[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: More than 2, but less than 3 GiB per process memory?



On Fri, Sep 23, 2005 at 10:43:40AM -0500, Ron Johnson wrote:
> On Fri, 2005-09-23 at 14:45 +0200, Malte Cornils wrote:
> > Hello,
> > 
> > we've been trying to make a program (ITK/VTK image processing for a university 
> > project) work. Unfortunately, the process needs slightly above 2 GiB of 
> > virtual memory.
> > 
> > Judging from the documentation I've seen, on 32bit systems I should be able to 
> > allocate up to 3 GiB of virtual memory (1 GiB of 32bit address space is 
> > reserved for the kernel).
> > 
> > However, our test case for this terminates when trying to allocate more than 2 
> > GiB of memory, even though we have a really big swap file.
> > 
> > Does anyone have a clue why 2 GiB is the limit?
> 
> http://www.puschitz.com/TuningLinuxForOracle.shtml#AddressMappingsOnLinux
> 0GB-1GB  User space   - Used for executable and brk/sbrk allocations
>                         (malloc uses brk for small chunks).
> 1GB-2GB  User space   - Used for mmaps (shared memory), shared libraries
>                         and malloc uses mmap (malloc uses mmap for large
>                         chunks).
> 2GB-3GB  User space   - Used for stack.
> 3GB-4GB  Kernel Space - Used for the kernel itself.
> 
> 
> >                                                  Would using a swap partition 
> > instead of a swap file help? The 64G HIGHMEM kernel config option does not 
> > seem to make a difference. On a real AMD64 system, the program works fine 
> > with >2 GiB RAM, as was expected. Is there any way to make this work, besides 
> > changing the algorithm to use less than 2 GiB of memory?
> 
> No.

Actually, some kludgery might help here.

It seems malloc restricts itself to the user space.
But if you were te replace malloc (or provide your own allocation
and freeing methods, which *miht* be possible in C++) you could do
the following.

Near the start of your program, allocate a *huge* array on the stack, like

char * hugepointer;
int main(...)
{
  char huge[1000000000];
  hugepointer = huge;
  restofprogram(...);
}

Then have your private malloc (or allocation methods)
allocate space within huge, and when it's full, use
calls to the regular malloc for the rest.

-- hendrik



Reply to: