Memory pool (understand the process of new developing space from the perspective of kernel)

Every time we pass new perhaps malloc apply 200 In bytes , Looks like OS It gives us the size we need , In fact, there is a slight difference . 

int* p = new int[50];
char* p = (char*)malloc(200);

We're calling new/malloc in the future , In fact, it will go from user mode to kernel mode , Suppose you want to apply for 200 Bytes , At this time, if there is no continuous 200 Bytes of space ,OS Memory processing algorithms may be called , Adjust the position of other variables , Make room for us A big piece Space .

When our users call , Just the results , No matter the process , however OS When calling the memory processing algorithm , It takes time , If we frequently apply for memory later ,OS Also call the processing algorithm frequently , This is inefficient , It's also time-consuming . therefore To improve efficiency , In the first application ,OS The amount of space allocated > The size of our application .

This space is on the user layer , Every time we have to pass new/malloc Application memory , All of them apply to this space , In this case , There is no need to switch states , It will not call memory processing algorithms frequently . This space we call Memory pool ！