Main menu

Navigation

Implicit error when changing the PE_LAYOUT

3 posts / 0 new
Last post
xyli@...
Implicit error when changing the PE_LAYOUT
Dear Madam/Sir, After I changed the PE_LAYOUT, the job was terminated due to unknown reasons. Any idea about this? Attached please kindly find the error message reprot. Here is the procedure I followed ./create_newcase --case ~/cases/FW1850_PE --compset FW1850 --res f09_f09_mg17 --machine tetralith ./xmlchange NTASKS_ATM=128 ./xmlchange NTASKS_CPL=2 ./xmlchange NTASKS_LND=1 ./xmlchange NTASKS_ICE=1 ./xmlchange NTASKS_OCN=1 ./xmlchange NTASKS_ROF=1 ./xmlchange NTASKS_GLC=1 ./xmlchange NTASKS_WAV=1 ./case.setup ./case.build ./case.submit Best regards, Xiang-Yu Li

Xiangyu Li

jedwards

The FW case at f09_f09_mg17 is a rather large model, along with the atm component it includes a land model and a coupler.  You are specifying way too few processes for those components.   I would suggest changing all components to have 120 tasks:

./xmlchange NTASKS=128

and trying again.   That should work but is still very small, if you have access to more processors you may want to use them, try 1024 for example. 

 

CESM Software Engineer

xyli@...
Thanks a lot, Jedwards. I will eventually push to 1024 CPUs for the production run. For now, I am playing around with the optimum PE-LAYOUT. I checked the Run Time for each component (CPL, ATM, LAND, ...) and found that the ATM module takes up most of the CPU time. So I designed a PE-LAYOUT according to the following Run Time statistics. It seems that assign 1024 CPUs for all the modules will waste a lot of CPU resources? TOT Run Time: 27680.976 seconds 892.935 seconds/mday 0.27 myears/wday CPL Run Time: 14.893 seconds 0.480 seconds/mday 492.72 myears/wday ATM Run Time: 27405.343 seconds 884.043 seconds/mday 0.27 myears/wday LND Run Time: 126.022 seconds 4.065 seconds/mday 58.23 myears/wday ICE Run Time: 5.684 seconds 0.183 seconds/mday 1291.01 myears/wday OCN Run Time: 0.430 seconds 0.014 seconds/mday 17053.84 myears/wday ROF Run Time: 10.341 seconds 0.334 seconds/mday 709.61 myears/wday GLC Run Time: 2.768 seconds 0.089 seconds/mday 2651.04 myears/wday WAV Run Time: 0.000 seconds 0.000 seconds/mday 0.00 myears/wday ESP Run Time: 0.000 seconds 0.000 seconds/mday 0.00 myears/wday CPL COMM Time: 384.699 seconds 12.410 seconds/mday 19.07 myears/wday

Xiangyu Li

Log in or register to post comments

Who's new

  • praveenmaniyatt@...
  • arjunbabun11@...
  • lama@...
  • sisi393@...
  • 1658093099@...