Jian Fang

PhD student at the Computer Engineering Lab of the Faculty of Engineering, Mathematics and Computer Science (EEMCS/EWI), Delft University of Technology.

I am currently working on using FPGAs to accelerate databases. This project is under the supervision of Peter Hofstee.


Using FPGAs to Accelerate Databases Operations with Huge Acceleration Bandwidth

Project Overview

OpenCAPI enables a low latency, high bandwidth interconnect by using 25Gb/s differential signaling. This results in a uni-directional bandwidth of  25GB/s for 8 lanes. In comparison, the widely adopted PCIe Gen 3 interconnect attains a bandwidth of roughly 8GB/s for a similar number of lanes. Aggregate OpenCAPI bandwidth can rival or exceed the bandwidth of DDR memory, making OpenCAPI-attached accelerators candidates for bandwidth-limited applications.

Our work focus on designing a heterogeneous architecture for databases with high memory bandwidth connected FPGAs. Based on this architecture, four high-bandwidth streaming accelerators for database queries are studied: decompress-filter, hash-join, multi-way join, and merge-sort. Each has different buffering requirements, which are challenging at this speed. Requirements vary over having to hide latency versus the number of read ports.

  1. Jian Fang (PhD Student)
  2. Jinho Lee (PostDoc)
  3. Jan Hidders (Prof.)
  4. H. Peter Hofstee (Prof.)
  5. Jianyu Chen (Master Internship)
  6. Yvo T.B. Mulder (Master, Graduated)
  7. Xianwei Zeng (Master, Graduated)
  8. Kangli Huang (Master, Graduated)
  9. Yang Qiao (Master, Graduated)

Research Interest

  • Heterogeneous Computer Architecture
  • High Performance Computing
  • Reconfigurable Computing
  • Database Systems
  • Graph Processing

Master Project

It is my pleasure to work together with the following master student, supervising or collaborating with them on their theses:

  • 1. Kangli Huang (Multi-way Hash Join Based on FPGAs, Graduated in Jan 2018)
  • 2. Xianwei Zeng (FPGA-Based High Throughput Merge Sorter, Graduated in Jan 2018)
  • 3. Yang Qiao (An FPGA-based Snappy Decompressor-Filter, Graduated in Jan 2018)
  • 4. Yvo T.B. Mulder (Feeding High-Bandwidth Streaming-Based FPGA Accelerators, Graduated in Jan 2018)

I am now supervising the following Master internship program together with Peter Hofstee and Zaid Al-Ars:

  • 1. Jianyu Chen ([sub-project: FPGA-Based Snappy Decompressor] from Project: Parqeut-to-Arrow converter on FPGAs)

You are welcome to join our team for your master thesis/honor/intern project. The topic mainly focuses on A) Big data systems or database systems performance analytic and optimization, B) database, machine learning, data mining acceleration on FPGAs. You should have some basic knowledge on computer architecture, system performance analytic, and basic mathematics. For topic A) you'd better to know some software program skill of C/C++. For topic B) you'd better to know some hardware design skill of VHDL/Verilog.

The project now use the Fabric server which contains 9 IBM Power8 S824L nodes server in TACC . The design targets on the IBM Power9 server with OpenCAPI, and you will have a chance to touch the latest systems in the world.

Invited Talks

Progress with Power Systems and CAPI  --  Exploiting Accelerator Diversity for Cognitive Workloads - Workshop at MICRO'50 (Cognitive Workloads@MICRO'50), October 2017, Boston, USA. (Slides)

Leveraging the bandwidth of OpenCAPI with reconfigurable logic -- Academic Discussion Group Workshop 2017 - ADG 2017, November 2017, Denver, USA. (Slides)

Publications Jian Fang

Please note: This page contains links to PostScript files of articles that may be covered by copyright. You may browse the articles at your convenience. (In the same spirit as you may read a journal or a proceeding article in a public library). Retrieving, copying, or distributing these files, however, may violate the copyright protection law. We recommend that the user abides international law in accessing this article list.


  1. J. Fang, Y.T.B. Mulder, K Huang, Y Qiao, X Zeng, P. Hofstee, J Lee, J Hidders, Adopting OpenCAPI for High Bandwidth Database Accelerators1649_adopting_opencapi_for_high_bandwidth_database_accelerators.pdf (November 2017), 3rd International Workshop on Heterogeneous High-performance Reconfigurable Computing (H2RC 2017), 17 November 2017, Denver, USA [Conference Paper]
  2. J. Fang, J Lee, P. Hofstee, J Hidders, Analyzing In-Memory Hash Joins: Granularity Matters1639_analyzing_inmemory_hash_joins_granularity_matters.pdf (September 2017), 8th International Workshop on Accelerating Analytics and Data Management Systems Using Modern Processor and Storage Architectures (ADMS 2017), 1 September 2017, Munich, Germany [Conference Paper]
  3. 2016

  4. J. Fang, J Hidders, K.L.M. Bertels, J Lee, P. Hofstee, A Locality-Aware Hash-Join Algorithm1557_a_localityaware_hashjoin_algorithm.pdf (July 2016), 12th International Summer School on Advanced Computer Architecture and Compilation for High-Performance and Embedded Systems (ACACES 2016), 10-16 July 2016, Fiuggi, Italy [Conference Proceedings]