Presentation
SLATE: Design of a Modern Distributed and Accelerated Linear Algebra Library
SessionLinear Algebra Algorithms
Event Type
Paper
TP
Algorithms
I/O
Linear Algebra
Parallel Programming Languages, Libraries, and Models
Performance
Task-based programming
TimeTuesday, 19 November 20194:30pm - 5pm
Location401-402-403-404
DescriptionThe SLATE (Software for Linear Algebra Targeting Exascale) library is being developed to provide fundamental dense linear algebra capabilities for current and upcoming distributed high-performance systems, both accelerated CPU-GPU-based and CPU-based. SLATE will provide coverage of existing ScaLAPACK functionality, including the Parallel BLAS, linear systems using LU and Cholesky, least squares problems, and eigenvalue and singular value problems. In this respect, it will serve as a replacement for ScaLAPACK, which after two decades of operation, cannot adequately be retrofitted for modern accelerated architectures. SLATE uses modern techniques such as communication-avoiding algorithms, lookahead panels to overlap communication and computation, and task-based scheduling, along with a modern C++ framework. Here we present the design of SLATE and initial reports of several of its components.
Download PDF
Archive