Program

9:00 Welcome

9:05  Shahar Kvatinsky, Technion [slides]
Real Processing-in-Memory  with Memristive Memory Processing Unit


Memory

9:55  Reactive NaN Repair for Applying Approximate Memory to Numerical Applications. Shinsuke Hamada, Soramichi Akiyama and Mitaro Namiki.

10:20 Smart Software Caches. Antonio Barbalace, Martin Decky and Javier Picorel.

10:40 Asymmetry-aware Page Placement for Contemporary NUMA Architectures. David Gureya, Rodrigo Rodrigues, Paolo Romano, Pramod Bhatotia, Vivien Quema, Joa Barreto.

11:05-11:35 Coffee break

Compilers and programming models


11.35  Local Operations Should Appear to Be Remote: Consistent Semantics Enable Transparent Distribution. John A. Thywissen, Arthur Michener Peters and Christopher Rossbach. 
12:00  PorcE: A Deparallelizing Compiler. Arthur Michener Peters, John A. Thywissen and Christopher Rossbach. 
12:25  Towards Fully Automatic Staged Computation. Mathias Gottschlag, Christian Schwarz, Marc Rittinghaus and Frank Bellosa.

12:50-14:00 Lunch

14:00 Gustavo Alonso, ETH [slides]
The impact of modern hardware on system design



Cloud acceleration

14:55  NFV Acceleration: The Role of the NIC. Haggai Eran, Dotan Levi, Liran Liss and Mark Silberstein.
15:15  CELDA: Cloud Edge Local Dataflow Architecture. Henrique Fingler and Christopher Rossbach. 

Accelerator programming models

15:40 Talk to My Neighbors Transport: Decentralized Data Transfer and Scheduling Among Accelerators. Amogh Akshintala, Vance Miller, Donald Porter and Christopher Rossbach. 

16:05-16:30 Coffee break

16:30 CPUs as Co-processors of GPUs: Running GPGPU Applications at the Full Speed with PullKernels. Yusuke Suzuki, Hiroshi Yamada, Shinpei Kato and Kenji Kono.

17:00  Derek Murray, Google [slides
Optimizing TensorFlow for Multi-core and Heterogeneous Architectures