The Performance and Architecture Laboratory (PAL) operates at the forefront of advanced system architectures and applications. PAL investigates the issues that contribute to optimal application and computer system performance for current extreme-scale systems and for future architectures. We develop, refine, and improve a variety of original techniques for performance modeling of architectures and applications. We apply these techniques to codes at LANL and to other significant workloads such as those of DARPA and the DOE Office of Science. We have active projects in performance analysis and modeling, system design for advanced architectures, high-performance interconnection networks, scheduling, and system software R&D.
Our core competencies are in parallel systems architecture, parallel computing, system software, networking, and system and application performance analysis and modeling.
Team members have received the Gordon Bell Prize for high performance parallel computing and the R&D 100 Award for technological innovation. Our publications have won best paper awards at some of the most prestigious conferences in the field, including Supercomputing and IPDPS.
Our team is involved in research and development activities, in the area of system and application performance, for some of the largest supercomputer installations in the world. We also operate a number of testbeds for hardware and software experiments.
We are contributing to major programmatic efforts at Los Alamos National Laboratory, and we collaborate with other national laboratories, with academia, and with computer hardware vendors.