|
[Sponsors] |
June 25, 2015, 08:11 |
Parallel partition memory allocation problem
|
#1 |
New Member
Tomas
Join Date: Jun 2015
Posts: 3
Rep Power: 11 |
Hi there,
I have problem with the manual setup of the memory allocation for my computation. I'm getting errors like: IPUSHX: WORK IS OUT OF BOUNDS, INCREASE MXRVAR DATA TREE ENDS: 31695886 STACK: 30391549 EURANUS KILLED BY PROCESS 1 IPUSHX: WORK IS OUT OF BOUNDS, INCREASE MXRVAR DATA TREE ENDS: 31695886 STACK: 31311613 EURANUS KILLED BY PROCESS 30 I know how to prevent this in the case of the normal (blocks based) parallel metod, but partition parallel method allocates memory automaticaly. In the manual are described two parameters: -nbint # and -nbreal #. The problem is that even if I use them in the run script nothing changes. Run script: #$ -m ae #$ -M aa@aa.com #$ -v PATH #$ -cwd # VAR=bez_sani_rad SIM=000-SST MESH=_mesh/bez_sani_rad4.igg # # # # cat $PE_HOSTFILE | awk '{print $1" "$2}' > ./hosts2 X1=`cat hosts2 | sed -n 1p | awk '{print $2}'` X2=`expr $X1 - 1` sed 0,/"local "$X1/s/"local "$X1/"local "$X2/ hosts2 > hosts rm hosts2 hostname cat $PE_HOSTFILE echo "CouCou" # fine91_2 -batch -partition -computation $VAR\_$SIM/$VAR\_$SIM.run -machine_file ./hosts -nbint 600 -nbreal 6000 mv $VAR\_$SIM/$VAR\_$SIM.batch $VAR\_$SIM/$VAR\_$SIM.batch.orig cat $VAR\_$SIM/$VAR\_$SIM.batch.orig | sed s/euranusTurbo_parallel/"euranusTurbo_parallel -mca=btl=openib,sm,self"/g > $VAR\_$SIM/$VAR\_$SIM.batch chmod a+x $VAR\_$SIM/$VAR\_$SIM.batch # # date $VAR\_$SIM/$VAR\_$SIM.batch > $VAR\_$SIM/$VAR\_$SIM.LOG 2>&1 date Could anyone help me with this and say what is wrong? Thank you |
|
June 25, 2015, 12:47 |
|
#2 |
New Member
Tomas
Join Date: Jun 2015
Posts: 3
Rep Power: 11 |
Problem solved:
#$ -m ae #$ -M aa@aa.com #$ -v PATH #$ -cwd # VAR=bez_sani_rad SIM=000-SST MESH=_mesh/bez_sani_rad4.igg # # RELAX=1.700 # # cat $PE_HOSTFILE | awk '{print $1" "$2}' > ./hosts2 X1=`cat hosts2 | sed -n 1p | awk '{print $2}'` X2=`expr $X1 - 1` sed 0,/"local "$X1/s/"local "$X1/"local "$X2/ hosts2 > hosts rm hosts2 hostname cat $PE_HOSTFILE echo "CouCou" # fine91_2 -batch -partition -computation $VAR\_$SIM/$VAR\_$SIM.run -machine_file ./hosts mv $VAR\_$SIM/$VAR\_$SIM.batch $VAR\_$SIM/$VAR\_$SIM.batch.orig cat $VAR\_$SIM/$VAR\_$SIM.batch.orig | sed s/euranusTurbo_parallel/"euranusTurbo_parallel -mca=btl=openib,sm,self"/g > $VAR\_$SIM/$VAR\_$SIM.batch chmod a+x $VAR\_$SIM/$VAR\_$SIM.batch # mv $VAR\_$SIM/$VAR\_$SIM.run $VAR\_$SIM/$VAR\_$SIM.run.orig cat $VAR\_$SIM/$VAR\_$SIM.run.orig | sed -e "s/INTEGER_RELAXATION 1.5000/INTEGER_RELAXATION $RELAX/" > $VAR\_$SIM/$VAR\_$SIM.run mv $VAR\_$SIM/$VAR\_$SIM.run $VAR\_$SIM/$VAR\_$SIM.run.orig cat $VAR\_$SIM/$VAR\_$SIM.run.orig | sed -e "s/REAL_RELAXATION 1.2000/REAL_RELAXATION $RELAX/" > $VAR\_$SIM/$VAR\_$SIM.run chmod a+x $VAR\_$SIM/$VAR\_$SIM.run # date $VAR\_$SIM/$VAR\_$SIM.batch > $VAR\_$SIM/$VAR\_$SIM.LOG 2>&1 date Last edited by Jepice; June 25, 2015 at 16:53. |
|
July 28, 2017, 18:20 |
|
#3 |
New Member
Join Date: Apr 2017
Posts: 1
Rep Power: 0 |
Hi Jepice.
I have the same problem at the moment. Can you tell me what exactly did you do in your .run file? I mean, I see you deleted nbint and nbreal in one line and added some lines. What exactly is it good for? In my case I tried to increase "Memory_Estimation_Real" and "Memory_Estimation_Integer" but that didn't help. It looks like this in my case: .run file: NUMECA_PROJECT_FILE VERSION 1110 PROJECT_TYPE STRUCTURED COMPUTATION_NAME <some name> SUB_PROJECT_FLAG 0 GRID_FILE /<some name>/<some name>/<some name>.igg RELATIVE_GRID_FILE _mesh/<some name>.igg TRB_FILE /usr/unknown RELATIVE_TRB_FILE undefined GEOMETRY_FILE /usr/unknown RELATIVE_GEOMETRY_FILE undefined SOLVER_PRECISION single PERFORMANCE_CURVE_AXIS 0 LAST_USED_COMPUTATION_INDEX 0 ************************************************** ********* ******* START COMPUTATION 1 SECTION ************** ************************************************** ********* NI_BEGIN computation VERSION 1110 NAME <computation name> POINTER 30885456 COARSE_FLAG 1 1 1 NUMBER_OF_GRID_LEVEL 3 RELATIVE_REFERENCE_RUN_FILE undefined POSITION -1 ACTIVE 0 NON_DETERMINISTIC_ACTIVE 0 generalisedTransitionModel 0 GRID_LEVEL_I 0 GRID_LEVEL_J 0 GRID_LEVEL_K 0 Efficiency 1 Efficiency_mean 1 Efficiency_variance 1 Perf_curve_Pressure 1 Perf_curve_Pressure_mean 1 Perf_curve_Pressure_variance 1 PressureRatio 1 PressureRatio_mean 1 PressureRatio_variance 1 Mass_flow 1 Mass_flow_mean 1 Mass_flow_variance 1 uncertainty_weight_factor 1 uncertainty_grid_type 1 AXISYMMETRIC no AXISYMMETRIC_AXIS 0 0 1 AUTO_GRID_MESH no START_FROM_INITIAL_SOLUTION_FILE no RESET_CONVERGENCE_HISTORY yes INITIAL_SOLUTION_FILE unknown INITIAL_RESIDUAL_FILE unknown FLUID_NAME AIR(Perfect) FLUID_TYPE Perfect Gas STEAM_TABLES_PATH unknown LAST_USED_FLUID_RECORD 0 TIME_STEPS_CHOICE 0 SWEEPS_DEFINITION 0 USER_SOLVER_PRECISION 0 MULTI_RANK_UPDATE 1 MAXIMUM_RANK 1 MULTIRK_PARTIAL 0 MULTI_RANK_INTERACTION 0 MAX_ORDER_INTERACTION 1 MAXIMUM_HARMONICS_NUMBER 0 ALL_HARMONICS_ON 1 NB_HARMONICS_PER_RANK 3 NB_HARMONICS_PER_RANK 3 MEMORY_ESTIMATION_REAL 104047572303 MEMORY_ESTIMATION_INTEGER 104047572303 SUBTASK_STATUS pending RUNNING_MODE STOPPED ...and so on. It would be nice if you could help me. Thank you and greetings. |
|
|
|
Similar Threads | ||||
Thread | Thread Starter | Forum | Replies | Last Post |
Problem with parallel computation (case inviscid onera M6) | Combas | SU2 | 11 | January 30, 2014 02:20 |
problem in the CFX12.1 parallel computation | BalanceChen | ANSYS | 2 | July 7, 2011 11:26 |
Fluent Parallel Memory Usage | David Christopher | FLUENT | 1 | March 3, 2011 03:20 |
Problem on Parallel Run Setup | Hamidur Rahman | CFX | 0 | September 23, 2007 18:11 |
Dynamical memory allocation for multiblock? | peter.zhao | Main CFD Forum | 2 | May 20, 2002 05:58 |