2019 05 23 sutrams - hydrogeologger/pyduino GitHub Wiki
layout: post title: sutrams issues
problem: segmentation fault
Opening input files...
Using AP Column Storage Routine
Reading parameters from ET.INP...
Initializing input data...
Reading Data Sets 1-3...
Initializing initial SUTRA storage...
Allocating hydraulic parameter zones and reading data if zone file is specified...
Reading Data Set 4...
Reading Data Sets 5-7...
Reading Data Sets 8-15...
Reading Data Set 22 (Mesh Connection Data)...
Calculating Problem Bandwidth...
Allocating solver storage and making AP column storage format...
Reading ICS data...
Allocating storage for Tecplot output...
Calculating and output of total storage allocated...
Calling main SUTRA subroutine...
Outputting initial/starting conditions to fLST...
Allocation/Initialization complete...Beginning Time Steps
---------------------------------------------------------
TIME STEP 1 OF 10000: ITERATION 1
TOTAL TIME: 1.0000 HRS
DELT : 1.0000 HRS
---------------------------------------------------------
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
#0 0x7e2d9b3c3d4a
#1 0x7e2d9b3c2f7d
#2 0x7e2d9a8e205f
#3 0x55a47862fc6d
#4 0x55a47860d5ed
#5 0x55a4785f30b8
#6 0x55a4785c2d3e
#7 0x7e2d9a8cf2e0
#8 0x55a4785c2d79
#9 0xffffffffffffffff
Segmentation fault
attempt:
run
uqczhan2@goliath:~/macondo/AP2Dtide1m_stack$ ulimit -a
core file size (blocks, -c) 0
data seg size (kbytes, -d) unlimited
scheduling priority (-e) 0
file size (blocks, -f) unlimited
pending signals (-i) 513344
max locked memory (kbytes, -l) 64
max memory size (kbytes, -m) unlimited
open files (-n) 1024
pipe size (512 bytes, -p) 8
POSIX message queues (bytes, -q) 819200
real-time priority (-r) 0
stack size (kbytes, -s) 8192
cpu time (seconds, -t) unlimited
max user processes (-u) 4096
virtual memory (kbytes, -v) unlimited
file locks (-x) unlimited
tried to increase the stack size to a very large value but still it does not resolve the issue.
ulimit -s 10000000000
second, we compiled the gfortran using different parameters:
F90FLAGS= -O3 -fno-stack-arrays -fmax-stack-var-size=10000000
where -fmax-stack-var-size=10000000 in gfortran is equivalent to -heap-arrays:0
still it does not work
third, we found ifort is available in goliath, so we used the given parameters in makefile to compile the code:
uqczhan2@goliath:~/macondo/sutrams/src_fno_stack_arrays$ ifort
-bash: ifort: command not found
uqczhan2@goliath:~/macondo/sutrams/src_fno_stack_arrays$ module avail
--------------------------------------------------------------- /usr/share/Modules/modulefiles ---------------------------------------------------------------
dot module-git module-info modules null use.own
---------------------------------------------------------------------- /etc/modulefiles ----------------------------------------------------------------------
ansys/17.0 cuda/10.0 fds/6.6.0 mpi/compat-openmpi16-x86_64 tensorflow/0.10.0
ansys/18.0 cuda/10.1 fds/6.7.0 mpi/intel-mpi/5.1.3.181 tensorflow/0.7.0
ansys/19.0 cuda/7.5 gcc/7.3.1 mpi/mpich-3.0-x86_64 tensorflow/1.2.1
ansys/19.2 cuda/8.0 gocart-adjoint/2.1.1 mpi/mpich-3.2-x86_64 tensorflow/1.9.0
biogeme/2.5 cuda/9.0 gridpro/6.5SP1 mpi/mpich-x86_64 theano/1.0.2
biogeme/3.1.2 dmd/2.074.1 gridpro/6.6 mpi/openmpi-x86_64 torch/2017-09-13
cntk/2016-02-08 fds/6.5.3 intel-xe/2016.2.181 openfoam/5.x torch-rnn/2017-07-31
uqczhan2@goliath:~/macondo/sutrams/src_fno_stack_arrays$ module load intel-xe/2016.2.181
uqczhan2@goliath:~/macondo/sutrams/src_fno_stack_arrays$ ifort
ifort: command line error: no files specified; for help type "ifort -help"
uqczhan2@goliath:~/macondo/sutrams/src_fno_stack_arrays$
in the makefile
F90=ifort
F90FLAGS= -O3 -heap-arrays:0 -fpe:0 -traceback
problem resolved