First steps to compiling Matlab to X10

Abstract

Matlab is a popular dynamic array-based language commonly used by students, scientists and engineers, who appreciate the interactive development style, the rich set of array operators, the extensive builtin library, and the fact that they do not have to declare static types. Even though these users like to program in Matlab, their computations are often very computeintensive and are potentially very good applications for high-performance languages such as X10.To provide a bridge between Matlab and X10, we are developing MiX10, a source-tosource compiler that translates Matlab to X10. This paper provides an overview of the initial design of the MiX10 compiler, presents a template-based specialization approach to compiling the builtin Matlab operators, and provides translation rules for the key sequential Matlab constructs with a focus on those which are challenging to convert to semantically-equivalent X10. An initial core compiler has been implemented, and preliminary results are provided.

A MATLAB subset to C compiler targeting embedded systems

“…Recently and orthogonally to this work, there have been approaches to translate MATLAB (and its main clones) to languages suitable to multicore and/or GPU architectures. For instance, MiX10 compiles MATLAB to IBM's X10 language, designed for high‐performance computing. Their work takes advantage of MATLAB's parfor and X10 parallel features.…”

Mix10

“…The set of benchmarks used for our experiments consists of benchmarks from various sources; Most of them are from related projects like FALCON [19] and OTTER [18], Chalmers university of Technology 9 , "Mathworks' central file excahange" 10 , and the presentation on parallel programming in MATLAB by Burkardt and Cliff 11 . This set of benchmarks covers the commonly used MATLAB features like arrays of different dimensions, loops, use of numerical functions like random number generation, trigonometric operations, and array operations like transpose and matrix multiplication.…”

“…It uses the Tamer IR it receives from the Tamer to drive the code generation, but for expressions referring to built-in MATLAB functions it interacts with the Built-in Handler which used the built-in template file we provide. We have described the functioning of the built-in handler and a very basic code generation strategy for ordinary sequential constructs at the X10 workshop [11]. In this paper we focus on the challenges in generating efficient X10 code whose performance is comparable to state-of-the-art tools that generate more traditional imperative languages like C and Fortran.…”

“…For readers who are unfamiliar with the "wild" nature of arrays in MATLAB, we recommend reading Section 5.5 of our previous paper, First steps to compiling MATLAB to X10[11].…”