INDUSTRY

Teja Technologies Inc., the leading supplier of embedded networking and communications software for multi-core processor architectures today announced that Los Alamos National Laboratory (LANL) has created a unique application that achieves line rate graphics compositing for large scale data visualizations using Teja NP, the award-winning software platform for Intel Network Processors (NPUs). Using Teja NP with Intel's IXP2850 NPUs, LANL was able to create this application in just six weeks. By offloading into the network the image compositing function from the data visualization clusters, LANL is able to demonstrate reduced network bandwidth by more than 50 percent while also reducing CPU workload. The resulting benefits to LANL researchers include the ability to perform additional computational tasks, higher bandwidth rendering, and more interactive visualizations. "As more and more users turn to Intel network processors for the flexibility and programmability benefits they offer, the need to quickly create complex software applications that operate at line rate has become paramount," stated David Stepner, Teja's president and CEO. "Los Alamos National Laboratory's use of Intel's NPU with the Teja NP Software Platform is an excellent example of an innovative application created in just six weeks by building their unique functionality on top of our existing software components and platform." The application LANL created using Teja NP was primarily designed to increase the performance of the process known as "scientific visualization," which is the conversion of large amounts of time-dependent data from complex, high performance computer simulations into viewable images that scientists can then evaluate. LANL has extremely complex, CPU-intense visualization requirements, as recent projects have included weapons simulations as well as efforts to better understand complex earth processes, including computer models to predict the spread of a wildfire, global climate processes, movement within the earth's mantle, and the effects of potential asteroid impacts. While interaction with the data generated by such research is critical to LANL's scientists, the sheer volume and complexity of the data sets make it extremely challenging to maintain sufficient performance. As a result, LANL chose the Intel IXP2850 NPU for its performance and programmability and turned to the Teja NP Software Platform, as it is specifically designed to allow developers to quickly create, test and implement software applications for the IXP2XXX NPUs. Teja accomplishes this through a unique approach to developing embedded networking applications that greatly simplifies the creation of software for parallel, multiprocessing architectures. A developer's primary challenge in designing high performance software for multiprocessor systems such as the Intel IXP2850 NPU is in properly partitioning the application code among the multiple processors, allocating shared memory, and building a framework to support scheduling, inter-processor communication and synchronization. Teja NP gives OEMs the ability to define a complete system's application logic in a C-based, hardware-independent model that is separately mapped to the various resources of the target hardware configuration, resulting in the automatic generation of optimized production code for the target system. Teja NP consists of an Application Development Environment, Network Processing Operating System (NPOS), and a library of foundation application building blocks available from Teja and Teja's Independent Software Vendors (ISVs), including TCP Termination, IPv4 Forwarding, IPv6 Forwarding, and ATM. The resulting benefits of using Teja NP include accelerated time-to-market, the reduction of engineering risk and, most importantly, reliable, high-performance products that provide OEMs and other developers with a first-to-market advantage.