Programmable semiconductor device maker Xilinx Inc., has acquired AccelChip Inc., Milpitas CA, a provider of software tools for building digital signal processing (DSP) systems using the MATLAB synthesis environment from The Mathworks Inc., Natick MA.
The AccelChip DSP synthesis tools and DSP library of algorithmic intellectual property (IP) will be offered as part of the Xilinx XtremeDSP products, including System Generator for DSP design tools, rich DSP libraries, and reference designs.
This enables DSP algorithm and system designers who use MathWorks MATLAB and Simulink to create high-performance systems using Xilinx reconfigurable DSPs.
Xilinx says this will accelerate the adoption of application-optimized solutions for digital communications; multimedia, video and imaging (MVI); and defense systems segments of the US$2-billion high-performance DSP market.
“AccelChip provides an excellent strategic fit with our existing DSP products and roadmaps,” Omid Tahernia, Xilinx vice president and general manager, DSP Division, says. “The combination of AccelChip and Xilinx System Generator will help accelerate delivery and adoption of our robust, application-optimized solutions by many more DSP developers.”
“It is clear that most DSP designers want a design flow that starts with MATLAB and Simulink, and also supports modeling in the C programming language,” AccelChip President and CEO Vin Ratford says. “The merging of our tools and libraries of DSP algorithms will provide a world-class platform for a host of new DSP applications, and will create new design wins for Xilinx reconfigurable DSP silicon.”
AccelChip’s recent 2006.1 software release adds automatic generation of C++ verification models from fixed-point MATLAB models, enabling rapid algorithm and architectural design exploration based on performance, power, area, and cost.
The Xilinx System Generator for DSP 8.1 suite extends hardware co-simulation capability to support high-speed Ethernet (10/100/Gigabit) connectivity as well as providing significant enhancements in generating DSP co-processors for embedded applications.
