Compiling high-level scripting languages to performant code

Dakkak, Abdul Majed

Permalink

https://hdl.handle.net/2142/108715 Copy

Description

Title

Compiling high-level scripting languages to performant code

Author(s)

Dakkak, Abdul Majed

Issue Date

2020-07-17

Director of Research (if dissertation) or Advisor (if thesis)

Hwu, Wen-Mei W

Doctoral Committee Chair(s)

Hwu, Wen-Mei W

Committee Member(s)

• Wickham-Jones, Charles Thomas
• Adve, Vikram S
• Padua, David A

Department of Study

Computer Science

Discipline

Computer Science

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• compiler
• high-performance
• high-level
• scripting
• symbolic
• typesystem
• wolfram
• mathematica
• cas

Abstract

The popularity of data- and scientific-oriented applications, the abundance of on-demand compute resources, and the scarcity of domain expert programmers have given rise to high- level scripting languages. These high-level scripting languages offer a fast way to translate ideas into code, but tend to incur a heavy performance overhead. To alleviate the performance penalty, each implementation of these languages often offer a compilation path to a subset of the language. In this thesis, we present the design and implementation of the Wolfram Language compiler, the production compiler for the Wolfram Language. We show how popular language features and runtime behavior, expected by Wolfram Language developers, are efficiently implemented within the compiler. We then show how the compiler provides a friction-less path to migrate programs from the interpreter to the compiler. We evaluate the compiler and show that the compiled code matches the performance of a highly tuned hand-written C code. Unlike existing techniques that compile a subset of the language, the compiler sup- ports the entirety of the Wolfram Language. We show why the compiler is a new model of development for programmers and showcase some applications of the compiler. The compiler has been released as a prominent feature of the Wolfram Engine, is readily available to developers, and is used by internal and external users to drive Wolfram Language features and implementations.

Graduation Semester

2020-08

Type of Resource

Thesis

Permalink

http://hdl.handle.net/2142/108715

Copyright and License Information

Copyright 2020 Abdul Dakkak

Owning Collections