196 | | specify a multiple of 8 for MPI in config.card and increase nodes and ppn in the Header of Job_. |
197 | | see here howto use more processors (info for all machines): |
| 199 | example for a global simulation and 64 processors |
| 200 | config.card you told libIGCM to run ORCHIDEE with one execution unit (63MPI and xios with 1MPI): |
| 201 | {{{ |
| 202 | OL= (orchidee_ol, orchidee_ol, 64MPI) |
| 203 | IOS= (xios_server.exe, xios.x, 1MPI) |
| 204 | }}} |
| 205 | |
| 206 | {{{ |
| 207 | ###################### |
| 208 | ## OBELIX LSCE ## |
| 209 | ###################### |
| 210 | #PBS -N test |
| 211 | #PBS -m a |
| 212 | #PBS -j oe |
| 213 | #PBS -q longp |
| 214 | #PBS -o Script_Output_test.000001 |
| 215 | #PBS -S /bin/ksh |
| 216 | #PBS -v BATCH_NUM_PROC_TOT=64 |
| 217 | #PBS -l nodes=8:ppn=8 |
| 218 | }}} |
| 219 | |
| 220 | see here more info how to use parallel processors (info for all machines): |