The GreenLight Instrument marks a significant milestone in the effort to "green" modern campus cyberinfrastructure. Campus networks, which enable computational and data-intensive approaches to research, typically are housed in ad hoc and sub-optimal energy environments in departmental facilities. As the amount of energy usage per compute rack increases (about 2 KW/rack in 2000 to an estimated 30 KW/rack in 2010), these computing clusters require more sophisticated and energy efficient cooling and power mitigation.
Development of the GreenLight Instrument is enabling communities of application scientists — drawn from metagenomics, ocean observing, microscopy, bioinformatics, and the digital media — to better understand how to measure and minimize energy consumption by leveraging UCSD's innovative energy/cooling sources and employing middleware that automates and optimizes compute/power strategies.
Using OptIPuter and Quartzite methods developed under earlier NSF awards, cumbersome computing clusters have been relocated to pre-fabricated campus "machine rooms," where they are remotely monitored for energy efficiency via the GreenLight Instrument GLIMPSE measurement system (http://glimpse.calit2.net). GreenLight is providing, via service-oriented architecture methodology, real-time sensor outputs that allow researchers to study the energy cost of at-scale scientific computing from anywhere in the world.
GreenLight enables researchers to make deep and quantitative explorations in advanced computer architectures, including alternative circuit fabrics, direct-graph execution machines, graphics processors, solid-state disks, and photonic networking. The resulting quantitative data allows engineers to compare "computational work per watt" across full-scale applications running on at-scale computing platforms and transform systems engineering for green cyberinfrastructure.
Development of the GreenLight Instrument is enabling communities of application scientists — drawn from metagenomics, ocean observing, microscopy, bioinformatics, and the digital media — to better understand how to measure and minimize energy consumption by leveraging UCSD's innovative energy/cooling sources and employing middleware that automates and optimizes compute/power strategies.
Using OptIPuter and Quartzite methods developed under earlier NSF awards, cumbersome computing clusters have been relocated to pre-fabricated campus "machine rooms," where they are remotely monitored for energy efficiency via the GreenLight Instrument GLIMPSE measurement system (http://glimpse.calit2.net). GreenLight is providing, via service-oriented architecture methodology, real-time sensor outputs that allow researchers to study the energy cost of at-scale scientific computing from anywhere in the world.
GreenLight enables researchers to make deep and quantitative explorations in advanced computer architectures, including alternative circuit fabrics, direct-graph execution machines, graphics processors, solid-state disks, and photonic networking. The resulting quantitative data allows engineers to compare "computational work per watt" across full-scale applications running on at-scale computing platforms and transform systems engineering for green cyberinfrastructure.