Realization method for parallelization of single-threading program based on analysis of data flow

Abstract

The invention discloses a realization method for the parallelization of the single-threading program based on the analysis of data flow. In the invention, by analyzing the data dependence among the instructions in the single-thread program, the single-threading program is transformed to a multi-threading program, the dependence among the instructions in the single-thread program includes a data dependence and a control dependence, wherein the control dependence is a dependence for the control condition value and a special data dependence. In the process of thread analyzing, the invention can take regard of the balance of the thread communication expense and the thread after analyzing. The invention has the advantage that the different parts of the single-threading program are executed in parallel, thus reducing the program executing time and improving the program executing efficiency. The method for the parallelization of the single-threading program is in particular suitable for the current multi-core structure.

Description

technical field [0001] The invention relates to the field of parallelization of single-threaded programs, in particular to a method for realizing parallelization of single-threaded programs based on data flow analysis. Background technique [0002] With the development of electronic technology, the manufacturing process of electronic chips has been reduced from the original 90nm to 65nm and 45nm, and now there are even 32nm chips. The improvement of the manufacturing process enables chips with a fixed area to integrate more transistors. So that a single chip can be designed into a more complex structure and more computing resources; at the same time, the higher chip integration makes the original practice of increasing the chip operating frequency no longer effective. On the one hand, the circuit delay of the chip makes the frequency of the chip It cannot be improved infinitely. On the other hand, too high frequency brings higher energy consumption, making heat dissipation o...

AI Technical Summary

Problems solved by technology

[0006] No matter how people try to parallelize the program, the dependence before and after the instructions in the program is the most important constraint limiting the parallelism of the program. The establishment of a new programming model may improve the dependence between the programs, but due to the process of the task The dependence between programs cannot be eliminated, so all the research on parallelism is based on the dependence

There are two kinds of dependencies within the program: control dependencies and data dependencies. However, if a certain piece of code needs to be executed in parallel, the reverse dependency is also a problem that must be considered. The reverse dependency exists in the parallelism of the instruction level and the thread level. In parallelism, such as out-of-order execution of processors, speculation and software pipelining

Images