Language memory management is an important aspect of language design. It is an important factor in determining language performance. Whether it is manual management in C language or garbage collection in Java, they have become the most important features of the language. Here we take the Python language as an example to illustrate the memory management method of a dynamically typed, object-oriented language.
Summary in one sentence: reference counting is mainly, clearing marks, and generational recycling is auxiliary(Recommended learning: Python video Tutorial)
Garbage collection in python (3 types)
Reference counting
When an object When the reference counter reaches 0, the object may be in memory but no longer accessible. No other operations can be performed during Python's garbage collection. If Python will recycle an object when its reference count becomes 0, then obviously Python's efficiency will be very poor. So when will Python recycle it? This is a good question.
Python will monitor how many new objects it has created and how many objects' reference counters have become 0. The difference between the two values will be compared with the threshold. If it is greater than the threshold, the memory will begin to be garbage. Recycling destroys objects with reference counters of 0.
Advantages: simple real-time performance, disadvantages: maintaining reference counting consumes resources and circular references.
Generational Recycling
In order to improve efficiency, there are many objects that still exist after being cleaned many times. It can be considered that such objects do not need to be recycled frequently and can be It is divided into different collections, and each collection has different recycling intervals. Simply put, this is python's generational recycling.
To be specific, garbage in Python is divided into 1, 2, and 3 generations. Objects in the 1st generation will be cleaned every time they are recycled. When the objects referenced after cleaning still exist, they will Enter the second-generation collection. In the same way, the objects that exist when the second-generation collection is cleaned will enter the second-generation collection.
How to allocate the cleaning time of each collection? The first generation garbage will be cleaned first. After the first generation garbage is cleaned 10 times, the second generation garbage will be cleaned once. After the second generation garbage is cleaned 10 times, the second generation garbage will be cleaned.
Mark Clear
Allocate on demand. When the memory is not enough, start from the register and the reference on the program stack, traverse the object, and mark the traversed object. Unmarked objects are then cleared from memory.
For more Python related technical articles, please visit the Python Tutorial column to learn!
The above is the detailed content of What is Python's memory management mechanism?. For more information, please follow other related articles on the PHP Chinese website!