PCRC Project Overview New Ideas ¥ Portable and scalable multi-platform runtime support Ð Address translations Ð Communication optimizations ¥ Integrated multilanguage support Ð Common HPF and HPC++ libraries Ð Common compiler data movement interface specification ¥ Interoperable Compilers Impact ¥ Software engineering of distributed systems Ð Framework for composing, executing and measuring software built from programs in multiple languages ¥ Efficient address translations and data movements for distributed memory systems ¥ Portable runtime support for multiprocessors ¥ Testbeds for extending compilers, runtime software, and languages for experimentation and research Schedule Research Issues and Innovations - 1 ¥ Portable and Scalable Multi-platform Runtime Support Efficiently support address translations and data movements when mapping a shared address space onto a multiple processor architecture Develop an integrated runtime support system to carry out address translation and communication optimizations; build this on top of a message passing interface ¥ Integrated Multilanguage Support Allow different programming languages to share data structures that are distributed across the memory hierarchy of a scalable parallel system Ð Develop common code and data descriptors, and routines that operate on them, for supporting data parallelism in HPLs Design a common compiler data movement interface specification that will provide a set of communication standards for compiled code, ranging from very low primitives that exploit special hardware properties to very high level primitives directly coupled the the common array and data structure formats Research Issues and Innovations - 2 ¥ Software Engineering System level test and validation of the Common Runtime System requires development of a methodology and framework for composing, executing and measuring systems built from programs in multiple languages Develop a methodology to allow ease of use, modification, specialization, and extension of the Common Runtime System: support new data structures, new language features, new runtime mechanisms and algorithms, and new message passing or distributed shared memory interfaces. ¥ Interoperable Compilers We will have seven compilers sharing the same runtime system and be available for use and experimentation in multilanguage systems Team Members ¥ Syracuse University ¥ Geoffrey C. Fox, PI ¥ Cooperating Systems ¥ Marina Chen ¥ Harvard University ¥ Thomas Cheatham ¥ Indiana University ¥ Dennis Gannon ¥ Rice University ¥ Ken Kennedy ¥ University of Maryland ¥ Joel Saltz ¥ University of Rochester ¥ Wei Li ¥ University of Texas, Austin ¥ Jim Browne