IXPUG Birds of a Feather (BOF) session at SC16
Wednesday, Nov. 16, 2016 from 12:15-1:15 p.m, Room 355-F
Optimizing Performance on Many-Core Processors: Unleashing the Power of the Intel® Xeon Phi and Beyond
Agenda: (PDFs linked for each presentation)
|Richard Gerber, IXPUG President|
Xeon Phi Refresher
|John Pennycook, Intel|
Performance evaluation of Weather Research and Forecast (WRF) on Intel Knights Landing Processor (KNL)
George S. Markomanolis
Saber Feki, KAUST
|Looking at Clustering and MCDRAM options on large applications||John Levesque, Cray|
|3D FEM for Solid Mechanics on Knights Landing and Omni Path Architecture||Toshihiro Hanawa, University of Tokyo|
|Automatic deep performance understanding of deep learning codes in multi- and many-core processors||Tareq Malas, Berkeley Lab|
Optimizing Molecular Dynamics and Stencil mini-applications for Intel’s Knights Landing
|Kavitha Chandrasekar, UIUC|
The Intel Xeon Phi Users Group (IXPUG) is soliciting proposals for 5-minute “lightning talks” on experiences configuring, using, and benchmarking Knight’s Landing clustering modes and high bandwidth memory (MCDRAM), to be presented at the IXPUG Birds of a Feather at SC16.
The Knight’s Landing architecture and supporting software gives users a number of choices to help their applications achieve enhanced performance. The IXPUG Birds of a Feather at SC16 will focus on user experiences on two related topics: usage of the KNL clustering modes (e.g. quadrant/hemisphere, SNC-4/SNC-2) and MCDRAM usage (e.g. flat vs. cache mode).
We are especially interested in real-world experiences with full applications.
- Oct. 26, 2016: Abstracts/proposals due
- Oct. 31, 2016: Notification of selected talks
- Nov. 7, 2016: Presentation drafts due
- Nov. 14, 2016: Final presentations due
- Nov. 16, 2016; SC16 Birds of a Feather