A MATLAB subset to C compiler targeting embedded systems

Bispo, João; Cardoso, João M. P.

doi:10.1002/spe.2408

Cited by 11 publications

(8 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…The repository used for this work is the one used to test the Matlab compiler by Bispo and Cardoso (), comprising 35,193 files organised by toolboxes and covering various application domains. The 28,000 m‐files were downloaded from Sourceforge , and 2,000 m‐files were downloaded from GitHub.…”

Section: Methodology For Data Analysis and Explorationmentioning

confidence: 99%

Analysis of a token density metric for concern detection in Matlab sources using UbiSOM

2018

View full text Add to dashboard Cite

Matrix and data manipulation programming languages are an essential tool for data analysts. However, these languages are often unstructured and lack modularity mechanisms. This article presents a knowledge discovery approach for studying manifestations of the lack of modularity support in that sort of languages. The study is focused on Matlab, as a well‐established representative of those languages. We present a technique for the automatic detection and quantification of concerns in Matlab and their exploration in a code base. The Ubiquitous Self Organizing Map (UbiSOM) is used to perform exploratory data analysis over concerns detected in a, possibly changing, repository of Matlab files. The UbiSOM is quite effective in detecting patterns of co‐occurrence of multiple concerns. To illustrate the technique, a repository comprising over 35,000 Matlab files is analysed. The results show that the use of Token Density metrics in conjunction with UbiSOM enables the detection of patterns of co‐occurrence of multiple concerns in m‐files.

show abstract

Section: Methodology For Data Analysis and Explorationmentioning

confidence: 99%

Analysis of a token density metric for concern detection in Matlab sources using UbiSOM

2018

View full text Add to dashboard Cite

show abstract

“…Compared with reality, the solution to the problem will inevitably have a certain deviation. Therefore, a conceptual evaluation of the performance of the machine model is performed [9]. Accuracy refers to the proportion of correct predictions (positive and negative) in all predictions, as shown in…”

Section: Introductionmentioning

confidence: 99%

Application of Big Data Analysis Techniques in Sports Training and Physical Fitness Analysis

Chen

Guo

et al. 2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

With the development of sports and information technology, people use mathematical tools and computer technology to study sports data and mine the intrinsic value of sports data. Statistical methods are the most widely used to achieve this goal. The research purpose of sports effect evaluation research is to understand the impact of sports on physical fitness through mining and analysis of sports data and to provide theoretical guidance for the public to participate in fitness activities scientifically and effectively. At present, in the study of combining individual performance test data, the research on the standardization of physical fitness monitoring data for sports training is relatively scarce. Therefore, based on the background of big data, this paper integrates the existing data standardization work and designs a plan for the standardization of physical fitness monitoring data for sports training. Combined with machine learning, data preprocessing is performed to obtain the data required by the machine model. The comprehensive physical fitness rating model and the recommendation model are established to realize the development of physical fitness monitoring service applications. In the experiment, compared with the three classical methods, the results show that the classification accuracy of this paper is 4% higher than that of other algorithms, which can more intuitively reflect the characteristic samples of sports training. In this paper, the data mining and analysis technology based on feature indicators in the mining and application of sports data has great application value for human fitness guidance and has certain research value and market application prospects.

show abstract

“…In this paper, we show how to automatically generate C and OpenCL implementations from high‐level programs written in a subset of MATLAB. Our work uses the MATISSE compiler framework, 9,10 and proposes an enhanced Intermediate Representation (IR) and optimization framework, with preliminary versions previously presented by Reis et al 11 . MATISSE is a source‐to‐source compilation framework that receives as input a program in a subset of MATLAB and is able to generate transformed MATLAB code, or to transform MATLAB into C code.…”

Section: Introductionmentioning

confidence: 99%

Compilation of MATLAB computations to CPU/GPU via C/OpenCL generation

Reis

Bispo

Cardoso

2020

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

In order to take advantage of the processing power of current computing platforms, programmers typically need to develop software versions for different target devices. This task is time-consuming and requires significant programming and computer architecture expertise. A possible and more convenient alternative is to start with a single high-level description of a program with minimum implementation details, and generate custom implementations according to the target platform. In this paper, we use MATLAB as a high-level programming language and propose a compiler that targets CPU/GPU computing platforms by generating customized implementations in C and OpenCL. We propose a number of compiler techniques to automatically generate efficient C and OpenCL code from MATLAB programs. One of such compiler techniques relies on heuristics to decide when and how to use Shared Virtual Memory (SVM). The experimental results show that our approach is able to generate code that provides significant speedups (eg, geometric mean speedup of 11× for a set of simple benchmarks) using a discrete GPU over equivalent sequential C code executing on a CPU. With more complex benchmarks, for which only some code regions can be parallelized, and are thus offloaded, the generated code achieved speedups of up to 2.2×. We also show the impact of using SVM, specifically fine-grained buffers, and the results show that the compiler is able to achieve significant speedups, both over the versions without SVM and with naïve aggressive SVM use, across three CPU/GPU platforms.

show abstract

A MATLAB subset to C compiler targeting embedded systems

Cited by 11 publications

References 38 publications

Analysis of a token density metric for concern detection in Matlab sources using UbiSOM

Analysis of a token density metric for concern detection in Matlab sources using UbiSOM

Application of Big Data Analysis Techniques in Sports Training and Physical Fitness Analysis

Compilation of MATLAB computations to CPU/GPU via C/OpenCL generation

Contact Info

Product

Resources

About