CSC801: Seminar in Computer Science (System Group)

When & Where

EB2 3211

Friday 11:00a - 12:00p

Course Description

The primary objective of the system seminar is to keep us abreast of interesting developments in system research with relatively little effort or discuss your innovative ideas with the rest of the group.

If the topic of this week is to present a "published" or an "accepted" paper, everyone reads the paper and needs to discuss with the whole class. You may also discuss your in-progress ideas with us. Presentations are low-key:

  • Keep it short (aim for 20 minutes -- most conference presentations are only this long) to allow more time for discussion.
  • 2-3 slides: presenting the why of the paper / your work -- the problem this paper/work is attacking.
  • 2-3 slides: presenting the what of the paper / your work -- the mechanisms/solutions that this paper/work is proposing.
  • 2-3 slides: presenting the how of the paper / your work -- the implementation details and most important results.
  • 1 slide: strengths. (If it's a published paper')
  • 1 slide: weaknesses.(If it's a published paper')
  • 1-2 slides: though-provoking questions to get discussion started.
  • It's useful if you cut and paste the graphs and important tables onto slides so we can put them up and discuss them as needed.
Questions? or would like to present? email hungwei_tseng at

Schedule and Slides

DateTopicPresentor Notes
2017/08/25 Kick off Hung-Wei Tseng & Marvin Damschen Marvin Damschen is from Karlsruhe Institute of Technology
2017/09/01 Summarizer: Trading Communication with Computing Near Storage Te I
2017/09/08 Cancelled
2017/09/22 PCERE: Fine-grained Parallel Benchmark Decomposition for Scalability Prediction. Mihail Popov, Chadi Akel, Florent Conti, William Jalby, and Pablo de Oliveira Castro. In Proceedings of the 29th IEEE International Parallel and Distributed Processing Symposium IPDPS 2015. IEEE, 2015 Tao Wang
2017/09/29 Enabling Inter-request Analysis for Request-based Applications Xiao Yu
2017/10/13 Code generation for efficient query processing in managed runtimes Stefan O'neil
2017/10/20 LSTM Optimizations on GPUs with Software and Hardware Scheduling Jin Wang Jin is from NVIDIA
2017/10/27 Scout: Guided Walk for Exploring Architecture Configuration in the Cloud Chin-Jung Hsu
2017/11/03 Hui Guan
2017/11/10 A League of Extraordinary Machines Jack Davidson (UVA) Jack Davidson is from University of Virginia
2017/11/17 Energy Efficient Architecture for Graph Analytics Accelerators Lin Ning and Ozcan Ozturk Ozcan Ozturk is from Bilkent Univ. Ankara, Turkey
2017/11/27 Yiran Chen (Duke) Not the usual schedule