MATLAB Coder Speeds Development Time
April 7, 2011
In an effort to facilitate a faster, more efficient systemdevelopment workflow, MathWorks hasannounced MATLABCoder, a tool that lets design engineers automatically generate readableand portable C or C++ code directly from their MATLAB algorithms.
Click here for larger image. |
The MATLAB technical computing software is widely deployedin the concept development of electronic and communications systems, to createthe algorithms at the core of these platforms. Eventually, however, the MATLABalgorithms need to be converted into C code for the prototyping stage and forintegration into the final product-a step that has been a manual, highlyiterative process and one that could take from days to weeks to perform,depending on the complexity of the program, according to Ken Karnofsky, seniorstrategist at MathWorks. With the introduction of MATLAB Coder, instead ofmanually converting the code and risk the possibility of introducing designerrors and time-to-market delays, the code is automatically converted fromMATLAB to C, greatly speeding up development tasks such as fixed-point designand prototyping, Karnofsky says.
"The difference with manual translation or recoding is thatyou have two versions of the program-one on MATLAB and another in C++," heexplains. "As engineers make refinements, they can get out of sync with eachother. In addition, iterations can be quite time consuming and expensive sinceproblems are discovered later in the process when bugs are harder to fix."
MATLAB Coder will aid in tasks where translation iscritical, including the creation of reusable libraries, real-time prototypingfor creating a standalone executable, to accelerate algorithms to implementcode for embedded processors. The tool supports MEX function generation forfixed-point and floating-point math in addition to an extensive set of MATLABlanguage features for algorithm development, including built-in matrix mathoperators and dynamic array handling.
MathWorks also used the MATLAB Coder announcement toconsolidate some of its other code generation offerings. Simulink Codercombines capabilities previously available in multiple MathWorks offerings, andEmbedded Coderoffers full code generation support for rapid prototyping, real-time testingand production of embedded systems, including links to embedded developmentenvironments.
In separate news, MathWorksannounced that TUV SUD certified Simulink Design Verifier and SimulinkVerification and Validation for use in ISO 26262 compliant developmentprocesses. The organization specializes in independent assessments andcertifications and this latest one will help automotive engineers developinghigh-integrity embedded systems establish the required confidence in the properfunctioning of these verification tools according to their use cases.
About the Author
You May Also Like