I wrote a simple wirefire hightmap render and a not quite so simple scene graph recently. While testing I noticed a memory leak, so I went through the entire project and eliminted what I believe to be every potential leak. However, now when I run my program it will start at about 8900k memory usasge (this is in the process tab of the windows task manager), slowly work it's way down to 8650k memory usage over the course of two minutes... then begin increasing at 4k per second from there on out. I'm certain that there is no object that is randomly being created that late into execution. This leads me to believe that either opengl or windows is doing something. I've yet to try running it for more then 10 or so minutes, so I'm not sure if it will eventully stop or not. Has anyone seen anything like this before? If not, does anyone know of a good freeware program I can use to try to track down where this memory is going? Thanks

Quote:Original post by Telastyn
I've used the app called LeakTracer before. It's free and easy to use.

My favourite memory manager has always been Paul Nettle's. Free, simple and works very well.
I'm really promoting Paul's stuff today...

Quote:Original post by Mercury

Quote:Original post by Telastyn
I've used the app called LeakTracer before. It's free and easy to use.

My favourite memory manager has always been Paul Nettle's. Free, simple and works very well.
I'm really promoting Paul's stuff today...

Seconded! I will promote Pual's stuff as long as it exists! I am currently using his memory manager as well as his logger for my project and they have saved me tons of time. It is of such high quality, I would pay for a license to use it if it ever came to that. That's how good I think it is. Anyways, Fib One, definitly take a look into that. I know it has helped me find the most obsecure bugs before that I would have never found before, such as when I had a 'delete' instead of a 'delete []'. It told me that I was using the wrong operator! Great stuff indeed.

- Drew

Thanks for the suggestions. I'll definitely have to give Paul Nettle's program a try when I can get back to my computer.