CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   Hardware (https://www.cfd-online.com/Forums/hardware/)
-   -   why Starccm+ ran out of memory on Centos7 (https://www.cfd-online.com/Forums/hardware/244987-why-starccm-ran-out-memory-centos7.html)

warnerchang September 8, 2022 03:47

why Starccm+ ran out of memory on Centos7
 
When I ran Starccm+ simulation on Centos7. Memory can easily explode, however, physical memory is about 128G. The log file is as follow:
Code:

#
# There is insufficient memory for the Java Runtime Environment to continue.
# Native memory allocation (mmap) failed to map 12288 bytes for committing reserved memory.
# Possible reasons:
#  The system is out of physical RAM or swap space
#  The process is running with CompressedOops enabled, and the Java Heap may be blocking the growth of the native heap
# Possible solutions:
#  Reduce memory load on the system
#  Increase physical memory or swap space
#  Check if swap backing store is full
#  Decrease Java heap size (-Xmx/-Xms)
#  Decrease number of Java threads
#  Decrease Java thread stack sizes (-Xss)
#  Set larger code cache with -XX:ReservedCodeCacheSize=
#  JVM is running with Unscaled Compressed Oops mode in which the Java heap is
#    placed in the first 4GB address space. The Java Heap base address is the
#    maximum limit for the native heap growth. Please use -XX:HeapBaseMinAddress
#    to set the Java Heap base and to place the Java Heap above 4GB virtual address.
# This output file may be truncated or incomplete.
#
#  Out of Memory Error (os_linux.cpp:2749), pid=4629, tid=0x00007f92f72fe700
#
# JRE version: Java(TM) SE Runtime Environment (8.0_202-b08) (build 1.8.0_202-b08)
# Java VM: Java HotSpot(TM) 64-Bit Server VM (25.202-b08 mixed mode linux-amd64 compressed oops)
# Failed to write core dump. Core dumps have been disabled. To enable core dumping, try "ulimit -c unlimited" before starting Java again
#

---------------  T H R E A D  ---------------

Current thread (0x00007f9248001000):  JavaThread "Logging Flush" daemon [_thread_new, id=46237, stack(0x00007f92f71fe000,0x00007f92f72ff000)]

Stack: [0x00007f92f71fe000,0x00007f92f72ff000],  sp=0x00007f92f72fdad0,  free space=1022k
Native frames: (J=compiled Java code, j=interpreted, Vv=VM code, C=native code)
V  [libjvm.so+0xad33a5]  VMError::report_and_die()+0x2e5
V  [libjvm.so+0x4e04c7]  report_vm_out_of_memory(char const*, int, unsigned long, VMErrorType, char const*)+0x67
V  [libjvm.so+0x90fe86]  os::pd_commit_memory(char*, unsigned long, bool)+0xf6
V  [libjvm.so+0x90774f]  os::commit_memory(char*, unsigned long, bool)+0x1f
V  [libjvm.so+0x9113c8]  os::pd_create_stack_guard_pages(char*, unsigned long)+0x48
V  [libjvm.so+0xa7bc0e]  JavaThread::run()+0x1ee
V  [libjvm.so+0x90d8c2]  java_start(Thread*)+0x102
C  [libpthread.so.0+0x7ea5]

I am new to this OS, anyone can offer some help?

flotus1 September 8, 2022 16:54

I guess the reason you asked the question this way is that the simulation ran fine on a machine with a different operating system, and the same amount or RAM?

Don't know if this is a specific issue for CentOS or any other distro, but CCM+ currently copies an instance of the CAD models to every single core it is run on.
If there are a lot of CAD models, or they are fairly large, this can cause problems when running on many cores in parallel.
The only known workaround so far is deleting all CAD models.
Maybe this is relevant to you, otherwise feel free to fill us in on any other information that might be relevant here.

warnerchang September 10, 2022 00:13

Thank you for your reply!

I ran this simulation on ubuntu is fine, however, running it on Centos causes memory explosion! The memory usage is increasing gradually.


All times are GMT -4. The time now is 18:43.