In the last article, I explained something about the concept of MemPool. Now I want to show you some optimization in design and a simple sample of MemPool that can be downloaded and tested.
Before testing the code imagine that we have a lot of chunks on the pool consecutively. in the last article, I introduced the body of a chunk just like a basket that contains the data. now the chunk can have a fixed design to have more readable code.

New Chunk Structure

As you can see we can place the address of the chunk in the head of the chunk instead of the end line of that and store the address of the behind chunk. all we try to do is make a chain of chunks by minimum size.

Here are the results of my test for the performance of this system for 100000 allocation/deallocation in release mode. I try to simulate a real environment for both the memory pool and OS memory system, but I don’t know how much it’s real. 😉
Also, the source code of this sample already exists at the end of this post. You may change and test it by yourself.

NOTE:
There is no force to use a general MemPool in your project! you can make a faster MemPool for each part of your application considering the needs. for instance, I write a fast hybrid MemPool which is faster. but the number of allocations for this hybrid MemPool is limited and must be defined in creation time.

Download: MemPool_test.rar