Greetings,

I'm having trouble with enlarging an array in a program I've written.

When I run the program, gdb spits out the following:

warning: (Internal error: pc 0x804e35e in read in psymtab, but not in 
symtab.)


Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 1024 (LWP 18873)]
warning: (Internal error: pc 0x804e35e in read in psymtab, but not in 
symtab.)

warning: (Internal error: pc 0x804e35e in read in psymtab, but not in 
symtab.)

warning: (Internal error: pc 0x804e35e in read in psymtab, but not in 
symtab.)

0x0804e35e in main_mp_advanceparticles_ () at main.f90:99
99      main.f90: No such file or directory.
         in main.f90
Current language:  auto; currently fortran
warning: (Internal error: pc 0x804e35e in read in psymtab, but not in 
symtab.)

warning: (Internal error: pc 0x804e35e in read in psymtab, but not in 
symtab.)


The complaint about line 99 matches the following block:
     FORALL( i=1:SIZE(gParticles) )
        gParticles(i)%lastPhi = gParticles(i)%phi
        gParticles(i)%lastr = gParticles(i)%r
        gParticles(i)%phi = gParticles(i)%phi + ( dtStep * GetOmega( 
gParticles(i)%r ) )
     END FORALL

Line 99 itself is the one updating the phi component.

The array gParticles is defined as:

TYPE(Particle), SAVE, ALLOCATABLE :: gParticles(:)


The size of the array is set in an input file. Once I make the array 
bigger than about 80000 elements, the segfault problems appear. What's 
confusing me is this:
This isn't an 'out of memory' problem - the array has already been 
initialised by this point. It's even been used to generate a PNG file. 
The code is compiling with as many warnings as I can get the compiler to 
manage, and the only 'problem' is an unused variable.

One final thing:
ifort -V gives
Version 8.0   Build 20031016Z

With fewer particles, the code runs perfectly. It looks like a memory 
error, but I dont' see how it can be - the array has already been used.

Any thoughts much appreciated.

TIA,

Richard

On Mon, 03 May 2004 11:58:12 +0200, Richard Edgar <rge21@astro.su.se> wrote:


>I'm having trouble with enlarging an array in a program I've written.
>
>One final thing:
>ifort -V gives
>Version 8.0   Build 20031016Z

I would suggest trying a more recent compiler to see if the behavior changes.
If you're still having problems, please contact Intel Premier Support.  Even
if you're using the non-commercial, limited-support version, we'd like to know
about problems you encounter.


Steve Lionel
Software Products Division
Intel Corporation
Nashua, NH

User communities for Intel Software Development Products
  http://softwareforums.intel.com/
Intel Fortran Support
  http://developer.intel.com/software/products/support/

Steve Lionel wrote:

>>I'm having trouble with enlarging an array in a program I've written.
>>
>>One final thing:
>>ifort -V gives
>>Version 8.0   Build 20031016Z
> 
> 
> I would suggest trying a more recent compiler to see if the behavior changes.
> If you're still having problems, please contact Intel Premier Support.  Even
> if you're using the non-commercial, limited-support version, we'd like to know
> about problems you encounter.

OK.... I'll nag the appropriate sysadmins.

Doing more research over on the Support Forums, it seems that the 
problem arises with -static and friends. Unfortunately, I have to have 
static linking, since this program is to be run on a barebones compute 
grid. Trying -nothreads and ulimit -s unlimited didn't help either

Thanks,

Richard

On Mon, 03 May 2004 15:59:33 +0200, Richard Edgar <rge21@astro.su.se> wrote:
> Steve Lionel wrote:
> 
>>>I'm having trouble with enlarging an array in a program I've written.
>>>
>>>One final thing:
>>>ifort -V gives
>>>Version 8.0   Build 20031016Z
>> 
>> 
>> I would suggest trying a more recent compiler to see if the behavior changes.
>> If you're still having problems, please contact Intel Premier Support.  Even
>> if you're using the non-commercial, limited-support version, we'd like to know
>> about problems you encounter.
> 
> OK.... I'll nag the appropriate sysadmins.
> 
> Doing more research over on the Support Forums, it seems that the 
> problem arises with -static and friends. Unfortunately, I have to have 
> static linking, since this program is to be run on a barebones compute 
> grid. Trying -nothreads and ulimit -s unlimited didn't help either

What about -static-libcxa -nothreads

Also make sure to use the i686 version of glibc on the
target instead of a i386 version.