This shows you the differences between two versions of the page.

Link to this comparison view

seminar_10_18_17 [2017/10/13 11:28]
ermao created
seminar_10_18_17 [2017/10/13 11:57] (current)
Line 1: Line 1:
-====== ​Solving the DRAM Scaling Challenge ​======+====== ​Human Mobility Sensing, Analytics, and Services for Societal Good ======
 Wednesday October 18, 2017\\ Wednesday October 18, 2017\\
Line 5: Line 5:
 Time: 4:​30PM-5:​30PM\\ Time: 4:​30PM-5:​30PM\\
 **[[https://​www.ece.cmu.edu/​directory/​department/​faculty/​S/​John_Shen_159.html|John Paul Shen (Carnegie Mellon University)]]**\\ **[[https://​www.ece.cmu.edu/​directory/​department/​faculty/​S/​John_Shen_159.html|John Paul Shen (Carnegie Mellon University)]]**\\
 =====Abstract===== =====Abstract=====
-Technology scaling of DRAM cells has enabled higher capacity memory for the last few decades. UnfortunatelyDRAM cells become vulnerable ​to failure as they scale down to a smaller sizeEnabling high-performanceenergy-efficientscalable memory systems without sacrificing the reliability ​is a major research ​challenge. Khan's work focuses ​on designing a scalable memory system by rethinking ​the traditional assumptions in abstraction and separation ​of responsibilities across system layers+Unprecedented in human historywe now have the necessary tools to pursue ambitious global-scale experimental research on human mobilityResearchers from the fields of computerinformationdata, behavior, and social sciences, finally have our “Large Hadron Collider” to sense, curate, and analyze an incredible amount of real-world and real-time human mobility data; this is enabled by ubiquitous wireless connectivity,​ billions of mobile devices, and global ride sharing/​hailing services. This talk presents ​vision for such large-scale ​research on human mobility with the aim of using big data sensing and analytics for developing services for the mobile population and societal benefits.
-In this talk, Khan will discuss three fundamental ways to enable DRAM scaling. First, we can enable scaling by letting the manufacturers build smaller cells without providing any strict reliability guarantee. ​ Khan envision manufacturers shipping DRAMs without fully ensuring correct operation, and the system being responsible for detecting and mitigating DRAM failures while operating in the field. However, designing such a system is difficult due to intermittent DRAM failures. In this talk, Khan will discuss a system design, capable of providing reliability guarantees even in the presence of intermittent failures. Second, ​ tolerating failures in the application can improve DRAM scalability. The fundamental challenge of such a system is how to assure, verify, and quantify the quality of the results. Khan envision a system that limits the impact of memory failures such that it is possible to statically determine the worst-case results from the maximum possible error in the input. Third, we can enable high-capacity memory leveraging the emerging non-volatile memory technologies that are predicted to be more scalable. Khan will present her vision to redefine the hardware and operating system interface to unify memory and storage system with non-volatile memory and discuss the opportunities and challenges of such a system. ​ 
 =====Bio===== =====Bio=====
-Samira Khan is an Assistant Professor at the University of Virginia (UVa). Prior to joining ​UVashe was a Post Doctoral Researcher ​at Carnegie Mellon University, funded by Intel Labs. Her research ​focuses ​on improving the performance,​ efficiency, ​and reliability of the memory systemShe received her PhD from the University of Texas at San AntonioDuring her graduate studies, she worked at Intel, AMD, and EPFL+After spending 15 years in the industry, John returned to CMU in 2015. Prior to that he was the founding director of the Nokia North America Research Lab, pursuing research in mobile computing and web services. Before ​joining ​Nokia in 2006he was Director of the Microarchitecture Research Lab at Intel, pursuing ​research on aggressive superscalar ​and multi-core processor designsBefore joining Intel in 2000, John was a tenured full Professor in the ECE Dept. at CMUHe has returned to this former position, and is currently directing a new CMU initiative on “Human Mobility Analytics and Services” (HUMANS).