[vtkusers] Memory allocation problems under Python wrapped VTK in Windows

[Lassi Paavolainen]

Hi,

On Fri, 16 Jan 2009, Jim wrote:

> Is this a problem of memory fragmentation?
>    
>   Jim

No it seems not to be as I can execute that non-VTK program and Python 
wrapped VTK code in any which order and the result is always the same.

Lassi

> Lassi Paavolainen <lopaavol at jyu.fi> wrote:
>   Hi,
> 
> I was wondering about a memory allocation problem that seems to be only in 
> some machines where we run our software so I decided to make a simple test 
> program for that (code attached at the end of the message).
> 
> I got some interesting results from the test program that I cannot 
> explain. I used same code in console application not using VTK and in VTK 
> class but got different results of maximum memory possible to dynamically 
> allocate. Result was also platform dependent in VTK class.
> 
> To the results. I used three different systems explained below:
> 
> Linux (64 bit): 2 GB of RAM, VTK 5.2, Python 2.5
> Windows Vista (32 bit): 3 GB of RAM, VTK 5.2, Python 2.6
> Windows XP Pro (32 bit): 8 GB of RAM (3 GB usable), VTK 5.2, Python 2.6
> 
> Exactly same code was used in both Windows machines (compiled in Vista). I 
> ran this code from command line and using Python interpreter to ran 
> wrapped VTK class and got following interesting results:
> 
> Linux:
> - bash: about 3300 MB
> - Python: about 3300 MB
> 
> Vista:
> - Command line: about 1720 MB
> - Python: about 1180 MB!!!
> 
> XP Pro:
> - Command line: about 1920 MB
> - Python: about 720 MB!!!
> 
> So my question is, how it is possible that VTK class ran from Python in 
> Windows can't allocate as much continuous memory as it is possible to 
> simple application? Same thing happens if I try to allocate memory using 
> for example vtkUnsignedCharArray. I'm guessing the problem is not in VTK 
> but Python wrapped VTK. Does anyone know where that memory limit defined 
> by Python wrapped VTK and platform is coming from?
> 
> Here is the memory allocation testing code. Probably not the best but 
> gives similar results as in real application.
> 
> #include 
> 
> #define START_ALLOC 1048576
> #define MAX_ALLOC 4294967296
> #define START_STEP 536870912
> 
> int main(int argc, char* argv[])
> {
> long long alloc = START_ALLOC;
> long long maxAlloc = MAX_ALLOC;
> long long minAlloc = 0;
> long long step = START_STEP;
> while(maxAlloc - minAlloc > START_ALLOC) 
> {
> std::cout << "Allocating " << alloc << " bytes of memory." << std::endl;
> try 
> {
> char *mem = new char[alloc];
> minAlloc = alloc;
> alloc += step;
> delete[] mem;
> }
> catch (std::bad_alloc &ba)
> {
> std::cout << "Couldn't allocate " << alloc / 1048576 << " MB of memory." << std::endl;
> maxAlloc = alloc;
> alloc = minAlloc;
> step /= 2;
> alloc += step;
> }
> }
> 
> std::cout << "Maximum size of memory allocated: " << minAlloc / 1048576 << " MB" << std::endl;
> 
> return 0;
> }
> 
> Regards,
> Lassi Paavolainen
> _______________________________________________
> This is the private VTK discussion list.
> Please keep messages on-topic. Check the FAQ at: http://www.vtk.org/Wiki/VTK_FAQ
> Follow this link to subscribe/unsubscribe:
> http://www.vtk.org/mailman/listinfo/vtkusers
> 
> 
>        

-- 
Lassi Paavolainen, M.Sc.
Software Engineer
BioImageXD (http://www.bioimagexd.net)
lopaavol at jyu.fi