Thread: Effect of Perm gen memory on CPU Utilization

Hello All,

I am Abhishek, working on a game project. Whenever we are starting our game application in production, it works perfectly around 3 days after 3 days, CPU utilization is increasing upto around 95% to 99% and application stops to respond. Before 6 days, I started application again and we doubled our Perm Gen memory size. The same problem happened but it happened after 6 days.

Now, i am not able to relate CUP utilization to Perm Gen memory. Is Perm Gen memory effect CPU utilization?

Thanks,
Abhishek

Why dont you use a profiler? Thats what they have been made for.

Dear @Cornix, Thanks for your reply...
I used Visual VM profiler and got to know that free space in Perm Gen memory is reducing slowly slowly and CPU utilization is increasing corresponding to this reduction in Perm gen memeory.
Thats why, i looking relation between Perm Gen memory and CPU utilization.
Thanks

Does your application actually need all that memory or might it be that you have a memory leak somewhere?

Dear @Cornix, I dont think memory leak is a reason because my jvm never crashed. in memory leak condition, after a interval jvm should crashed due to insufficient memory and garbage collector is also running.

Thread moved. Please don't post requests for help in Member Introductions.

Welcome to the Forum! Please read this topic to learn how to post code correctly and other useful tips for newcomers.

I don't believe your assertion that a JVM crash must always accompany a memory leak is correct. When doubling the size of PermGen memory doubles the time the program is stable, and a profiler shows you that "free space in Perm Gen memory is reducing slowly slowly," I'd say you have a PermGen memory leak.

Understanding the possible reasons for the results you've experienced requires an intimate understanding of your program's design, use of 3rd party libraries and their possible contribution, etc., an intimacy I would only expect your design/programming team to have. There are good "PermGen memory leak" articles on the 'net that discuss possible causes and remedies that you should refer to.

Good luck, and I hope you'll share your findings with us when you've determined root cause and a solution.

Dear @Gregrannon, i have one doubt that if PermGen memory is full, will it increase CUP Utilization or not?

Typically, a decrease in critical computer resources causes CPU utilization to increase until instability and then failure occur, but I can't say specifically that a decrease in PermGen memory is causing the behavior you've observed when running your program on your test platform(s). Did the profiling you've done indicate any other possible causes for the increased CPU utilization?

And why does it matter? If you know you have a PermGen memory leak and that the leak will eventually cause your program to fail, why not find the leak and fix it?