Home > Products > MATLAB Based Products > Agility RMS
AGILITY RMS: Rapid MATLAB Simulator
Accelerated MATLAB Algorithm Execution
The increasing complexity of signal processing algorithms drives the need for faster simulation for validation. Agility RMS accelerates algorithms written in M-code. The accelerated simulation speed approaches the simulation speed of hand-coded C but is automatically generated from the source MATLAB®. Agility RMS automatically converts user's MATLAB to C into a compiled MEX-file, delivering compiled code performance seamlessly within the MATLAB environment.
Agility RMS (Rapid MATLAB Simulator) features fast simulation of MATLAB code and fixed-point analysis tools for quantization and datapath optimization.
Blazing Fast Simulation
Agility RMS converts M-code functions into a compiled library that uses
the MATLAB standard MEX interface. This library produces results identical
to the original algorithm, but executes much faster. Agility RMS is integrated
seamlessly within MATLAB.
Agility RMS uses proprietary, state-of-the-art compilation technology, including interprocedural analysis and optimizations, type inferencing and advanced loop and dependency analysis. The result: faster execution of MATLAB algorithms.
Because MATLAB does use compiled code for some critical functions, like
FFT, the performance gain RMS will provide varies. The graph to the right
gives some guidelines for levels of acceleration for different styles of
MATLAB code.
In addition, a table of example acceleration factors for some common signal processing applications is shown.
Easy to Use
Agility RMS is designed to be used with existing MATLAB code. Based on an execution profile of a program, it will automatically convert candidate functions into a MEX-file. RMS will also recommend code modifications for additional acceleration.
Agility MCS supports a large subset of the MATLAB language, including many of the commonly used features for modeling signal processing systems:
- Operators: arithmetic, matrix, logical, and bitwise operators
- Data types: real/complex, logical, structures
- Shapes: scalars, vectors, matrices, n-dimensional arrays
- Persistent and global variables
- Flow control constructs: if-then-else, for loops, while loops, switch
- Common mathematical functions (e.g. sqrt, sin, atan)
- Numerous built-in functions
- Many plotting functions
- Most of the Signal Processing and Communications toolboxes and Image Processing toolbox.
Additional Benefits for Fixed-Point Designers
Without Agility fast simulation technology, simulation of fixed-point algorithms can be unacceptably slow. As a result, developers usually do not quantize their design until they have migrated out of the MATLAB environment.
In addition to blazing fast simulation, Agility RMS also includes analysis tools to dramatically ease the quantization process. These fixed-point analysis tools:
- Display the types of all variables in a MATLAB design
- Recommend quantization based on simulation runs Flag overflows and other potential quantization errors
With Agility RMS, developers can now quantize MATLAB designs and run full system simulations inside the MATLAB environment. Developers can continue to use the MATLAB development environment to develop, optimize, and verify their MATLAB algorithms.
Agility RMS lets users view both floating-point and fixed-point statistics of their MATLAB code.
Key Features
- 5-20x acceleration of MATLAB code
- Seamlessly integrates into MATLAB environment
- Analysis tools ease quantization
Required Products
- MATLAB R12.1 or newer
Supported Products
- MATLAB Communications Toolbox
- MATLAB Image Processing Toolbox
- MATLAB Signal Processing Toolbox
Operating Systems
- Microsoft Windows (2000 or XP)
- 32 or 64 bit Linux (red Hat Enterprise version 3 or 4)
Complementary Products
- Agility MCS (Matlab to C Synthesis)