architecting technology enabled phase change memory systems
speaker: dr. tao li, associate professor
department of electrical and computer engineering
university of florida
主讲人:李涛博士, 副教授
电子/电机和计算机工程系,佛罗里达大学 (美国)
报告时间:10月13日下午1:30
报告地点:玉泉校区 曹光彪大楼218室
报告地点:玉泉校区 曹光彪大楼218室
abstract
as the number of cores per cpu chip continues to increase and computer workloads become more memory intensive, providing highly efficient memory architecture is crucial for achieving the desirable performance scalability of future multi-core based computer systems. dynamic random access memory (dram), which has been used as main memory for decades, is facing performance, power and scalability walls in the era of multi-core and technology scaling. this trend will prevent the integration of ultra-high density dram devices into future multi-core on-chip memory hierarchy.
the cutting-edge phase change memory (pcm) is attracting increasing attention as a promising candidate for next generation memories. compared with dram, pcm has very low standby power, which could contribute to significant power reduction in memory hierarchy. the unique, heat-driven programming mechanism of pcm suggests that it has potential to tolerate higher chip temperature, a desirable feature for on-chip processor and memory integration using die stacking technologies. as technology scales, the volume of pcm’s phase change material shrinks as well, providing a truly scalable solution. in addition, phase change material can further be programmed into multiple resistance levels between the high and low, which can be used to effectively store more than 1-bit of data within a single cell.
nevertheless, there are many challenges need to be addressed for unleashing the above potentials. in this talk, i will discuss holistic and synergetic approaches that involve novel circuit, microarchitecture and operating system supports with the objective to fully exploit the potential of emerging phase change memory to significantly improve multi-core memory architecture performance, power-efficiency and reliability. this will enable multi-core processors and systems to stay on track of their scaling roadmap, and hence benefit numerous high-impact and emerging applications.
biography (http://www.taoli.ece.ufl.edu)
dr. tao li is currently an associate professor in the department of electrical and computer engineering at the university of florida. he received a ph.d. in computer engineering from the university of texas at austin. his research interests include computer architecture, energy-efficient and dependable microprocessors, memory systems and networks-on-chip, the impacts of emerging technologies and applications on hardware and operating/run-time systems, and evaluation of computer systems. dr. tao li received 2009 national science foundation faculty early career award, 2008, 2007, 2006 ibm faculty awards, 2008 microsoft research safe and scalable multi-core computing award and 2006 microsoft research trustworthy computing curriculum award.
主讲人简介(http://www.taoli.ece.ufl.edu)
李涛博士,2004年德州奥斯汀分校获得计算机工程博士学位,目前为佛罗里达大学电子与计算机工程系副教授。李涛博士于2009年获美国国家科学基金会杰出青年教授奖,2008年,2007年,2006年均获ibm学院奖,2008年获得微软研究院安全及可扩展多核计算机奖,2006年还获得微软研究院可信计算课程研究奖。研究方向包括计算机系统结构,高效可靠微处理器及存储系统,新型技术及应用对硬件和操作系统/系统软件和编译的影响,以及计算的性能、功耗、可靠性及安全性问题等。目前研究项目包括多核系统的系统资源智能化管理、纳米级系统结构设计中的可靠性优化、gpgpu体系结构、高性能低功耗服务器的数据中心体系结构设计、新型技术驱动的架构设计等,研究项目受到美国国家自然科学基金、美国航天局、半导体技术研究中心、ibm及微软等支持。在如micro, hpca, dsn, pact,alplos, sigmetrics 等计算机体系结构类顶级国际会议和ieee/acm transactions 杂志上发表学术论文多篇。