Closed gjaeger closed 4 years ago
I assume this is not the same zone issue?
Try compiling and running the _db (debug) version. It should give us better traceback and a clue for what is going on.
I assume this is not the same zone issue?
Yes, it's not the same issue.
with debug version:
$ fds_mpi_gnu_linux_64_db evac_memory_test0.fds
Starting FDS ...
MPI Process 0 started on ias7139
Reading FDS input file ...
EVAC: Emesh 1 FF1stFloor has 4 door flow fields
FDS+Evac pressure method : FFT
WARNING: Exit LeftExit XB adjusted to mesh FF1stFloor
Old XB: 0.0000 0.0000 5.2500 6.4500 1.4500 1.5500
New XB: 0.0000 0.0000 5.2500 6.5000 1.4500 1.5500
WARNING: Exit RightExit XB adjusted to mesh FF1stFloor
Old XB: 50.0000 50.0000 25.2500 26.4500 1.4500 1.5500
New XB: 50.0000 50.0000 25.2500 26.5000 1.4500 1.5500
WARNING: Exit TopExit XB adjusted to mesh FF1stFloor
Old XB: 3.2500 4.4500 40.0000 40.0000 1.4500 1.5500
New XB: 3.2500 4.5000 40.0000 40.0000 1.4500 1.5500
Program received signal SIGSEGV: Segmentation fault - invalid memory reference.
Backtrace for this error:
#0 0x7f5ad1a8770d in ???
#1 0x7f5ad1a86925 in ???
#2 0x7f5ad069af1f in ???
#3 0x55a62769b7ae in determine_fixed_elements
at ../../Source/hvac.f90:1928
#4 0x55a6276cf538 in __hvac_routines_MOD_proc_hvac
at ../../Source/hvac.f90:885
#5 0x55a62789b401 in __read_input_MOD_read_data
at ../../Source/read.f90:119
#6 0x55a627b8e070 in fds
at ../../Source/main.f90:115
#7 0x55a627c32c80 in main
at ../../Source/main.f90:5
Segmentation fault (core dumped)
Yep that is a bug. Surprised we haven't had that pop up before. DETERMINE_FIXED_ELEMENTS is called from PROC_HVAC but we aren't checking to see if there are any HVAC elements. The Intel compiler really should throw an error for this as well.
Should be fixed now.
Thanks. It works.
$fds_mpi_gnu_linux_64 evac_memory_test0.fds
Starting FDS ...
MPI Process 0 started on ias7139
Reading FDS input file ...
EVAC: Emesh 1 FF1stFloor has 4 door flow fields
FDS+Evac pressure method : FFT
WARNING: Exit LeftExit XB adjusted to mesh FF1stFloor
Old XB: 0.0000 0.0000 5.2500 6.4500 1.4500 1.5500
New XB: 0.0000 0.0000 5.2500 6.5000 1.4500 1.5500
WARNING: Exit RightExit XB adjusted to mesh FF1stFloor
Old XB: 50.0000 50.0000 25.2500 26.4500 1.4500 1.5500
New XB: 50.0000 50.0000 25.2500 26.5000 1.4500 1.5500
WARNING: Exit TopExit XB adjusted to mesh FF1stFloor
Old XB: 3.2500 4.4500 40.0000 40.0000 1.4500 1.5500
New XB: 3.2500 4.5000 40.0000 40.0000 1.4500 1.5500
Fire Dynamics Simulator
Current Date : January 17, 2020 16:41:12
Revision : FDS6.7.3-364-gf8717c88f-master
Revision Date : Thu Jan 16 15:58:20 2020 -0500
Compiler : Gnu gfortran 9.2.1-17ubuntu1~18.04.1)
Compilation Date : Jan 17, 2020 10:44:27
MPI Enabled; Number of MPI Processes: 1
OpenMP Enabled; Number of OpenMP Threads: 16
MPI version: 3.1
MPI library version: Open MPI v4.0.2, package: Open MPI artss@ias7139 Distribution, ident: 4.0.2, repo rev: v4.0.2, Oct 07, 2019
Job TITLE : test for RAM: fire mesh
Job ID string : evac_memory_test0
Time Step: -199, Evacuation Initialization Time: -1.990 s
Time Step: -149, Evacuation Initialization Time: -1.490 s
Time Step: -99, Evacuation Initialization Time: -0.990 s
Time Step: -49, Evacuation Initialization Time: -0.490 s
Time Step: 1, Simulation Time: 0.10 s
...
Time Step: 101, Simulation Time: 10.00 s
STOP: FDS completed successfully (CHID: evac_memory_test0)
We can close the issue.
After compiling FDS (current master version) with GNU Fortran and OpenMPI, I get a segmentation fault when executing evac files (like evac_memory_test0.fds).
Linux ias7139 4.15.0-74-generic #84-Ubuntu SMP Thu Dec 19 08:06:28 UTC 2019 x86_64 x86_64 x86_64 GNU/Linux
GNU Fortran (Ubuntu 9.2.1-17ubuntu1~18.04.1) 9.2.1 20191102
mpirun (Open MPI) 4.0.2
My result:
The fire simulation (like couch.fds) works.
What's my mistake?