different array: ``` 0: 0: Subscript...
Trying this test with DEBUG=TRUE, I see an error. `SMS_D.f09_g16.I1850ELMCN.cori-knl_gnu.elm-bgcinterface` ``` 22: At line 288 of file /global/cscratch1/sd/ndk/wacmy/m58-aug26/components/elm/src/main/elm_interface_funcsMod.F90 22: Fortran runtime error: Array bound mismatch for dimension 1 of array...
Using the nvidia compiler on CPU-only nodes of Perlmutter (`pm-cpu`), most tests pass, but these fail compare: ``` ERS_Ld3.ne4pg2_oQU480.F2010.pm-cpu_nvidia.eam-thetahy_sl_pg2 ERS_Ld3.ne4pg2_oQU480.F2010.pm-cpu_nvidia.eam-thetahy_sl_pg2_ftype0 ``` I see that this issue has similar fails https://github.com/E3SM-Project/E3SM/issues/4768...
One of the 4 ERS tests in the MMF suite (`e3sm_mmf_integration`) fails compare on CPU with nvidia compiler. `ERS_Ln9.ne4pg2_ne4pg2.F2010-MMF1.pm-cpu_nvidia.eam-mmf_crmout` All of those tests pass with `gnu` compiler. These ERS MMF...