Published October 15, 2016 | Version v1
Conference paper Open

Accelerating Data Center Applications with Reconfigurable DataFlow Engines

Authors/Creators

  • 1. School of Electronics, Electrical Engineering and Computer Science
  • 2. Institute of Electronics, Communications & Information Technology

Description

Energy-efficient, programmable accelerators are an important value proposition for pushing the limits on the computation capacity and density of future data centers. Integrating the accelerators in the software stacks of cloud-based data analytics frameworks is, however, an open issue. The goal of this work is to achieve seamless integration of accelerators in data analytics programming frameworks.

Files

accelerate_170127115257.pdf

Files (105.4 kB)

Name Size Download all
md5:0fe896aeaaceccdc42edebb3a117ac42
105.4 kB Preview Download

Additional details

Funding

European Commission
VINEYARD - Versatile Integrated Accelerator-based Heterogeneous Data Centres 687628