MultiGPU Systems / High-Performance Linear Algebra Software / AMD / NVIDIA / GPU / CPU / INTEL / Distributed-Memory Multicore Systems / /
Currency
AMD / /
Facility
O. Villa / Vienna Computing Library / Microelectronics TU Wien Karl Rupp MCS Division Argonne National Laboratory Siegfried Selberherr Institute / Microelectronics TU Wien Chin-Teng Lin Institute of Electrical / Convenient Linear Algebra Philippe Tillet Institute / library ATLAS / ATLAS Library / /
IndustryTerm
straight-forward brute-force search / dot products / heterogeneous devices / matrix-matrix products / target device / work / target hardware / memory bank / large search space / clever manual memory management / matrix-vector products / signal processing / signal processing algorithms / search space / /
OperatingSystem
Fermi / /
Organization
Performance-Portable / Scalable / and Convenient Linear Algebra Philippe Tillet Institute / Microelectronics TU Wien Karl Rupp MCS Division Argonne National Laboratory Siegfried Selberherr Institute / Microelectronics TU Wien Chin-Teng Lin Institute of Electrical / U.S. Securities and Exchange Commission / National Chiao Tung University / /
Person
Markus Puschel / Jeremy Johnson / Manuela Veloso / Karl Rupp / Robert W. Johnson / David Padua / Jianxin Xiong / Morgan Kaufmann / /
Towards Performance-Portable, Scalable, and Convenient Linear Algebra Philippe Tillet Institute for Microelectronics TU Wien