Write a Blog >>
LCTES 2017
Wed 21 - Thu 22 June 2017 Barcelona, Spain
co-located with PLDI 2017

Convolutional neural networks (CNNs) are widely adopted in artificial intelligent systems. In contrast to conventional computingcentric applications, the computational and memory resources of CNN applications are mixed together in the network weights. This incurs a significant amount of data movement, especially for highdimensional convolutions. Although recent embedded 3D-stacked Processing-in-Memory (PIM) architecture alleviates this memory bottleneck to provide fast near-data processing, memory is still a limiting factor of the entire system. An unsolved key challenge is how to efficiently allocate convolutions to 3D-stacked PIMto combine the advantages of both neural and computational processing.

This paper presents Memolution, a compiler-based memoryefficient data allocation strategy for convolutional neural networks on PIM architecture. Memolution offers thread-level parallelism that can fully exploit the computational power of PIMarchitecture. The objective is to capture the characteristics of neural network applications and present a hardware-independent design to transparently allocate CNN applications onto the underlining hardware resources provided by PIM. We demonstrate the viability of the proposed technique using a variety of realistic convolutional neural network applications. Our extensive evaluations show that, Memolution significantly improves performance and the cache utilization compared to the baseline scheme.

Thu 22 Jun

10:30 - 12:10: LCTES 2017 - Session 3: Non-Volatile Memory/Processor and RTOS at Vertex WS208
Chair(s): Hanjun KimPOSTECH
LCTES-2017-papers10:30 - 10:55
Yi WangShenzhen University, Mingxu ZhangShenzhen University, Jing YangHarbin Institute of Technology
LCTES-2017-papers10:55 - 11:20
Xianzhong DingShandong University, Zhiyong ZhangShandong University, Zhiping JiaShandong University, Lei JuShandong University, Mengying ZhaoShandong University, Huawei HuangThe University of Aizu
LCTES-2017-papers11:20 - 11:45
Chen Pan, Mimi XieOklahoma State University, Yongpan LiuTsinghua University, Yanzhi WangSyracuse University, Jason XueCity University of Hong Kong, China, Yiran ChenUniversity of Pittsburgh, Jingtong HuOklahoma State University
LCTES-2017-papers11:45 - 12:10
Christian Dietrich, Daniel LohmannFriedrich-Alexander-Universität, Germany