The International Conference for High Performance Computing, Networking, Storage and Analysis
X10 at Petascale.
Authors: Olivier Tardieu (IBM Research), Benjamin Herta (IBM Research), David Cunningham (IBM Research), David Grove (IBM Research), Prabhanjan Kambadur (IBM Research), Vijay A. Saraswat (IBM Research), Avraham Shinnar (IBM Research), Mikio Takeuchi (IBM Research), Mandana Vaziri (IBM Research)
Abstract: X10 is a high-performance, high-productivity programming language
aimed at large-scale distributed and shared-memory parallel
applications. It is based on the Asynchronous Partitioned Global
Address Space (APGAS) programming model, supporting the same
fine-grained concurrency mechanisms within and
across nodes.
We demonstrate that X10 delivers solid performance at petascale by
running (weak scaling) eight application kernels on an IBM Power 775
supercomputer utilizing up to 55680 Power7 cores (1.7 Pflop/s). We
sketch advances in distributed termination detection, distributed
load balancing, and use of high-performance interconnects that
enable X10 to scale out to thousands of nodes.
