Opening remarks

8:30 - 8:45

Keynote

8:45 - 9:45

What first principles of algorithms and architectures says about heterogeneity

Prof. Richard Vuduc, Georgia Institute of Technology Slides

Abstract:

In this talk, we will "pop up a level" and ask whether first principles of algorithm and architecture design can tell us anything about heterogeneity. In particular, I'll posit "strawman" cost models that connect properties of an algorithm with physical properties of machines, like energy, power, and area. I will then explain what these models imply. For instance, I'll use them to evaluate the time- and energy-efficiency of heterogeneous designs as well as for predicting the benefits of other forms of heterogeneity, such as "communication" heterogeneity. And although I'll show a bunch of formulas, calculate things, and plot them, I stress that this talk is about ideas, rather than a well-developed set of results. As such, your questions, healthy skepticism, (constructive!) feedback, and offers of collaboration may be even more welcome than usual! :-) (Read...)

Break 9:45 - 10:15

Session 1: Programing Model and Performance Optimizations

10:15 - 11:55

Session Chair: Toshio ENDO, Tokyo Institute of Technology

• Synchronization and Ordering Semantics in Hybrid MPI+GPU Programming
  Ashwin Aji, Pavan Balaji, James Dinan, Wuchun Feng and Rajeev Thakur. Slides
• Tightly Coupled Accelerators Architecture for Minimizing Communication Latency among Accelerators
  Toshihiro Hanawa, Yuetsu Kodama, Taisuke Boku and Mitsuhisa Sato. Slides
• Analyzing Optimization Techniques for Power Efficiency on Heterogeneous Platforms
  Yash Ukidave and David Kaeli. Slides
• Composing multiple Starpu applications over heterogeneous machines: a supervised approach
  Andra-Ecaterina Hugo, Abdou Guermouche, Pierre-Andre Wacrenier and Raymond Namyst. Slides

Lunch 11:55 am - 1:30 pm

Session 2: Accelerated Applications

1:30 - 3:10 pm

Session Chair: Yong Chen, Texas Tech University

• Fast, Scalable Parallel Comparison Sort on Hybrid Multicore Architectures
  Dip Sankar Banerjee, Parikshit Sakurikar and Kishore Kothapalli. Slides
• Tridiagonalization of a symmetric dense matrix on a GPU cluster
  Ichitaro Yamazaki. Slides
• A Multi-level Optimization Method for Stencil Computation on the Domain that is Bigger than Memory Capacity of GPU
  Guanghao Jin, Toshio Endo and Satoshi Matsuoka. Slides
• Accelerating the 3D Elastic Wave Forward Model on GPU and MIC
  Yang You, Haohuan Fu, Guangwen Yang and Xiaomeng Huang. Slides

Break 3:10 - 3:40 pm

Session 3: Emerging Hybrid Systems

3:40 - 5:20 pm

Session Chair: David Kaeli, Northeastern University

• Improving GPU Performance Prediction with Data Transfer Modeling
  Michael Boyer, Jiayuan Meng and Kalyan Kumaran. Slides
• Use of SIMD Vector Operations to Accelerate Application Code Performance on Low-Powered ARM and Intel Platforms
  Gaurav Mitra, Beau Johnston, Alistair P. Rendell, Eric McCreath and Jun Zhou. Slides
• Using MIC to accelerate a typical data-intensive application: the Breadth-first Search
  Gao Tao, Lu Yutong and Suo Guang. Slides
• Dynamic Load Balancing of the Adaptive Fast Multipole Method in Heterogeneous Systems
  Robert Overman, Jan Prins, Laura Miller and Michael Minion. Slides