c31 lecture 16

Impact on people negligible 1950 2000 desktop

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: 900-1950: prototype devices Impact on people: negligible 1950-2000: desktop computer, enterprise servers, first networks Impact on people: medium (simple automation of tasks) 2000’s: the revolution Worldwide connection, devices connected, human knowledge accessible anywhere by a click, 9 The future of computing The desktop PC is dead (it was a transition only) Home/house servers will become more frequent Each device can compute/communicate TV, home appliances, cars, wearable devices, … For the next decade (at least), three main categories 10 TITAN (2013) #1 supercomputer (as of today), cost: $100M Uses ~20,000 AMD processors + ~20,000 Nvidia GPUs Space: ~4000 sqf Power: 8.2 MW Computing speed: 27 petaFLOPs (or 27,000,000,000,000,000 operations per second) 11 iPhone 5 (2012) Handheld computing and communicating device Cost: ~$200. Weight: 112 grams Power: < 3 Watts peak Computing power: 25 GF/s (or 25,000,000,000 operations per second) About 1/4th of a Xbox 360 12 Google server farm (2012) Built of XXX (millions?) commodity Linux servers Objective: best performance per dollar Surface? Numerous buildings at least Power: estimated aggregated 500-600 MW Computing power: estimated 20-100 petaFLOPs 13 The programming issue… Who will program applications to run on those devices? And to do what? Supercomputers Scientists, engineers, You? Cloud You? Handheld device You! 14 Scientific breakthrough / life changer Now computers have become powerful enough to Now computers have become powerful enough to model real-life conditions Less/no need for physical testing (i.e., drug testing) Computer-aided design: full-scale systems (factory, nuclear reactor, etc.) Weather modeling: full-earth with good resolution Physics experiments: atom/quantum level Now networks (wired, wireless) have become fast enough to communicate “any” data Now storage has become large enough to store “all” the human knowledge 15 Anton (2010’s) Dedicated supercomputer for molecular dynamics Goal: do drug development “entirely” with a computer Challenge: study the protein interactions for “long...
View Full Document

This note was uploaded on 04/03/2014 for the course CS 31 taught by Professor Melkanoff during the Fall '00 term at UCLA.

Ask a homework question - tutors are online