Mobile friendly
| Time | Event |
|---|---|
| 9:00 - 9:10 | Welcome and Introduction Antonino Tumeo, Vito Giovanni Castellana, John Feo |
| 9:10 - 10:00 | Keynote 1 - Chair: Marco Minutoli (PNNL) Photonic Interconnects for Extreme Scale Computing Madeleine Glick (Columbia University) |
| 10:00 - 10:30 | Coffee Break |
| 10:30 - 11:45 | Session 1: Software optimizations for irregular applications - Chair: Roger Pearce (LLNL) Software prefetching for unstructured mesh applications Hadade, Jones, Wang, Di Mare A Fast and Simple Approach to Merge and Merge Sorting using Wide Vector Instructions Watkins, Green Impact of Traditional Sparse Optimizations on a Migratory Thread Architecture Rolinger, Krieger |
| 11:45 - 12:35 | Session 2: Graph Processing - Chair: Scott Beamer (UCSC) There are Trillions of Little Forks in the Road. Choose Wisely! -- Estimating the Cost and Likelihood of Success of Constrained Walks to Optimize a Graph Pruning Pipeline -- Tripoul, Halawa, Reza, Sanders, Pearce, Ripeanu Scale-Free Graph Processing on a NUMA Machine Aasawat, Reza, Ripeanu |
| 12:35 - 14:10 | Lunch Break |
| 14:10 - 15:00 | Keynote 2 - Chair: Vito Giovanni Castellana (PNNL) Versal: The new Xilinx Adaptive Compute Acceleration Platforms (ACAP) Kees Vissers (Xilinx) |
| 15:00 - 15:30 | Coffee Break |
| 15:30 - 16:45 | Session 3: Irregular Kernels on GPUs - Chair: Mahantesh Halappanavar (PNNL) Mix-and-Match: A Model-driven Runtime Optimisation Strategy for BFS on GPUs Verstraaten, Varbanescu, de Laat A Block-oriented, Parallel and Collective Approach to Sparse Indefinite Preconditioning on GPUs Thuerck, Naumov, Goesele, Garland High-Performance GPU Implementation of PageRank with Reduced Precision based on Mantissa Segmentation Grützmacher, Anzt, Scheidegger, Quintana-Ortí |
| 16:45 - 17:30 | Debate - Moderator: John Feo (PNNL) Proposition: "Data analytic workloads are a complex mix of heterogeneous data, algorithms, and computing models. Thus, reconfigurable computing architectures are an attractive alternative to today's state-of-the-art fixed heteogenous designs. Market forces expect real-time decision systems to generate sales and awareness in an ever increasing interconnected modern society. But, the high volume and high velocity of data streams particular to these problems makes reconfiguration impractical. Unable to support important commercial data analytic workloads, the market for reconfigurable systems is too small to be profitable." Panelists: Sanjukta Bhowmick (University of Omaha-Nebraska), Aydin Buluç (LBNL), Howie Huang (George Washington University), Martin Schulz (University of Munich), |
Authors interested in the topics of the workshop are also invited to submit papers to the thematic Special Issue on Innovations in Systems for Irregular Applications of ACM Transactions on Parallel Computing (TOPC).
We invite, in particular, authors of papers accepted to previous editions of the workshop, as well as authors of accepted papers to the upcoming edition, to submit exteded versions of their paper.
For more information on the TOPC Special Issue on Innovations in Systems for Irregular Applications, including details on submissions and important dates, please consult the following link: http://hpc.pnl.gov/TOPCSI/.
Due to the number of requests received, we have moved the submission deadline to November 30, 2018!