Accelerating Data Center Applications with Reconfigurable DataFlow Engines
Published in Heterogeneous High-Performance Reconfigurable Computing (H2RC) 2016 Workshop, 2016
This paper investigates the integration of reconfigurable DataFlow Engines (DFEs) into data center applications, focusing on enhancing computational efficiency and energy performance. The authors address the challenges of integrating hardware accelerators into existing data analytics frameworks, aiming to bridge the gap between high-performance computing resources and cloud-based data processing tasks.
Key contributions include:
Integration of DFEs into Data Analytics Frameworks: The paper discusses methods for incorporating DFEs into traditional data analytics frameworks, enabling seamless acceleration of computational tasks.
Performance Evaluation: The authors present performance evaluations demonstrating the effectiveness of DFEs in improving computational efficiency and energy performance in data center applications.
Challenges and Solutions: The paper identifies key challenges in integrating DFEs into existing systems and proposes solutions to overcome these obstacles, facilitating the adoption of hardware acceleration in data centers.
Recommended citation: Barbhuiya, S., Wu, Y., Murphy, K., Vandierendonck, H., Karakonstantis, G., & Nikolopoulos, D. S. (2016). Accelerating Data Center Applications with Reconfigurable DataFlow Engines. In *Proceedings of the Heterogeneous High-Performance Reconfigurable Computing (H2RC) 2016 Workshop*. https://h2rc.cse.sc.edu/2016/papers/paper_9.pdf
Download Paper | Download Slides | Download Bibtex