Main Topics

BigDataBench is a comprehensive big data benchmark suite (published on HPCA’14). The source code of BigDataBench is available from http://prof.ict.ac.cn/BigDataBench This tutorial presents what is BigDataBench and how to use BigDataBench. Please feel free to download Handbook of BigDataBench [BigDataBench Handbook]

Location and Date

We will give a tutorial on BigDataBench at Micro 2014 in Cambridge, UK.

Dec. 13 (Sat) ,09:00 - 12:00 (Half Day)

ROOM:Games

Organizers and Presenters

Organizer: Jianfeng Zhan Chinese Academy of Sciences, and University of Chinese Academy of Sciences
Presenter: Jianfeng Zhan Chinese Academy of Sciences, and University of Chinese Academy of Sciences
Presenter: Zhen Jia Chinese Academy of Sciences, and University of Chinese Academy of Sciences
Presenter: Gang Lu Chinese Academy of Sciences, and University of Chinese Academy of Sciences

Abstract

BigDataBench is an open-source big data benchmark suite, publicly available from BigDataBench. After identifying diverse data models and representative big data workloads, BigDataBench proposes several benchmarks specifications to model five important application domains, including search engine, social networks, ecommerce, multimedia data analytics and bioinformatics. BigDataBench implements the same benchmarks specifications using variety of competitive techniques. The current version BigDataBench 3.1 includes 14 real-world data sets and the corresponding scalable big data generation tools, and 33 big data workloads. To allow flexible setting and replaying of mixed workloads, BigDataBench provides the multi-tenancy version; To save the benchmarking cost, BigDataBench reduces the full workloads to a subset according to workload characteristics from a specific perspective. It also provides both MARSSx86 and Simics simulator versions for architecture communities.

Schedule

09:00-09:20 What is BigDataBench [Slides]
09:00-09:40 BigDataBench methodology [Slides]
09:40-10:00 BigDataBench workloads and scalable data sets/how to use them [Slides-1 ,Slides-2 ]
10:00-10:30 — Coffee break —
10:30-10:50 How to generate large-scale data from small-scale real-world one [Slides ]
10:50-11:00 Multi-tenancy version of BigDataBench [Slides ]
11:00-11:30 Subsetting big data workloads from BigDataBench [Slides ]
11:30-12:00 How to use the simulator versions? [Slides ]

Biographies

Jianfeng Zhan
Jianfeng Zhan is a Professor of Computer Science and Engineering at Institute of Computing Technology, Chinese Academy of Sciences and University of Chinese Academy of Sciences. His research interests include computer architecture, operating systems, data management, parallel and distributed systems. He has published over 60 papers in major journals and international conferences related to these research areas, and filed 40 patents. From 2004 to 2010, he leaded the R&D efforts of innovative cluster and cloud systems software for the dawning-series super computers (which ranked top 2 and top 10 on the top 500 list in 2010 and 2004, respectively). Among them, GridView was transferred to Sugon, which is a premier supercomputing company in China, and becomes its popular software product. Currently, he is leading the research efforts for datacenter and big data software stacks, including BigDataBench---an open source big data benchmarking project, and RainForest--- an operating system for warehouse-scale computing. He received the second-class Chinese National Technology Promotion Prize in 2006, the Distinguished Achievement Award of the Chinese Academy of Sciences in 2005, IISWC Best paper award in 2013, and Huawei Contribution Prize in 2013, respectively. More details about Prof. Zhan are available at http://prof.ict.ac.cn/jfzhan.

Zhen Jia
Zhen Jia is a PhD candidate in Computer Science at the Institute of Computing Technology, Chinese Academy of Sciences and University of Chinese Academy of Sciences. His research focuses on parallel and distributed systems, benchmarks, and data center workload characterization. He received his BS in 2010 from Dalian University of Technology in China.

Gang Lu
Gang Lu received his Bachelor’s Degree in 2010 from Huazhong University of Science and Technology in China in computer science. He is currently working toward PhD degree in computer science at Institute of Computing Technology, Chinese Academy of Sciences. His current research interests include operating system and distributed and parallel systems.

Relate Links

BigDatabench http://prof.ict.ac.cn/BigDataBench/