Slashdot Mirror
Ask Jonathan Koomey About 'Koomey's Law'
A few weeks back, we posted a story here that described Koomey's Law, which (in the spirit of Moore's Law) identifies a long-standing trend in computer technology. While Moore's prediction centers on the transistor density of microprocessors, Jonathan Koomey focuses instead on computing efficiency — in a nutshell, computing power per watt, rather than only per square nanometer. In particular, he asserts that the energy efficiency of computing doubles every 1.5 years. (He points out that calling this a "law" isn't his idea, or his doing — but it's sure a catchy turn of phrase.) Koomey has agreed to respond to your questions about his research and conclusions in the world of computing efficiency. Please observe the Slashdot interview guidelines: ask as many questions as you want, but please keep them to one per comment.
What is your take on the interpretation of Futurists -- like Raymond Kurzweil -- in regards to extrapolating these 'laws' out to extreme distances?
My work here is dung.
Find an arbitrary pattern or trend, then name it after yourself.
Have gnu, will travel.
This one doesn't seem to have fundamental physical limits, so long as we eventually transition to reversible computing, in which the computer does not use up useful energy because every process it uses is fully reversible (i.e. the original state could be inferred).
All the limits on computation (except regarding storage) that you hear about (e.g. Landauer limit) are on irreversible computing, which is how current architecture works. It is the irreversibility of an operation that causes it to increase entropy.
Information theory is life. The rest is just the KL divergence.
A lot of consumer grade machines have begun focusing on multicore chips with a lower frequency to provide the same or better perceived computing performance than a high frequency single core chip. What happens when a technology like this subverts our craving for higher transistor density? Can you argue that your "law" is immune to researchers focusing on some hot new technology like a thousand core processor or a beefed up system on a chip in order to improve end user experience over pure algorithm crunching speed?
My work here is dung.
I would like to see not only the Babbage engine on your curve, but also the abacus and slide rule. Maybe the physical Rod, too, which used to be used in surveying. (Hey, you try calculating property area using pencil, paper, and a deed.)
I will create a sig when innovation restarts in the U.S.
Especially in the era of cloud computing. Why does it matter? You know, even if you are not aware of it because Amazon provides you with Infrastructure as a Service, you are running virtual machines, which, wait for it: are running on real hardware :)
And of course amazon has real limitations with how many machines can be fed on a single datacenter and how much this power consumption cost is.
Note that power consumption can easily be 50% of the datacenter, that for every watt spent computing there is about 2.5 as much spent in infrastructure and cooling, and that power consumption costs can easily be greater than the cost of hardware itself. (Sorry, too lazy to provide citations but you can look it up, not hard to find)
So yeah, it does matter pretty much. :)
OK J.K here is the list of moral / ethical arguments about the path we're on, as seen in your law. You saw the path clearly enough to define a time based law. Are there any issues I'm not seeing on our current path?
1) Lower energy consumption at point of use
2) Higher energy consumption at manufacturing point
3) faster cpu = bigger programs = more bugs = lower quality of life
4) faster cpu = stronger DRM possibilities
5) Better processing * battery life = better medical devices
6) Better processing * battery life = better 1984 style totalitarian devices
7) Lower energy consumption = less air conditioning demand = decreasing average lattitude of data centers = population shifts or whatever or something?
8) More money required for both hw and sw development = good for big corps and bad for the little guy
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
Hey J.K. have you run into a law relating battery capacity (either per Kg or L) vs proc speed over time? I bet there is some kind of interesting curve for mobile devices. Or, maybe not, donno thats why I'm asking a guy with previous success at data analysis in a closely related field...
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
This law originally having been proposed by Niklaus Wirth.
While sarcastic your question is an important one: as computing power has increased, the tendency of coders to just ride over badly coded underlayers rather than redesign them competently and efficiently has increased. Why bother cutting out bloat that causes an 80% penalty on system efficiency when you can just use a more efficient chipset to get the same result?
So my question is whether Koomey has put any thought into similarly quantifying the opposing software bloat factor, and what he sees the total balance of system works out to.
Someone had to do it.
How can anyone take "cloud computing" seriously? It's really just a much less efficient version of the age old distributed computing paradigm. All it does is enable people who cannot wrap their heads around complex clustering topics to write extremely wasteful applications, and give management a new buzz-word dejour.
Someone had to do it.
Mr. Koomey, if we take your numbers from the attached article, which may not have been quoted correctly...
Feynman indicated that there was approximately 100 billion times efficiency improvement possible, and 40,000 times improvement has happened so far.
If we take Feynman's number at face value, this means that if computing efficiency improvements continue at the current rate (doubling every 18 months,) we will reach the theoretical maximum in 2043.
Based on that, do you believe that we will see a dramatic reduction in efficiency improvements in the next 10-20 years as we approach the theoretical limit, or do you think Feynman was conservative in his estimate?
Thanks!
Let's say you get 2 hrs into your sim and you realize you made a mistake in coding, forgot something,
saw initial results and realized you could trim things up to make it run better or turn out better results...
Or if you were taking the resources you were about to use more seriously, you might not be so quick to start a run before properly testing your code.
Hey, I'm all for commodity distributed computing. Would be nice if you could sell back, but even still. However, the branding of a suite of decade-old technologies as "cloud computing" has been rather silly. That's all I'm saying.
Someone had to do it.