Target Specific Optimizations

C Language Extensions / Intrinsics

Infineon® TriCore®

SupportC89/C90 ANSI Target Function LibraryInfineon TriCore ISO Target Function LibraryInfineon TriCore Target Function Library (ERT Only)
ANSI SupportYesYesYes
ISO Support YesYes
Saturated Arithmetic Support  Yes

ISO/IEC 9899:1999 Math Library.   The target function library Infineon TriCore ISOuses the TASKING® ISO/IEC 9899:1999 Math Library to implement floating-point mathematical function blocks (e.g. trigonometric functions, log functions). Using these target optimizations improves the performance of applications performing floating-point mathematical operations.

When using these target optimizations, the regular Real-Time Workshop® implementation for many ANSI floating-point mathematical operations is replaced by the ISO equivalent. These functions behave identically to the regular Real-Time Workshop implementation and can be verified using processor-in-the-loop cosimulation.

You can use the Infineon TriCore ISO target function library with ERT or GRT system target files.

To enable the math library for the optimization of floating-point mathematical operations, select Infineon TriCore ISO for the Real-Time Workshop option Target function library (on the Interface pane of the Configuration Parameters dialog box).

Saturated Arithmetic.   The target function library Infineon TriCore includes all ISO optimizations and also saturated arithmetic optimizations. The target function library Infineon TriCore is only available for ERT system target files.

You can use TASKING compiler extensions and intrinsic functions for saturated arithmetic. These target optimizations can increase execution speed up to 18 times for saturated arithmetic operations. The use of these target optimizations will improve the performance of most applications performing saturated arithmetic operations. It is therefore recommended to enable the optimizations.

When using these target optimizations, the regular Real-Time Workshop implementation for many saturated arithmetic operations are replaced by calls to target optimized inlined functions. The behavior of these functions is identical to the regular Real-Time Workshop implementation and can be verified using processor-in-the-loop cosimulation (see Processor-in-the-Loop (PIL) Cosimulation).

To enable TASKING compiler extensions and intrinsic functions for the optimization of saturated arithmetic, select Infineon TriCore for the Real-Time Workshop option Target function library (on the Interface pane of the Configuration Parameters dialog box).

General

Depending on your toolset, your the TASKING compiler may support C language extensions or intrinsics to help optimize in some of the following areas:

Please see your TASKING documentation for details. You can use these language extensions in your own Simulink® blocks and / or custom code.

Target Optimized Libraries for Infineon® XC166 and Infineon® TriCore®

The following optimized libraries are available for the processors supported by Embedded IDE Link™ TS software, and can be used to create optimized Simulink blocks:

Target Optimized FIR / FFT Blocks for the Infineon® TriCore®

Example FIR / FFT blocks that call target optimized Infineon TriLib routines are available on MATLAB Central. These blocks can be over a hundred times faster than the regular blocks in the Signal Processing Blockset™ product.

Search MATLAB Central for details.

  


 © 1984-2008- The MathWorks, Inc.    -   Site Help   -   Patents   -   Trademarks   -   Privacy Policy   -   Preventing Piracy   -   RSS