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

what should I do if I use more cpus after running 2 years?

W

wfc1102@163_com

New Member
Here is my model setting:EXP1 is a hybrid B20TRC5CN run based on/starting from b40_20th_1d_b08c5cn_139jp 1979-01-01 using CESM1.2.1.With 144 tasks,  it finished sucessfully after 2 years (STOP_N=2, STOP_OPTION=nyears). Now, there are 288 tasks available on the sever now and I want my experiment continue to run extra 10 years. What need I do? Here is my 2 thoughts, but not sure which is right/better: 1) If cpus changed as below, the model need be rebuilded besides (xmlchange CONTINUE_RUN=TRUE, STOP_N=10, cesm_setup -clean; cesm_setup;).------pp=288./xmlchange -file env_mach_pes.xml -id NTASKS_ATM -val $pp ./xmlchange -file env_mach_pes.xml -id NTASKS_LND -val $pp ./xmlchange -file env_mach_pes.xml -id NTASKS_ICE -val $pp ./xmlchange -file env_mach_pes.xml -id NTASKS_OCN -val $pp ./xmlchange -file env_mach_pes.xml -id NTASKS_CPL -val $pp ./xmlchange -file env_mach_pes.xml -id NTASKS_GLC -val $pp ./xmlchange -file env_mach_pes.xml -id NTASKS_ROF -val $pp ./xmlchange -file env_mach_pes.xml -id NTASKS_WAV -val $pp------2) create a new case, EXP2, based on/starting from EXP1 1981-01-01. RUNTYPE=branch, pp=288.------ Appreciate any suggestions or comments. 
 
J

jedwards

CSEG and Liaisons
Staff member
Method one would work fine but in a B case that isn't generally the best performing configuration.assuming you have 32 tasks per node I would do something like this:NTASKS_ATM=NTASKS_CPL=256NTASKS_LND=NTASKS_ROF=128NTASKS_ICE=128, ROOTPE_ICE=128NTASKS_OCN=32, ROOTPE_OCN=256 this will allow the ocn to run concurently with atm and should improve overall performance. 
 
J

jedwards

CSEG and Liaisons
Staff member
Method one would work fine but in a B case that isn't generally the best performing configuration.assuming you have 32 tasks per node I would do something like this:NTASKS_ATM=NTASKS_CPL=256NTASKS_LND=NTASKS_ROF=128NTASKS_ICE=128, ROOTPE_ICE=128NTASKS_OCN=32, ROOTPE_OCN=256 this will allow the ocn to run concurently with atm and should improve overall performance. 
 
You must log in or register to reply here.
Top