Skip Ribbon Commands
Skip to main content
Navigate Up
Sign In
:

Home > Industry News > Posts > GE Brings Improved Developer Productivity and System Performance to High Performance Embedded Computing Applications
February 14

GE Brings Improved Developer Productivity and System Performance to High Performance Embedded Computing Applications

Keywords: Multiple processing, Parallel Processing, Open Standards, Vector Signal, Image Processing, HPEC, SIGINT.

GE Intelligent Platforms announced the availability of Mentor Graphics Sourcery VSIPL++ which has been specifically optimized for the company's range of rugged embedded computing solutions.

The Sourcery VSIPL++ library for GE computing platforms is available from Mentor Graphics, who developed the capability in conjunction with GE.  It integrates with GE's AXISLIB-AVX performance-optimized VSIPL library with more than 600 DSP and math functions specifically targeted at Intel's latest 2nd generation Core platform.

VSIPL and VSIPL++ – Vector Signal Image Processing Library for C and C++ respectively – are open standard application programming interfaces (APIs) used by developers who desire a portable computational middleware for performance-intensive signal- and image processing applications.  GE Intelligent Platforms' AXISLIB-AVX provides a full VSIPL Core 1 profile library with commonly used functions for radar, sonar, signals intelligence (SIGINT), and electronic warfare (EW) applications.  Mentor Graphics' Sourcery VSIPL++ library is a portable implementation of the VSIPL++ API that provides a high-level abstraction layer for math libraries such as AXISLIB-AVX.  With Sourcery VSIPL++ and AXISLIB-AVX, developers can write program code only once and then reuse it on multiple platforms, such as the 2nd Generation Intel Core i7-based SBC624 and DSP280 OpenVPX COTS processors from GE Intelligent Platforms.

VSIPL++ is designed to provide higher performance relative to VSIPL, support for object orientation, support for multi-processor systems, extensibility and simpler syntax.  Multiple processing and parallel processing are increasingly features of the most demanding, most sophisticated military/aerospace high performance embedded computing (HPEC) applications, and VSIPL++ makes the development of such applications easier, faster and less costly.

The Sourcery VSIPL++ library provides a set of more than 240 C++ signal processing functions, while the proprietary compile-time and run-time dispatch engines optimize performance by mapping high level API function calls to the best performing low level hardware-optimized library functions available on the target platform.  The Sourcery VSIPL++ library provides detailed code execution and performance summary information logging to help users understand how applications are running across computing resources.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ARC Advisory Group

Comments

There are no comments for this post.
 

 Related Reports