Home > Products & Services > Development Tools > Code Development > SpeedDSP

Code Development Tools

SpeedDSP - Optimized Libraries for SHARC

Additional SpeedDSP Info SpeedDSP datasheet SpeedDSP function list SpeedDSP function timing BittWorks datasheet BittWorks software tools

Features

• Supports ADSP-TS101S, 2116x and 2106x SHARC DSPs
• Speed DSP blockset availalbe to integrate with SharcLAB MATLAB® interface
• SIMD optimization for the ADSP-2116x
• Highly optimized C-callable routines
• Includes the following libraries:
  • FFT Library
  • Filter Library
  • Math Library
  • Matrix Library
  • Sorting Library
  • Utility Library
  • Windowing Library
  • Vector Library

SpeedDSP is a collection of highly optimized routines for the ADSP-TS101S TigerSHARC and ADSP-21xxx SHARC DSP chips including SIMD optimizations for the ADSP-2116x family of DSPs. The functions are written in SHARC assembly language and are callable from high-level languages such as C. SpeedDSP includes functions for manipulating large arrays of floating-point numbers and for performing Fast Fourier Transforms (FFTs), windowing, statistics, sorting, histogramming, trigonometry, and timing. Since the functions in the library are coded in SHARC assembly language and take full advantage of the SHARC architecture, they are much faster than high-level language implementations, delivering optimum speed and performance. SpeedDSP integrates easily with the Analog Devices C compiler and is completely compatible with the program/data memory specifiers and the complex data type.

SpeedDSP Blockset for Sharc®LAB MATLAB Target

A blockset is available to integrate BittWare's SpeedDSP optimized math libraries with BittWare's SharcLAB, allowing for highly-optimized DSP code generation. SharcLAB is a MATLAB® target for real-time prototyping and deployment of DSP applications on SHARC DSPs. Developed exclusively for BittWare by SDL, SharcLAB allows you to use MATLAB to graphically generate code for BittWare's SHARC DSP boards. SharcLAB integrates seamlessly with MATLAB, Simulink®, Stateflow®, and Real-Time Workshop® to allow a nearly automatic transition from MATLAB-based algorithm development to executable DSP code.

GT-3U-CPCI: Ruggedized Hybrid Signal Processing 3U CompactPCI Board ATLANTiS Framework: I/O Switching and Processing Device BittWorks: Complete DSP Software Development Suite Extending the Flexibility of MicroTCA using Reconfigurable AMCs