NNSim: Fast performance estimation based on sampled simulation of gpgpu kernels for neural networks

Jintaek Kang, Kwanghyun Chung, Youngmin Yi, Soonhoi Ha

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Scopus citations

Abstract

Existent GPU simulators are too slow to use for neural networks implemented in GPUs. For fast performance estimation, we propose a novel hybrid method of analytical performance modeling and sampled simulation of GPUs. By taking full advantage of repeated computation of neural networks, three sampling techniques are devised: Inter-Kernel sampling, Intra-Kernel sampling, and Streaming Multiprocessor sampling. The key technique is to estimate the average IPC through sampled simulation, considering the effect of the warp scheduler and memory access contention. Compared with GPGPU-Sim, the proposed technique reduces the simulation time by up to 450 times with less than 5.0% of accuracy loss.

Original languageEnglish
Title of host publicationProceedings of the 55th Annual Design Automation Conference, DAC 2018
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Print)9781450357005
DOIs
StatePublished - 24 Jun 2018
Event55th Annual Design Automation Conference, DAC 2018 - San Francisco, United States
Duration: 24 Jun 201829 Jun 2018

Publication series

NameProceedings - Design Automation Conference
VolumePart F137710
ISSN (Print)0738-100X

Conference

Conference55th Annual Design Automation Conference, DAC 2018
Country/TerritoryUnited States
CitySan Francisco
Period24/06/1829/06/18

Keywords

  • Analytical Model
  • GPU Simulator
  • Sampled Simulation

Fingerprint

Dive into the research topics of 'NNSim: Fast performance estimation based on sampled simulation of gpgpu kernels for neural networks'. Together they form a unique fingerprint.

Cite this