forrtl: severe (157...
 
Notifications
Clear all

forrtl: severe (157): Program Exception - access violation in LS Dyna with ICFD  

 

alpha754293
(@alpha754293)
New Member
Joined: 2 months ago
Posts: 4
November 14, 2020 9:33 pm  

Hello.

 

I am trying to run this example from dynaexamples.com:

 

https://www.dynaexamples.com/icfd/advanced-examples/windturb

 

I am using MPP LS-DYNA R12.0.0, Platform MPI.

 

The specific command that I am using via a Cygwin shell is:

 

$ time -p "/cygdrive/c/Program Files (x86)/IBM/Platform-MPI/bin/mpirun.exe" -np 8 ls-dyna_mpp_d_R12_0_0_winx64_ifort170_pmpi.exe memory=1000m i=main.k 2>&1 | tee main.txt

 

So I am already using double precision and it fails when it tries to start the ICFD solver.

 

Here is the log:

 

$ time -p "/cygdrive/c/Program Files (x86)/IBM/Platform-MPI/bin/mpirun.exe" -np 8 ls-dyna_mpp_d_R12_0_0_winx64_ifort170_pmpi.exe
i=main.k 2>&1 | tee main.txt
mpirun: Drive is not a network mapped - using local drive.
Date: 11/14/2020 Time: 19:37:18

___________________________________________________
| |
| Livermore Software Technology, LLC |
| |
| 7374 Las Positas Road |
| Livermore, CA 94551 |
| Tel: (925) 449-2500 Fax: (925) 449-2507 |
| www.lstc.com |
|_________________________________________________|
| |
| LS-DYNA, A Program for Nonlinear Dynamic |
| Analysis of Structures in Three Dimensions |
| Version : mpp d R12.0.0 Date: 06/19/2020 |
| Revision: 148931 Time: 17:26:40 |
| |
| Features enabled in this version: |
| Distributed Memory Parallel |
| CESE CHEMISTRY EM ICFD STOCHASTIC_PARTICLES |
| FFTW (multi-dimensional FFTW Library) |
| BigID option |
| |
| Licensed to: |
| |
| Platform : WINX64 SSE2 (PLATFORM MPI 8.01) |
| OS Level : Windows 7/8/10 & Server 2012/2016 |
| Compiler : Intel Fortran XE 2017 MSVC++ 2017 |
| Hostname : aes06 |
| Precision : Double precision (I8R8) |
| SVN Version: 148978 |
| |
| Unauthorized use infringes LST copyrights |
|_________________________________________________|

LS-DYNA CPU Licenses Requested: 8 Available: 8

Input file: main.k

The native file format : 64-bit small endian
Memory size from default : 100000000, 0
Memory option : AUTO selected
Memory size from default : 100000000, 100000000

on UNIX computers note the following change:

ctrl-c interrupts ls-dyna and prompts for a sense switch.
type the desired sense switch: sw1., sw2., etc. to continue
the execution. ls-dyna will respond as explained in the users manual

type response
----- ------------------------------------------------------------
quit ls-dyna terminates.
stop ls-dyna terminates.
sw1. a restart file is written and ls-dyna terminates.
sw2. ls-dyna responds with time and cycle numbers.
sw3. a restart file is written and ls-dyna continues calculations.
sw4. a plot state is written and ls-dyna continues calculations.
sw5. ls-dyna enters interactive graphics phase.
swa. ls-dyna flushes all output i/o buffers.
swb. a dynain is written and ls-dyna continues calculations.
swc. a restart and dynain are written and ls-dyna contiunues calculations.
swd. a restart and dynain are written and ls-dyna terminates.
lpri toggle implicit lin. alg. solver output on/off.
nlpr toggle implicit nonlinear solver output on/off.
iter toggle implicit output to d3iter database on/off.
prof output timing data to prof.out and continue.
conv force implicit nonlinear convergence for current time step.
ttrm terminate implicit time step, reduce time step, retry time step.
rtrm terminate implicit at end of current time step.


******** notice ******** notice ******** notice ********
* *
* This is the LS-DYNA Finite Element code. *
* *
* Neither LST nor the authors assume any responsibility for *
* the validity, accuracy, or applicability of any results *
* obtained from this system. Users must verify their own *
* results. *
* *
* LST endeavors to make the LS-DYNA code as complete, *
* accurate and easy to use as possible. *
* Suggestions and comments are welcomed. Please report any *
* errors encountered in either the documentation or results *
* immediately to LST through your site focus. *
* *
* Copyright (C) 1990-2020 *
* by Livermore Software Technology, LLC *
* All rights reserved *
* *
******** notice ******** notice ******** notice ********

Beginning of keyword reader 11/14/20 19:37:46

11/14/20 19:37:47
Open include file: mesh_blade_slide_diff.k
ICFD::WARNING: Surface elements with aspect ratio > 3.
This may result in volume elements with poor quality.
For details check file icfd_surf_q.dat.

Memory required to process keyword : 283824
Additional dynamic memory required : 45783447


MPP execution with 8 procs

Initial reading of file 11/14/20 19:37:54
ICFD::Mesh -> Start Mesh Generation
ICFD::Mesh: Initializing.
ICFD::Mesh: Checking surface mesh integrity.
ICFD::Mesh: Orienting and creating normals.
ICFD::Mesh: Inserting nodes and creating elements.
ICFD::Mesh: Elements Generated:
*************
ICFD::Mesh: Smoothing and improving quality.
ICFD::Mesh: Creating sub-domains.
ICFD::Mesh: Creating boundary layer mesh.
10%
20%
30%
40%
50%
60%
70%
80%
90%
ICFD::Mesh: Checking surface mesh integrity.
ICFD::Mesh: Orienting and creating normals.
ICFD::Mesh: Inserting nodes and creating elements.
*************
ICFD::Mesh: Smoothing and improving quality.
ICFD::Mesh: Creating sub-domains.
ICFD::Mesh: Creating boundary layer mesh.
10%
20%
30%
40%
50%
60%
70%
80%
90%
ICFD::Mesh: Creating FEM data structures.
ICFD::Mesh -> Mesh Built

Performing Decomposition -- Phase 1 11/14/20 19:51:35

Performing Recursive Coordinate Bisection (RCB)

Memory required for decomposition : 5794
Additional dynamic memory required : 1231m


Performing Decomposition -- Phase 2 11/14/20 19:51:37
Performing Decomposition -- Phase 3 11/14/20 19:51:40

input of data is completed

initial kinetic energy = 0.00000000E+00

Memory required to begin solution (memory= 284K memory2= 5803)
Minimum 5795 on processor 1
Maximum 5803 on processor 0
Average 5796

Matrix Assembly dynamically allocated memory
Maximum 277

Additional dynamically allocated memory
Minimum 184M on processor 2
Maximum 1286M on processor 0
Average 323M

Total allocated memory
Minimum 184M on processor 2
Maximum 1286M on processor 0
Average 323M

initialization completed
******************************************
THE INCOMPRESSIBLE CFD (ICFD) SOLVER IS ON
******************************************
ICFD:: Transient analysis
End Time = 2.5002E+01
ICFD:: Second Order BDF time integration
ICFD:: Second Order Fractional Step Solve
ICFD:: OSS Stabilization Used
ICFD:: Divergence Cleaning using internal sub-stepping
ICFD:: 3D Analysis
ICFD:: CFD Solve Only
ICFD:: LES Smagorinsky Turbulence Model Activated
ICFD:: Van Driest damping function actived
----------------------------
ICFD:: Material ID 1 with:
Density = 1.2250E+00
Viscosity = 1.8700E-05
----------------------------
ICFD:: Surface Part ID 1 with :
Boundary condition : Nonslip
Boundary condition : Imposed Move
Material ID 1
----------------------------
ICFD:: Surface Part ID 3 with :
Boundary condition : Periodic Master
Boundary condition : Imposed Move
Material ID 1
----------------------------
ICFD:: Surface Part ID 4 with :
Boundary condition : Periodic Slave
Material ID 1
----------------------------
ICFD:: Surface Part ID 5 with :
Boundary condition : Freeslip
Material ID 1
----------------------------
ICFD:: Surface Part ID 8 with :
Boundary condition : Nonslip
Material ID 1
----------------------------
ICFD:: Surface Part ID 9 with :
Boundary condition : Prescribed pressure
Material ID 1
----------------------------
ICFD:: Surface Part ID 10 with :
Boundary condition : Prescribed linear velocity
Boundary condition : Prescribed normal velocity
Material ID 1
----------------------------
ICFD:: Surface Part ID 11 with :
Boundary condition : Nonslip
Boundary condition : Imposed Move
Material ID 1
----------------------------
ICFD:: Surface Part ID 12 with :
Boundary condition : Nonslip
Boundary condition : Imposed Move
Material ID 1
----------------------------
ICFD:: Surface Part ID 13 with :
Boundary condition : Nonslip
Boundary condition : Imposed Move
Material ID 1
----------------------------
ICFD:: Volume Part ID 40 with :
Surface Part ID = 1
Surface Part ID = 3
Surface Part ID = 11
Surface Part ID = 12
Surface Part ID = 13
Boundary condition : Imposed Move
Material ID 1
----------------------------
ICFD:: Volume Part ID 30 with :
Surface Part ID = 4
Surface Part ID = 5
Surface Part ID = 8
Surface Part ID = 9
Surface Part ID = 10
Material ID 1
******************************************
ICFD:: Fluid -> Assembly...
1 t 0.0000E+00 dt 1.00E-03 flush i/o buffers 11/14/20 19:53:13
1 t 0.0000E+00 dt 1.00E-03 write d3plot file 11/14/20 19:53:13
forrtl: severe (157): Program Exception - access violation
Image PC Routine Line Source
ls-dyna_mpp_d_R12 0000000149298B24 PFEM_D3PLOT_EXTRA 7714 d3plot.f
ls-dyna_mpp_d_R12 0000000143ADEF34 PRTDAT 2562 dyn36.F
ls-dyna_mpp_d_R12 0000000140F439B6 FEM3D_OUTPUT 29914 fem3d_a.F
ls-dyna_mpp_d_R12 0000000140ED364C FEM3D 5269 fem3d.F
ls-dyna_mpp_d_R12 0000000140EAE2B8 SOLTN 425 fem3d.F
ls-dyna_mpp_d_R12 0000000140FDEB98 OVERLY 252 dyn0.F
ls-dyna_mpp_d_R12 000000013FDC0C79 MAIN__ 800 dynm.F
ls-dyna_mpp_d_R12 000000014BC8C5FE Unknown Unknown Unknown
ls-dyna_mpp_d_R12 000000014CAB751C Unknown Unknown Unknown
kernel32.dll 00000000776F652D Unknown Unknown Unknown
ntdll.dll 000000007792C541 Unknown Unknown Unknown
MPI Application rank 0 exited before MPI_Finalize() with status 157
real 1000.91
user 0.84
sys 2.07

 

I don't think that this example should take 1000m words to run, but I tried to set it incredibly high to try to ensure that there aren't problems.

It's running on a HP Z420 workstation, with an 8-core Intel Xeon E5-2690 (v1), with 128 GB of RAM, in Windows 7.

 

(It has been my experience that almost all of the other examples from dynaexamples.com will run fine, as-is, so I am not really sure why I am having troubles trying to run this example.)

 

Any ideas or suggestions as to how I can resolve this would be greatly appreciated.

1605389604-Capture.png

Quote
DavidBenoit
(@davidbenoit)
New Member
Joined: 2 months ago
Posts: 1
November 20, 2020 2:30 pm  

I started having this access violation 157 error a few months ago, never had it before.

I have contacted support from LSTC about this problem a few weeks ago and they told me it was a problem mostly/only happening with SMP on windows platform. 

However this error still happens every once in a while even after i switched to MPP. It seems like a lot of forum posts have been coming up in the last weeks about this problem. I am wondering if it could be a problem coming from new versions of LS DYNA solver


ReplyQuote
Negative Volume
(@negativevolume)
Senior Member Admin
Joined: 2 years ago
Posts: 367
December 10, 2020 8:36 pm  

@alpha754293 and @davidbenoit

Yeah this is mainly an issue with SMP and implicit simulations. The most concrete solution that I found to the issue was to use a version of LS-Dyna with double precision. So in the solver name you want to use the version that has "_d_" not "_s_". Like @davidbenoit, there are several threads on this forum regarding the topic that may have more detail for you. 


ReplyQuote
alpha754293
(@alpha754293)
New Member
Joined: 2 months ago
Posts: 4
December 10, 2020 9:23 pm  

@negativevolume

Yeah, I've read those threads already.

But as is it shown, this is why this is different because a) this is for ICFD, as opposed to an implicit solution and b) as noted, I am already using the double precision, MPP solver, which is recommended for ICFD solutions.

So...beyond that, it would appear that the "usual" tips to address/solve this has already been employed and I am still getting a forrtl (157) error, shown above.

So I didn't know if there was something else that I can try to resolve this issue.

Thanks.


ReplyQuote
Negative Volume
(@negativevolume)
Senior Member Admin
Joined: 2 years ago
Posts: 367
January 4, 2021 2:46 pm  

@alpha754293 I see. sorry for the delayed response as I've been away on vacation. Personally, I don't run across this problem often, but from doing some research I found that the above tips were the most common solutions to this issue. Aside from that, have you tried using an older solver? v12.0 is like brand new so there may be issues. Maybe try 10.2 or 11.0?

*Edit: I just saw in the example that it requires v12.0 so my suggestion wouldn't make sense. You can always try to contact LSTC directly about the problem, especially since you are just trying to run one of their example simulations. 


ReplyQuote
alpha754293
(@alpha754293)
New Member
Joined: 2 months ago
Posts: 4
January 4, 2021 6:09 pm  

@negativevolume

Thank you for your help.

 

I wanted to see how much experience the community may have with LS-DYNA's ICFD solver.

 

Thanks.


ReplyQuote
Share: