A Novel Approach for a Specific Optimization Problem in Compilers
Hai Lin, Baoliang Mu
Abstract
Open-access reader
Hai Lin, Baoliang Mu
Abstract
Open-access reader
In this paper, we consider the problem of constant propagation.If some variable can only take a constant value in a given program, then we can replace all the occurrences of that variable with the constant value.That does not affect the semantics of the program, the program can be accelerated at run time.In order to present our method, first we propose a tiny language.We give the full syntax and semantics of that language.All the optimization is done on the source code of this language.We then present our optimization method as a set of inference rules.
A significance statement is not available in the OpenAlex record.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
In this paper, we consider the problem of constant propagation.If some variable can only take a constant value in a given program, then we can replace all the occurrences of that variable with the constant value.That does not affect the semantics of the program, the program can be accelerated at run time.In order to present our method, first we propose a tiny language.We give the full syntax and semantics of that language.All the optimization is done on the source code of this language.We then present our optimization method as a set of inference rules.
Key concepts: Computer science, Compiler, Programming language, Dead code elimination, Program optimization, Program transformation, Constant (computer programming), Optimizing compiler