E-Book, Englisch, 785 Seiten
Altman Accelerating MATLAB Performance
Erscheinungsjahr 2015
ISBN: 978-1-4822-1130-6
Verlag: Taylor & Francis
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)
1001 tips to speed up MATLAB programs
E-Book, Englisch, 785 Seiten
ISBN: 978-1-4822-1130-6
Verlag: Taylor & Francis
Format: PDF
Kopierschutz: Adobe DRM (»Systemvoraussetzungen)
Useful to novices and professionals, this book explains how to improve MATLAB® program speed. Packed with programming tips, the text discusses vectorization, optimization, memory management, and caching, as well as using MEX, GPU, the Parallel Computing Toolbox, and the Profiler. It also considers tradeoffs in tuning, horizontal vs. vertical scalability, and perceived vs. actual performance, among other aspects. Nearly all the material is based on fully documented and supported MATLAB® functionality, plus the book offers access to an active, dedicated website with code examples and online references.
Autoren/Hrsg.
Weitere Infos & Material
Introduction to Performance Tuning
Why Bother?
When to Profile and When Not to Bother
The Iterative Performance-Tuning Cycle
When Should We Stop Optimizing the Code?
Trade-Offs: Performance vs. Maintainability, Robustness, Development Time, and Repeatability
Vertical vs. Horizontal Scalability
Perceived vs. Actual Performance
What Should We Measure and Tune?
Profiling MATLAB® Performance
Profiling Tools Available in MATLAB®
Profiling Techniques
Real-Time Profiling Limitations
Using the Profiler vs. Tic/Toc
MATLAB®'s JIT and Its Effect on Profiling
Standard Programing Techniques
Loop Optimizations
Caching
Smart Checks Bypass
Exception Handling and Performance
Improving Externally-Connected Systems (Database, Filesystem, and Network)
Using Parallelization and the GPU
Parallelization in MATLAB®
The Parallel Computing Toolbox
Using the GPU
Using Multiple Cores and CPUs
Techniques for Effective Parallelization in MATLAB®
Data Analysis Techniques
Selecting the Right Tool for the Job
Outliers Removal
Controlling the Target Accuracy
Coordinate Transformation
Choosing Correct Optimization Parameters
Reducing Problem Complexity
MATLAB®-Specific Techniques
Effects of using Different Storage Types
Vectorization
Object-Oriented MATLAB® and Performance
Using Internal Helper Functions
I/O Aspects
Graphics and GUI Aspects
Using MEX Code
Deployed (Compiled) MATLAB® Programs
Memory-Related Techniques
Why Memory Affects Performance
Pre-Allocation
In-Place Data Manipulations
Using Global and Persistent Variables
Memory Packing
MATLAB® Graphs and GUI
Fast Graph Generation
Updating Graphs in Real Time
GUI Responsiveness
Avoiding Common Pitfalls
Topics and Resources for Further Learning
