Shulou(Shulou.com)06/01 Report--

This article mainly explains "what is the principle of PyPy". The content of the explanation is simple and clear, and it is easy to learn and understand. Please follow the editor's train of thought to study and learn "what is the principle of PyPy"?

Project background

PyPy is a project created by Python developers to better Hack Python. In addition, PyPy is more flexible and easier to use and test than CPython to develop ways to implement specific functions in different situations, which can be easily implemented. The goal of the project is to make it easier for PyPy to adapt to individual projects and facilitate tailoring than the Python implemented in C #.

Current status of the project

PyPy is developed by Armin Rigo, the dynamic compiler of Python language, and is the successor project of Psyco. The purpose of PyPy is to do dynamic compilation that Psyco does not do.

PyPy started out as a research project. However, the development is very mature, and after the release of version 1.0 Release in mid-2007, the focus is on whether there will be a version available for production in 2008.

It supports all the core parts of the Python language and most of the Python language standard library function modules, and passed the test suite of the Python language. The difference with CPython can be seen in its compatibility page.

PyPy also provides JIT compiler and sandboxie functions, so it runs faster than CPython and can safely run some untrusted code. PyPy also has a separate version that supports tasklets.

In addition, PyPy also has a nightly build for developers to test.

PyPy began to receive assistance from the European Union as a Specific Targeted Research Projects (field-specific research project) from December 2004 to March 2007.

The principle of PyPy

We compare the principle of Python with the execution process of PyPy through an example [2]

You had a python code:

Def add (x, y):

Return x + y

Then the CPython executes something like this (pseudo code):

If instance_has_method (x,'_ _ add__') {

There are a lot of judgments about different types of y in return call (x,'_ _ add__', y) / / x.judgments

} else if isinstance_has_method (super_class (x),'_ _ add__' {

Return call (super_class,'_ _ add__', y)

} else if isinstance (x, str) and isinstance (y, str) {

Return concat_str (x, y)

} else if isinstance (x, float) and isinstance (y, float) {

Return add_float (x, y)

} else if isinstance (x, int) and isinstance (y, int) {

Return add_int (x, y)

} else...

You can see that because of the dynamic type of Python, it takes so many judgments in a simple function to execute correctly. And then it's not over. Do you think the function in which two integers are added is the x + y in C language? Naive . In fact, an int in Python is probably a structure like this (also pseudocode, the real situation is more complicated than this):

Struct {

Prev_gc_obj * obj

Next_gc_obj * obj

Type int

Value int

... Other fields

}

Then each int is such a structure, or dynamically assigned to put on the heap, the value in it can not be changed, that is to say, if you add 2000 to the structure of 1000, you get 3000 of this structure, and you have to go to heap to malloc a structure. CPython does this honestly every time, even if you only pass two integers every time you call the add function. Then when pypy executes the add function a hundred times, I find that your TM only passes in two integers each time, so why do I bother to do so many calculations for you each time, so I generate a function similar to C on the spot:

Int add_int_int (int x, int y) {

Return x + y

}

Then compile it into machine code on the spot, and then the next time you call add (1, 2), you directly call this "Native" function. Do you think your Pypy is fast? The above process is called Just In Time compilation, or JIT, which is definitely faster than CPython. Of course, JIT also has many problems, such as compiling itself takes a lot of time, if this code is only executed once, it takes 1s, but it takes 10s to compile it, then the loss of JIT outweighs the gain. So many JIT implementations interpret execution and then determine that a piece of code is often executed before compiling. And divided into multi-layer JIT, relatively primary to the compiled machine code does not do more complex optimization and so on.

Thank you for your reading, the above is the content of "what is the principle of PyPy", after the study of this article, I believe you have a deeper understanding of what the principle of PyPy is, and the specific use needs to be verified in practice. Here is, the editor will push for you more related knowledge points of the article, welcome to follow!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.