Main Content

fxpopt

Optimize data types of a system

collapse all in page

Syntax

result = fxpopt(model, sud, options)

Description

example

result = fxpopt(model, sud, options) optimizes the data types in the model or subsystem specified by sud in the model, model, with additional options specified in the fxpOptimizationOptions object, options.

Examples

collapse all

Open Script

This example shows how to optimize the data types used by a system based on specified tolerances.

To begin, open the system for which you want to optimize the data types.

model = 'ex_auto_gain_controller';
sud = 'ex_auto_gain_controller/sud';
open_system(model)

Create an fxpOptimizationOptions object to define constraints and tolerances to meet your design goals. Set the UseParallel property of the fxpOptimizationOptions object to true to run iterations of the optimization in parallel. You can also specify word lengths to allow in your design through the AllowableWordLengths property.

opt = fxpOptimizationOptions('AllowableWordLengths', 10:24, 'UseParallel', true)

opt = 

  fxpOptimizationOptions with properties:

           MaxIterations: 50
                 MaxTime: 600
                Patience: 10
               Verbosity: High
    AllowableWordLengths: [10 11 12 13 14 15 16 17 18 19 20 21 22 23 24]
             UseParallel: 1

   Advanced Options
         AdvancedOptions: [1×1 struct]

Use the addTolerance method to define tolerances for the differences between the original behavior of the system, and the behavior using the optimized fixed-point data types.

tol = 10e-2;
addTolerance(opt, [model '/output_signal'], 1, 'AbsTol', tol);

Use the fxpopt function to run the optimization. The software analyzes ranges of objects in your system under design and the constraints specified in the fxpOptimizationOptions object to apply heterogeneous data types to your system while minimizing total bit width.

result = fxpopt(model, sud, opt);

Starting parallel pool (parpool) using the 'local' profile ...
Connected to the parallel pool (number of workers: 4).
	+ Preprocessing
	+ Modeling the optimization problem
		- Constructing decision variables
	+ Running the optimization solver
Analyzing and transferring files to the workers ...done.
		- Evaluating new solution: cost 180, does not meet the tolerances.
		- Evaluating new solution: cost 198, does not meet the tolerances.
		- Evaluating new solution: cost 216, does not meet the tolerances.
		- Evaluating new solution: cost 234, does not meet the tolerances.
		- Evaluating new solution: cost 252, does not meet the tolerances.
		- Evaluating new solution: cost 270, does not meet the tolerances.
		- Evaluating new solution: cost 288, does not meet the tolerances.
		- Evaluating new solution: cost 306, meets the tolerances.
		- Evaluating new solution: cost 324, meets the tolerances.
		- Evaluating new solution: cost 342, meets the tolerances.
		- Evaluating new solution: cost 360, meets the tolerances.
		- Evaluating new solution: cost 378, meets the tolerances.
		- Evaluating new solution: cost 396, meets the tolerances.
		- Evaluating new solution: cost 414, meets the tolerances.
		- Evaluating new solution: cost 432, meets the tolerances.
		- Updated best found solution, cost: 306
		- Evaluating new solution: cost 304, meets the tolerances.
		- Evaluating new solution: cost 304, meets the tolerances.
		- Evaluating new solution: cost 301, meets the tolerances.
		- Evaluating new solution: cost 305, does not meet the tolerances.
		- Evaluating new solution: cost 305, meets the tolerances.
		- Evaluating new solution: cost 301, meets the tolerances.
		- Evaluating new solution: cost 299, meets the tolerances.
		- Evaluating new solution: cost 299, meets the tolerances.
		- Evaluating new solution: cost 296, meets the tolerances.
		- Evaluating new solution: cost 299, meets the tolerances.
		- Evaluating new solution: cost 291, meets the tolerances.
		- Evaluating new solution: cost 296, does not meet the tolerances.
		- Evaluating new solution: cost 299, meets the tolerances.
		- Evaluating new solution: cost 300, meets the tolerances.
		- Evaluating new solution: cost 296, does not meet the tolerances.
		- Evaluating new solution: cost 301, meets the tolerances.
		- Evaluating new solution: cost 303, meets the tolerances.
		- Evaluating new solution: cost 299, meets the tolerances.
		- Evaluating new solution: cost 304, does not meet the tolerances.
		- Evaluating new solution: cost 300, meets the tolerances.
		- Updated best found solution, cost: 304
		- Updated best found solution, cost: 301
		- Updated best found solution, cost: 299
		- Updated best found solution, cost: 296
		- Updated best found solution, cost: 291
		- Evaluating new solution: cost 280, meets the tolerances.
		- Evaluating new solution: cost 287, meets the tolerances.
		- Evaluating new solution: cost 288, does not meet the tolerances.
		- Evaluating new solution: cost 287, does not meet the tolerances.
		- Evaluating new solution: cost 283, meets the tolerances.
		- Evaluating new solution: cost 283, does not meet the tolerances.
		- Evaluating new solution: cost 262, does not meet the tolerances.
		- Evaluating new solution: cost 283, does not meet the tolerances.
		- Evaluating new solution: cost 282, does not meet the tolerances.
		- Evaluating new solution: cost 288, meets the tolerances.
		- Evaluating new solution: cost 289, meets the tolerances.
		- Evaluating new solution: cost 288, meets the tolerances.
		- Evaluating new solution: cost 290, meets the tolerances.
		- Evaluating new solution: cost 281, does not meet the tolerances.
		- Evaluating new solution: cost 286, does not meet the tolerances.
		- Evaluating new solution: cost 287, meets the tolerances.
		- Evaluating new solution: cost 284, meets the tolerances.
		- Evaluating new solution: cost 282, meets the tolerances.
		- Evaluating new solution: cost 285, does not meet the tolerances.
		- Evaluating new solution: cost 277, meets the tolerances.
		- Updated best found solution, cost: 280
		- Updated best found solution, cost: 277
		- Evaluating new solution: cost 272, meets the tolerances.
		- Evaluating new solution: cost 266, meets the tolerances.
		- Evaluating new solution: cost 269, meets the tolerances.
		- Evaluating new solution: cost 271, does not meet the tolerances.
		- Evaluating new solution: cost 274, meets the tolerances.
		- Evaluating new solution: cost 275, meets the tolerances.
		- Evaluating new solution: cost 274, does not meet the tolerances.
		- Evaluating new solution: cost 275, meets the tolerances.
		- Evaluating new solution: cost 276, does not meet the tolerances.
		- Evaluating new solution: cost 271, meets the tolerances.
		- Evaluating new solution: cost 267, meets the tolerances.
		- Evaluating new solution: cost 270, meets the tolerances.
		- Evaluating new solution: cost 272, meets the tolerances.
		- Evaluating new solution: cost 264, does not meet the tolerances.
		- Evaluating new solution: cost 265, does not meet the tolerances.
		- Evaluating new solution: cost 269, meets the tolerances.
		- Evaluating new solution: cost 270, meets the tolerances.
		- Evaluating new solution: cost 269, meets the tolerances.
		- Evaluating new solution: cost 276, meets the tolerances.
		- Evaluating new solution: cost 274, meets the tolerances.
		- Updated best found solution, cost: 272
		- Updated best found solution, cost: 266
	+ Optimization has finished.
		- Neighborhood search complete.
		- Maximum number of iterations completed.
	+ Fixed-point implementation that met the tolerances found.
		- Total cost: 266
		- Maximum absolute difference: 0.087035
		- Use the explore method of the result to explore the implementation.

Use the explore method of the OptimizationResult object, result, to launch Simulation Data Inspector and explore the design containing the smallest total number of bits while maintaining the numeric tolerances specified in the opt object.

 explore(result);

You can revert your model back to its original state using the revert method of the OptimizationResult object.

 revert(result);

Input Arguments

collapse all

Name of the model containing the system that you want to optimize.

Data Types: char

Model or subsystem whose data types you want to optimize, specified as a character vector containing the path to the system.

Data Types: char

fxpOptimizationOptions object specifying additional options to use during the data type optimization process.

Output Arguments

collapse all

Result of the optimization, returned as an OptimizationResult object. Use the explore method of the object to open the Simulation Data Inspector and view the behavior of the optimized system. You can also explore other solutions found during the optimization that may or may not meet the constraints specified in the fxpOptimizationOptions object, options.

Version History

Introduced in R2018a

See Also

Classes

Functions

Topics