The Scalable System Software department traces its roots back to the early days of distributed memory massively parallel processing (MPP) systems of the late 1980’s. During this time, Sandia established the viability of MPP systems, such as the nCUBE-10 and the Intel Paragon, in solving mission-critical applications using modeling and simulation. The department grew out of the need to design, develop, and deploy more efficient system software focused on the meeting the performance and scalability demands of these applications running on the largest and fastest computing systems in the world. The group became firmly established in the early 1990’s when researchers at Sandia partnered with the University of New Mexico to develop a customized system software environment based on a lightweight compute node operating system designed specifically for large-scale, distributed memory, message-passing machines. This initial lightweight kernel environment was successfully deployed on several large production systems at Sandia and eventually evolved into the operating system that ran on the compute nodes of the world’s first general-purpose parallel computer to achieve a teraFLOPS, the Intel ASCI Red system. As parallel computing architectures and applications and have continued to evolve, the department has expanded into several other system software areas around operating systems, but has continued to focus on addressing the needs of extreme-scale systems and applications.
Scalable System Software