Cold Cash, Cool Climate


Cold Cash, Cool Climate
by Jon Koomey, PhD

Main Points:

  • Jon Koomey's talk is about computing energy efficiency. He had worked at Lawrence Berkeley National Laboratory and now is at the Steyer-Taylor Center. 
  • The computing trend that will change everything- revolutionary change is going to be driven by the energy efficiency of computing and devices. Enabling proliferation of gadgets that are cheap, smart, small, connected, and low power. 
  • Example: Proteus Biomedical makes a device that goes inside a placebo pill and the digestive system juices sends a signal to a device on your skin that tells the doctor when you take your medication.
  • Research question: How has the energy efficiency of computing changed over time?
  • Number of components in an IC vs. Manufacturing Cost per component. The minimum point of the curve is the point of lowest cost of production. The progression of the minimum points of these curves is Moore's Law.
  • Work for servers- energy use, cost, and performance. 
  • Bill Nordhaus' work on economics of climate mitigation and history of computing. He compiled a huge trend of computing and cost. 
  • Time (x-axis) vs. Computations per kWh (y-axis) - NERSC super computers, Eniac, and IBM computers compiled. This is a clear upward trend with time. Around the time on the shift to transistors in the 1960's you see a clear jump. 
  • Computations per kWh= number of computations per hour at full load/measured electricity consumption at full load. This says nothing about the machine powering down, just at highest output. 
  • Every point has been normalized to Nordhaus' database- there is lots of data on computational data by the Office of Naval Research. 
  • Doubling time for performance per computer was every 1.5 years in the PC era.
  • Data came from: Microsoft computer archives, LBNL, Erik Klein computer archives, Computer History Museum in Mountain View
  • Efficiency of Computing: R^2 = 0.983 for computers and 0.970 for PC's
    • Doubling every year and a half since the 1940's
      • computers = 1.6 years
      • PC's  = 1.5 years
      • Vacuum tubes= 1.35 years
    • 100x improvement every decade
    • Enabled the existence of laptops and smart phones
    • Big Jump from tubes to transistors
  • In 2009 laptops outsold desktops for the first time, and they will continue to do so moving forward. This term is called the "installed base"
  • Josh Smith has developed sensors that scavenge energy from stray radio and TV signals (active mode is using 60 microwatts). 
  • Back in 1985, Richard Feynman calculated a theoretical limit to the efficiency of a transistor. He assumed the max to be a 3 atom transistor. Sometime in the next few decades we need to radically change the way we view computing. Earlier this year, researchers at Purdue created a reliable 1 atom transistor (uses energy levels within atoms to do switching). Only problem is that it runs at liquid helium temperatures.
  • Big unanswered questions:
    • Are there technological innovations (software or hardware) that could allow us to substantially exceed the historical trend in the electrical efficiency of computation
    • What roadblocks might prevent these trends from continuing after the current innovation pipeline is exhausted?
    • Whats next after Richard Feynman's theoretical limit?
  • Big picture implications:
    • Low power is more important than high efficiency
    • Revolution is being driven by the confluence of trends allowing low-power
      • computing
      • communications
      • sensors
      • controls
    • Energy harvesting and storage also critical
    • Idle modes more important than active
  • Conclusions
    • Quantitative results
      • In the PC era (1976-2009) performance per computer and computations per kWh doubled every 1.5 years
      • From ENIAC to the present, computations per kWh doubled every 1.6 years
    • Performance and efficiency improvements inextricably linked
    • Still far from theoretical limits
    • Big implications for mobile technologies
    • Focus is now on low power

Watch the Google lecture by Jon Koomey on The Computing Trend That will Change Everything!

 Cold Cash, Cool Climate makes many references to the 2009 MIT Joint Program On the Science and Policy of Climate Change in which a particular "no-policy" course of action is juxtaposed against different levels of action (Level 1, 2, 3, etc.). Read the report (here).

Comments

Post a Comment

Popular posts from this blog

Consider a Spherical Cow

Image
Consider a Spherical Cow: A Course in Environmental Problem Solving by John Harte Main Points: What is the "spherical cow" joke ? The phrase comes from a joke about theoreticians I first heard as a graduate student. Milk production at a dairy farm was low so the farmer wrote to the local university, asking help from academia. A multidisciplinary team of professors was assembled, headed by a theoretical physicist, and two weeks of intensive on-site investigation took place. The scholars then returned to the university, notebooks crammed with data, where the task of writing the report was left to the team leader. Shortly thereafter the farmer received the write-up and opened it to read on the first line: "Consider a spherical cow..." The Approach : The spherical cow approach to problem solving involves the stripping away of unnecessary detail, so that only the essentials remain. Of course, approaching the complex world from the spherical cow perspective
Read more

Should We Risk It?

Image
Should We Risk It? by Daniel M. Kammen & David M. Hassenzahl Main Points: " I began by trying to quantify technical risks, thinking that if they were 'put into perspective' through comparison with familiar risks we could better judge their social acceptability. I am ashamed now of my naivety, although I have the excuse that this was more than 20 years ago, while some people are still doing it today. "- Harry Otway (1992) Background: Daniel Kammen teaches a course at UC Berkeley called "Environmental Classics" where they rotate through key papers/books like Amory Lovins' paper "A Road Not Taken", the Wedges paper. They read the original items- the critiques that came at the time and the critiques that come after. Only Silent Spring  becomes a fixture in the course year after year.  Take a look at Kammen's lab RAEL and BREG  for sustainability tools and what tools we need to alter the dialogue around clean energy or sus
Read more

Read one Book a Week

I once participated in an energy club in grad school. The club was composed mainly of grad students from several departments and disciplines - engineering, business, and law. Our mission was to foster a diverse energy community and enable interdisciplinary collaboration, knowledge exchange, and professional development. Over the course of two years, the club grew in size from 20 to over 100 students, creating its first alumni mentoring program, taking tours of several energy companies, hosting its first energy case competition, installing the largest solar panel array in the local neighborhood, building a vertical axis wind turbine, and collaborating with several organizations around the nation. I loved it. Mainly because of all the great friends who made my time in the club truly memorable. Much of what we were trying to do was to understand the energy environment from those who worked, taught, and preached about it daily. While talks and small group dinners got us interested, I fe
Read more