Loading...
詹剑锋教授 主要成就 个人自述

詹剑锋教授(下载简历)

欢迎光临我的个人主页。我是中国科学院计算技术研究所研究员(自2012年起),博士生导师(自2010年起),中国科学院大学岗位教授以及欧洲自然科学院外籍院士。我同时担任了先进计算机系统研究中心主任。

我的研究专注于创建评价学这一统一的学科,我将其称为"评价科学与工程",并创造了一个全新的英文术语——"Evaluatology"(我的博士生李鸿霄对这一术语的创造也有贡献)。在高性能计算和数据中心系统软件方面,我也取得了一些成就,并成功将35项专利转让给领先的企业。正如某上市公司的招股书所提及的,我作为主要发明人提出的《一种自主构建集群操作系统内核的方法及智能构建器》专利,连同其他两项专利和一项软件著作权(属于其他贡献者),已转让给该公司,占其总股本的34.44%。

此外,我还是一位创业者,分别创立或发起了国际测试委员会(International Open Benchmark Council)、Top 100科技评价活动、国际期刊《BenchCouncil Transactions on Benchmarks, Standards and Evaluations》以及多个国际会议。

除了学术研究和创业事业,我也是一个热爱生活的诗人。在2024年,我将与女儿共同出版诗集《在旷野同行》。

我先后获得国家科技进步二等奖(2005 年)、 中 国 科 学 院 杰 出 成 就 奖 (2006 年 ) , IEEE International Symposium on Workload Characterization (IISWC) 2013 最 佳 论 文 奖 、 IEEE Transactions on Parallel and Distributed Systems (TPDS) 2020 年优秀编辑奖等荣誉。

我的联系方式如下:
地址:北京市海淀区中关村科学院南路6号中科院计算所1038室。
电子邮件:jianfengzhan.benchcouncil@gmail.com

主要成就

Evaluation is a crucial aspect of human existence and plays a vital role in various fields[1]. However, it is often approached in an empirical and ad-hoc manner, lacking consensus on universal concepts, terminologies, theories, and methodologies. I formally introduce the discipline of evaluatology, which encompasses the science and engineering of evaluation. Derived from the essence of evaluation, I propose five axioms focusing on key aspects of evaluation outcomes as the foundational evaluation theory. These axioms serve as the bedrock upon which a universal framework for evaluation is proposed, encompassing concepts, terminologies, theories, and methodologies that can be applied across various disciplines. Building upon the science of evaluation, I propose a formal definition of a benchmark as a simplified and sampled evaluation condition that guarantees different levels of equivalency. This concept serves as the cornerstone for a universal benchmark-based engineering approach to evaluation across various disciplines, which I refer to as benchmarkology.

Under my leadership, the big data and AI benchmarking projects, BigDataBench [2,3] and AIBench [4,5], have had significant impacts. BigDataBench is widely adopted in academia and industry worldwide, with the article published at HPCA'14 receiving 726 citations.

I am the founding chair of the International Open Benchmark Council (BenchCouncil) (https://www.benchcouncil.org ). Renowned scholars such as Matsuoka, Satoshi, Andrew A. Chien, Jeffrey S. Vetter, and John Shalf have ranked BenchCouncil's contributions alongside SPEC, founded in 1988, and before MLCommons, which was founded by Google and other industry and academic leaders such as Stanford and Facebook [2]. In their citation, they state, "Other commonly used benchmarking sets have been curated by organizations such as the Standard Performance Evaluation Corporation (SPEC), the International Open Benchmark Council (BenchCouncil), or MLCommons—with proxies for CPUs, distributed systems, accelerators, artificial intelligence—or have been developed by companies such as Intel (HiBench), Baidu (DeepBench), and Google (PerfKit)."

[1] Jianfeng Zhan, Lei Wang, Wanling Gao et al. Evaluatology: The Science and Engineering of Evaluation. arXiv preprint arXiv:2404.00021, 2024.
[2] Wang, Lei, Jianfeng Zhan, et al. "Bigdatabench: A big data benchmark suite from internet services." In 2014 IEEE 20th international symposium on high-performance computer architecture (HPCA), pp. 488-499. IEEE, 2014. Google Citations: 711. The first author, Lei Wang, is my Ph. D student.
[3] Jia, Zhen, Lei Wang, Jianfeng Zhan, et al. "Characterizing data analysis workloads in data centers." IISWC 2013 Best Paper Award.
[4]Hao, Tianshu, …Jianfeng Zhan. "Edge AIBench: Towards comprehensive end-to-end edge computing benchmarking." In Benchmarking, Measuring, and Optimizing: First BenchCouncil International Symposium, Bench 2018, Seattle, WA, USA, December 10-13, 2018. Google Citations: 81. The first author, Tianshu Hao, is my Ph. D student.
[5] Tang, Fei, Wanling Gao, Jianfeng Zhan, et al. "AIBench training: Balanced industry-standard AI training benchmarking." In 2021 IEEE International Symposium on Performance Analysis of Systems and Software (ISPASS), pp. 24-35. IEEE, 2021. Google Citations: 21. The first author, Fei Tang, is my Ph. D student.
[6] Matsuoka, Satoshi, Jens Domke, Mohamed Wahib, Aleksandr Drozd, Andrew A. Chien, Raymond Bair, Jeffrey S. Vetter, and John Shalf. "Preparing for the Future—Rethinking Proxy Applications." Computing in Science & Engineering 24, no. 2 (2022): 85-90.

I spearheaded the research and development of the cluster operating system [7] for the Dawning 4000A, which achieved the 10th position on the Top 500 supercomputer list in June 2004 (source: https://www.top500.org/system/173564/). As mentioned in Sugon's initial public offering prospectus, my team's patent titled "A Method for Autonomously Constructing Cluster Operating System Kernels and an Intelligent Constructor," along with two other patents and one software IP, were transferred to Sugon, accounting for 34.44% of its total equity share [8]. Sugon is China's leading supercomputer company. In recognition of my contributions, I received the Outstanding Science and Technology Achievement Prize from the Chinese Academy of Sciences in 2005 and the Second Prize of The State Scientific and Technological Progress Award from the Central People's Government of the People's Republic of China in 2006.

Furthermore, I played a pivotal role in the design and implementation of cutting-edge cloud computing systems, namely PhoenixCloud [9,11] and DawningCloud [10], on the Dawning 5000 and 6000 Supercomputers. These innovative solutions revolutionized the support for high-performance computing and the consolidation of heterogeneous workloads on cloud systems. Additionally, I transferred 34 OS patents to Huawei, a world-class communication and IT company.

[7] Jianfeng Zhan, and Ninghui Sun. "Fire Phoenix cluster operating system kernel and its evaluation." In 2005 IEEE International Conference on Cluster Computing, pp. 1-9. IEEE, 2005.
[8] Sugon' initial public offering prospectus. Page 1-1-62. https://pdf.dfcfw.com/pdf/H2_AN201410230007390902_1.pdf .Accessed at September 10, 2023
[9] Jianfeng Zhan, Lei Wang, Bibo Tu, Yong Li, Peng Wang, Wei Zhou, and Dan Meng. "Phoenix cloud: Consolidating heterogeneous workloads of large organizations on cloud computing platforms." The first workshop on Cloud Computing and its Applications (CCA'08), 2008. https://www.alcf.anl.gov/events/cca-cloud-computing-and-its-applications
[10] Wang, Lei, Jianfeng Zhan, Weisong Shi, and Yi Liang. "In cloud, can scientific communities benefit from the economies of scale?" IEEE Transactions on Parallel and Distributed Systems 23, no. 2 (2011): 296-303. Google Citations: 214. The first author Lei Wang is my Ph. D student at the publication.
[11] Jianfeng Zhan, et al. "Cost-aware cooperative resource provisioning for heterogeneous workloads in data centers." IEEE Transactions on Computers 62, no. 11 (2012): 2155-2168. Google Citations: 69.

个人自述(部分内容引自我和女儿合著的诗集《在旷野同行》)

“我的祖先从鱼米之乡江西饶州府鄱阳县一路迁徙,来到安徽和湖北接壤的一个偏僻的山村:詹家新屋。她在铁林寨的山脚下,门前有条没有名字的河。我去过鄱阳县附近一带,那儿的老奶奶和我的奶奶说着相同的语言,我都能听懂。”引自《故乡的河》,我1976年正月24日(阴历)出生于此。

“我在河口初中度过了快乐自由的时光。现在的孩子说不能输在起跑线上,我那时候连起跑线是什么都不知道。我读初一时,父亲早逝(87年3月3日,阴历),母慈子不孝,我浑浑噩噩地度过了初中时光,几乎不读书。参加中考,作文题目是“记一个雷锋似的人物”,我故意偏题,硬生生地杜撰了一个早恋的班长。我中考成绩比分数线少了几分,居然还考了全校第三。母亲为了父亲的遗嘱,倾尽家里所有的财产,花了一千元帮我买了弥陀高中(一个非常普通的中学)的学籍,由此遭到了不少人“买功名”的嘲笑。”引自《少年》。

“从十三到十六岁,我慌里慌张地在弥陀高中开始了青春的序章。那时候,饥饿是不少同学最大的困扰。”引自《青春》。

我十六岁考入西南交通大学,就读铁道与道路工程(1992-1996)和固体力学专业(1996-1999),先后获得工学学士与工学硕士学位。1999年-2002年就读于中国科学院软件技术研究所,获得计算机软件与理论专业博士学位。严格上来说,我没有上过一门计算机课程。不过三生有幸,最后还在计算机研究圈扎下了根,研究方向覆盖体系结构、操作系统和算法。不过,从92年到99年,我在西南交大除了接受铁道工程师和固体力学的一些理论训练外,我花了不少时间倒腾VAX小型计算机和近20多台PC和工作站组成的分布式系统。其实我没有受到詹天佑影响,一切只是命运的随机安排,我当年的大学志愿填的是喂养熊猫和信息工程,铁道与道路工程是阴差阳错的专业调剂的结果。这段经历确实培养了我对计算机系统的兴趣和想象力。而在读大学之前和之中,我热衷于在大自然闲逛和冥想。我对自己的推测是对的,读一个工科确实能在无聊和颓废时拯救我。感谢西南交大,近20台计算机在当时是非常奢侈的资源,以至于我1999年来中科院软件所读博时,感觉回到了原始社会,非常不适应,写博士论文时,几乎一半靠想象的翅膀飞翔。02年来中科院计算所工作,听了徐志伟老师的报告,才知道我在读研期间独立做的工作正是热门的元计算或者网格。我当年读硕士,一个人折腾分布式系统来做有限元计算,很新潮也很艰难(当年的网管,一到放假,就把网络全拆了,故意给我们设置各种障碍)。读研究生期间,我独自写了一篇英文文章寄给Computer & Structure(可是SCI哦)评审。Reviewer说,文章立意很好,但英文句子太糟糕(生不逢时,没有遇到Grammarly的时代)。我在读硕士期间,在系统仿真学报和计算机仿真上发表了两篇论文。以我现在的视角来看,当年的硕士论文工作还是有些勇气的,题目并不容易,关键是只有我一个人工作(导师的专业是固体力学),而且当时正遭受着厄运。写下这些文字时,我非常怀念英年早逝的导师孔祥安教授,愿他在天堂安好。