[gschaider]

Hi!

I was running a simpleFoam-case with a symmetryPlane. When trying to run it with a LRR-turbulence it always crashed with this message:

DILUPBiCG: Solving for epsilon, Initial residual = 0.0535534, Final residual = 0.00270083, No Iterations 1
new cannot satisfy memory request.
This does not necessarily mean you have run out of virtual memory.
It could be due to a stack violation caused by e.g. bad use of pointers or an out of date shared library

Although the machine in question should have sufficient memory for that case.

To make the bug reproducible I changed the upperWall of the pitzDaily-case to symmetryPlane. The case ran until the 6th time-step and the crashed.

Then I tried to run the case with valgrind:
a) valgrind printed some messages (see below, I will spare you the complete dump)
b) the solver did not crash

Basically there are two kinds of messages:
- "Invalid read of size 4" or size 8 on various places
- "Address 0xXXXXXX is 8 bytes inside a block of size 24 free'd" in operator delete(void*) in vg_replace_malloc.c

Don't know enough about valgrind to interpret that. All the stack-traces have somewhere SymmTensor<double>-stuff in them.

I could reproduce this behaviour with
- OpenFOAM 1.4.1 and 1.4 (as distributed, only with patches from the message-board applied)
- On 64 and 32 bit Linux-machines

I could run a modified pitzDaily-case with a 1.3-installation (but that uses Tensor<double>, so the problem seems to be with SymmTensor)

Bernhard

[henry]

Try with this fix,
symmTensorField.C to replace OpenFOAM-1.4.1/src/OpenFOAM/fields/Fields/symmTensorField/symmTensorField.C

I setup the pitzDaily-case with a symmetryPlane as you suggest above and it now runs.

[gschaider]

Yeah. This fixes it. Thanks a lot.