Enabling technologies and software for scientific computing
The Innovative Computing Laboratory (ICL) aspires to be a world leader in enabling technologies and software for scientific computing. Our vision is to provide high performance tools to tackle science’s most challenging problems and to play a major role in the development of standards for scientific computing in general.
ICL is a research laboratory in the College of Engineering at the University of Tennessee and serves as the cornerstone laboratory of the Center for Information Technology Research (CITR), one of UT’s nine Centers of Excellence.
The Distributed Parallel Linear Algebra Software for Multicore Architectures (DPLASMA) package is the leading implementation of a dense linear algebra package for distributed heterogeneous systems. It is designed to deliver sustained performance for distributed systems where each node features multiple sockets of multicore processors, and if available, accelerators like GPUs or Intel Xeon Phi. DPLASMA achieves this objective by deploying the Parallel Linear Algebra Software for Multicore Architectures (PLASMA) algorithms on distributed memory systems by leveraging the state-of-the-art PaRSEC runtime.
In addition to traditional ScaLAPACK data distribution, DPLASMA provides interfaces for users to expose arbitrary data distributions. The algorithms transparently operate on local data, or introduce implicit communications to resolve dependencies, removing the burden of initial data re-shuffle, and providing to the user a novel approach to address load balance.