Tuesday 19 July 2011

Software reduces cost, risk of high performance embedded computing

Delivers flexibility and portability, allows performance potential of multicore processing technology to be realized

GE Intelligent Platforms has released AXIS Version 5.0. AXIS – Advanced Multiprocessor Integrated Software – is GE’s comprehensive development environment that accelerates the design, creation, testing and deployment of complex DSP and multiprocessing platforms for real-time applications such as radar, sonar, communications and image processing.

AXIS increases productivity, maximizes performance, reduces cost, optimizes software code portability and minimizes program risk and time-to-market. It also substantially eases the upgrade path for technology insertions. Version 5 provides support for GE’s latest multicore and many-core hardware platforms including single board computers featuring 2nd Generation Intel® Core™ i7 processors as well as those featuring the latest Power Architecture processors from Freescale, and those featuring NVIDIA’s CUDA GPGPU (general purpose processing on graphics processing units) technology. It also includes a number of new features designed to maximize ease of use, flexibility and performance.

“AXIS Version 5 allows us to provide our customers with a highly integrated hardware and software platform that uniquely enables them to take maximum advantage of the enormous benefits of the latest processing technology from Freescale, Intel and NVIDIA,” said Michael Stern, Product Manager, GE Intelligent Platforms. “Using AXIS, developers can create leading-edge high performance computing applications in less time and at less cost than would otherwise be possible, while maximizing return on investment by creating solutions that are more easily portable from one hardware generation or architecture to another.”

The AXIS software suite includes an integrated graphical user interface. AXISView, provides task level programming for algorithm development and system visualization. The developer exploits an iterative application development cycle allowing the user to benchmark routines and tune the system architecture to meet performance requirements.

In addition, AXISFlow provides a high performance interprocessor communication library for high throughput, low latency data movement across multiple fabrics (PCI Express®, Gigabit Ethernet, 10 Gigabit Ethernet, Serial RapidIO etc.) scaling from one to many CPU nodes across a single board, multiple boards in a single chassis and multiple system chassis.

AXISLib provides DSP and math function libraries to support very high performance signal and data processing routines with a standard VSIPL wrapper as well as a faster RSPL API for optimum performance and portability across multiple processor platforms. Generic C versions of the libraries are also available.

No comments:

Post a Comment