tiller man wrote:
> For the last few weeks, I have been allocating extremely large amounts of memory in the programs I have been writing. I keep running into the same problem.
>
> The amount of memory I am allocating is almost certainly larger than the amount of physical memory that the machine has available. Not being a linux expert, I assumed that the machine would be able to handle this fairly gracefully through virtual memory allocation. But it doesn't.
>
> And thats where the problem is. It doesn't fail in the call to malloc() or realloc(), it appears to continue running. However, the machine locks up at this point.
> (Well so it would seem, I am not sure if it still running the process at all, or just being extremely slow, as I only print out debugging information every 1000 allocations or so).
>
> Now, I currently access this machine(AMD64 with 1GB of memory I believe) through ssh, as it is not located directly at my workplace. After the process has allocated a certain amount of memory, I can't kill it, as it is completely unresponsive. I can't open another ssh session with the machine either. I just get no response from anything. As a result, I have to email someone working next to the machine to tell them to restart the machine every time this happens.
>
> Obviously, this is extremely inconvenient for all involved. Does anyone have any idea what could be causing this and / or possible solutions? My knowledge of Linux is fairly limited, so I am not able to provide much information (especially as I cannot do anything when it locks up), so I will just provide the output of uname -a as an indicator (please advise me on providing more useful information, I know this isnt much to go on).
>
How big is your swap file? If you run out of Physical memory your data
gets written to swap, and if you are using a lot of swap, it may become
VERY sluggish.
--
Charles MacDonald Stittsville Ontario
cmacd [ at ] TelecomOttawa [ dot ] net Just Beyond the Fringe
http://www.TelecomOttawa.net/~cmacd/
No Microsoft Products were used in sending this e-mail.