Filters








128 Hits in 4.5 sec

Towards energy-proportional datacenter memory with mobile DRAM

Krishna T. Malladi, Benjamin C. Lee, Frank A. Nothaft, Christos Kozyrakis, Karthika Periyathambi, Mark Horowitz
2012 SIGARCH Computer Architecture News  
proportionalEnergy efificient ~ 40pJ/bit  2x to 5x reduction over DDR3 9 5x LPDDR2 disadvantages  Channel bandwidth   Pin bandwidth is 2x lower  halves peak datarate per rank  Datacenter  ...  8MB shared L3 cache  2 memory channels:16 GB capacity using 2Gb DDR3, LPDDR2 chips  Validated Pin-driven simulator Datacenter workloads 17  5-6x lower DRAM power  Low active-idle states  ... 
doi:10.1145/2366231.2337164 fatcat:wpgphvlhmnfc5psm4yaafupzly

Towards energy-proportional datacenter memory with mobile DRAM

Krishna T. Malladi, Frank A. Nothaft, Karthika Periyathambi, Benjamin C. Lee, Christos Kozyrakis, Mark Horowitz
2012 2012 39th Annual International Symposium on Computer Architecture (ISCA)  
proportionalEnergy efificient ~ 40pJ/bit  2x to 5x reduction over DDR3 9 5x LPDDR2 disadvantages  Channel bandwidth   Pin bandwidth is 2x lower  halves peak datarate per rank  Datacenter  ...  8MB shared L3 cache  2 memory channels:16 GB capacity using 2Gb DDR3, LPDDR2 chips  Validated Pin-driven simulator Datacenter workloads 17  5-6x lower DRAM power  Low active-idle states  ... 
doi:10.1109/isca.2012.6237004 dblp:conf/isca/MalladiNPLKH12 fatcat:2xrf7dcx2raj7hkgxmwlkqwbui

Heterogeneous-Reliability Memory: Exploiting Application-Level Memory Error Tolerance [article]

Yixin Luo, Sriram Govindan, Bikash Sharma, Mark Santaniello, Justin Meza, Aman Kansal, Jie Liu, Badriddine Khessib, Kushagra Vaid, Onur Mutlu
2018 arXiv   pre-print
Toward this end, in our DSN 2014 paper, we make three main contributions to enable highly-reliable servers at low datacenter cost.  ...  This paper summarizes our work on characterizing application memory error vulnerability to optimize datacenter cost via Heterogeneous-Reliability Memory (HRM), which was published in DSN 2014, and examines  ...  Less reliable mobile DIMMs have been proposed [109, 156] as a replacement for ECC DIMMs in servers to improve energy e ciency.  ... 
arXiv:1602.00729v2 fatcat:l32lpfieifgq7hf6ifclekjxvi

Rethinking DRAM design and organization for energy-constrained multi-cores

Aniruddha N. Udipi, Naveen Muralimanohar, Niladrish Chatterjee, Rajeev Balasubramonian, Al Davis, Norman P. Jouppi
2010 Proceedings of the 37th annual international symposium on Computer architecture - ISCA '10  
DRAM vendors have traditionally optimized the cost-perbit metric, often making design decisions that incur energy penalties.  ...  Modern technology trends are also placing very different demands on the memory system: (i) queuing delays are a significant component of memory access time, (ii) there is a high energy premium for the  ...  Energy efficiency in datacenters has already been highlighted as a national priority [50] . Many studies attribute 25-40% of total datacenter power to the DRAM system [11, 33, 34, 37] .  ... 
doi:10.1145/1815961.1815983 dblp:conf/isca/UdipiMCBDJ10 fatcat:tpmfvqphcnhtposip6fzcwzdwm

Rethinking DRAM design and organization for energy-constrained multi-cores

Aniruddha N. Udipi, Naveen Muralimanohar, Niladrish Chatterjee, Rajeev Balasubramonian, Al Davis, Norman P. Jouppi
2010 SIGARCH Computer Architecture News  
DRAM vendors have traditionally optimized the cost-perbit metric, often making design decisions that incur energy penalties.  ...  Modern technology trends are also placing very different demands on the memory system: (i) queuing delays are a significant component of memory access time, (ii) there is a high energy premium for the  ...  Energy efficiency in datacenters has already been highlighted as a national priority [50] . Many studies attribute 25-40% of total datacenter power to the DRAM system [11, 33, 34, 37] .  ... 
doi:10.1145/1816038.1815983 fatcat:sm5tb6gxqjdallpjohmpkd5k3e

Challenges and opportunities for efficient computing with FAWN

Vijay Vasudevan, David G. Andersen, Michael Kaminsky, Jason Franklin, Michael A. Kozuch, Iulian Moraru, Padmanabhan Pillai, Lawrence Tan
2011 ACM SIGOPS Operating Systems Review  
We conclude with an outline of the longer-term implications of FAWN that lead us to select a tightly integrated stacked chipand-memory architecture for future FAWN development.  ...  This paper presents the architecture and motivation for a clusterbased, many-core computing architecture for energy-efficient, dataintensive computing.  ...  We would also like to thank Ken Mai for his help with understanding memory trends.  ... 
doi:10.1145/1945023.1945029 fatcat:ambe2hj4eze6thujboysdlenpu

Understanding the critical path in power state transition latencies

Sam Likun Xi, Marisabel Guevara, Jared Nelson, Patrick Pensabene, Benjamin C. Lee
2013 International Symposium on Low Power Electronics and Design (ISLPED)  
Increasing demands on datacenter computing prompts research in energy-efficient warehouse scale systems.  ...  With fast transitions, server activation can provide better performance at lower energy than dynamic voltage and frequency scaling.  ...  Building energy-proportional systems at this scale is a crucial part of improving datacenter energy-efficiency [3] . An energy-proportional system consumes power in proportion to its workload.  ... 
doi:10.1109/islped.2013.6629316 dblp:conf/islped/XiGNPL13 fatcat:tbtlogm3nzgdba44skxrh3txum

The Datacenter as a Computer: An Introduction to the Design of Warehouse-Scale Machines, Second edition

Luiz André Barroso, Jimmy Clidaras, Urs Hölzle
2013 Synthesis Lectures on Computer Architecture  
The relative contribution of the memory system to overall energy use has decreased over the last 5 years with respect to CPU energy use, reversing a trend of higher DRAM energy profile throughout the previous  ...  Second, memory technology shifted away from power hungry FB-DIMMs to DDR3 systems with better energy management. Third, DRAM voltage has dropped from 1.8 V down to 1.5 V and below.  ... 
doi:10.2200/s00516ed2v01y201306cac024 fatcat:435o455inbcmrakl6l7jp4gope

Causes of Energy Wastage in Cloud Data Centre Servers : A Survey

Kenga Mosoti Derdus, Vincent Oteke Omwenga, Patrick Job Ogao
2019 International Journal of Scientific Research in Computer Science Engineering and Information Technology  
Before one embarks on addressing the problem of energy wastage in a datacenter, it is important to understand the causes of energy wastage in datacenter servers.  ...  Unfortunately, datacenters consume a lot of energy, which currently stands at 3% of global electrical energy consumption.  ...  According to [40] , proportional computing is an energy efficiency technique where energy consumption by servers is proportional to the workload.  ... 
doi:10.32628/cseit1953139 fatcat:6jar3cefuzfq7gkometgbfotnm

Evaluating architecture impact on system energy efficiency

Shijie Yu, Hailong Yang, Rui Wang, Zhongzhi Luan, Depei Qian, Xiaosong Hu
2017 PLoS ONE  
The energy is tracked at subcomponent level such as Central Processing Unit (CPU) cores, uncore components and Dynamic Random-Access Memory (DRAM) through exploiting the power measurement ability exposed  ...  for High Performance Computing (HPC) and datacenter environment hosting tens of thousands of servers.  ...  The increased memory access latency due to NUMA leads to more energy consumed at the DRAM.  ... 
doi:10.1371/journal.pone.0188428 pmid:29161317 pmcid:PMC5697812 fatcat:dgvz6oyixfc2bmjjovd47piu5y

Characterizing Application Memory Error Vulnerability to Optimize Datacenter Cost via Heterogeneous-Reliability Memory

Yixin Luo, Sriram Govindan, Bikash Sharma, Mark Santaniello, Justin Meza, Aman Kansal, Jie Liu, Badriddine Khessib, Kushagra Vaid, Onur Mutlu
2014 2014 44th Annual IEEE/IFIP International Conference on Dependable Systems and Networks  
Toward this end, in this paper, we make three main contributions to enable highly-reliable servers at low datacenter cost.  ...  For example, we found that while traditional error protection increases memory system cost by 12.5%, some applications can achieve 99.00% availability on a single server with a large number of memory errors  ...  While server availability also depends on other factors, such 5 We also expect memory system energy savings to be proportional to memory capacity savings.  ... 
doi:10.1109/dsn.2014.50 dblp:conf/dsn/LuoGSSMKLKVM14 fatcat:57c4bkj3brha7jbzxosrrfwucu

Underprovisioning backup power infrastructure for datacenters

Di Wang, Sriram Govindan, Anand Sivasubramaniam, Aman Kansal, Jie Liu, Badriddine Khessib
2014 Proceedings of the 19th international conference on Architectural support for programming languages and operating systems - ASPLOS '14  
Our results show that one may be able to completely do away with DGs, compensating for it with additional UPS energy capacities, to significantly cut costs and still be able to handle power outages lasting  ...  While there has been prior work to underprovision the power distribution infrastructure for a datacenter to save costs, the ability to underprovision the backup power infrastructure, which contributes  ...  This can be more energy proportional for current technology.  ... 
doi:10.1145/2541940.2541966 dblp:conf/asplos/WangGSKLK14 fatcat:5frja3uexvdzldiriyn4qfhkd4

Power-Thermal Modeling and Control of Energy-Efficient Servers and Datacenters [chapter]

Jungsoo Kim, Mohamed M. Sabry, Martino Ruggiero, David Atienza
2015 Handbook on Data Centers  
This continuous growth in demand for computing has resulted in larger collections of servers machines, referred to as clusters or server farms, being hosted in denser datacenters thus having a higher computational  ...  Kim was also affiliated with ESL-EPFL during the period this research was developed. 1 The scaling theory he and his colleagues formulated in 1974 postulated that MOSFETs continue to function as voltage-controlled  ...  In order to be generally applicable it must be combined in synergistic ways with innovative energy-proportional server design and cooling solutions, as well as holistic datacenter thermal control.  ... 
doi:10.1007/978-1-4939-2092-1_29 fatcat:ksswhsbiqzhbbgfqgmd4bmohk4

Memory power management via dynamic voltage/frequency scaling

Howard David, Chris Fallin, Eugene Gorbatov, Ulf R. Hanebutte, Onur Mutlu
2011 Proceedings of the 8th ACM international conference on Autonomic computing - ICAC '11  
Energy efficiency and energy-proportional computing have become a central focus in enterprise server architecture.  ...  With an average of 0.17% slowdown, we show 10.4% average (20.5% max) memory power reduction, yielding 2.4% average (5.2% max) whole-system energy improvement.  ...  Energy Efficiency Techniques: As described previously, energy-proportional computing [2] is the broad research direction that characterizes this work: scaling energy use proportionally to workload demand  ... 
doi:10.1145/1998582.1998590 dblp:conf/icac/DavidFGHM11 fatcat:tbqzcplkf5bpjm3jik7aayyc6e

Server Engineering Insights for Large-Scale Online Services

Christos Kozyrakis, Aman Kansal, Sriram Sankar, Kushagra Vaid
2010 IEEE Micro  
the advent of new memory technologies such as phase-change memory, some long-term projections now indicate that all memory in computer systems could soon be nonvolatile, especially considering DRAM's  ...  scale and high availability requirements, online services run on large scale-out server farms and require system design principles to be applied not only at the server level but also holistically at the datacenter  ...  His research interests include energy-proportional computing, wireless and mobile systems, and sensor-actuator networks.  ... 
doi:10.1109/mm.2010.73 fatcat:kzfv67hr2ff4bp7fuix5zmbh3a
« Previous Showing results 1 — 15 out of 128 results