There is a glaring error in the text. You've quoted computation/energy in units of flops per klowatt-hour. But flops (floating point operations per second) is a unit of RATE of computation. kWh (kilowatt hours) is a unit of energy. So, either your numbers must actually be in units of flop/kWh (flop = floating point operation), or flops/kW (kW = kilowatt). From the graph, I'd guess it's the former, but the two possibilties differ by a factor of 3600.
Great article. Its great to hear about new technologies that can improve efficiencies. I think as a society advances technologically it is inevitable that it uses more energy.
In the 1950's a typical middle class house in America had a 100 amp service. Today 200 amp is the norm and in some cases barely sufficient.
So it is imperative that along with using more energy that we try to improve the efficiencies as much as we can. This should boil down to a simple economic issue in that we want to do more with the same amount of power or less. This saves us money as well as having many other desirable side affects such as leaving a smaller footprint on the environment.
An article such as this one helps to elevate this topic in the general societal conscience so thanks for writing it. I would also say that looking at all the comments that you have received a job well done.
Progress in ICs has not just been about processing power per kWh, but also about functionality per sq mm. This is why another trend building steam is 3D packaging and novel hybridization of functions and circuits. MEMS started this trend thru need to combine sensor + ASIC but nothing stops more and more circuit blocks from being added to the mix.
To a certain extend, too much reliance on the internet and mobile communication burns too much energy. Maybe we all have to think about how life can be simpler not by using so many hi-tech stuff (maybe we finally lose our job!) to save the earth.
1. There are actually many climate scientist who do not accept that man and CO2 are the primary drivers of climate change. The evidence is not at all convincing, nor is the IPCC the last or even the most current, word on the topic.
2. Not all of the global warming alarmists are actually climate scientists.
Indeed, even the much revered IPCC reports are written by governmental bureaucrats. (this last is from the guidelines of the IPCC).
3. Only those who have not read any of the papers or evidence can blithely announce that the matter is resolved. Besides since when is science a consensus model. Even Einstein stated that one well run experiment could disprove his theories.
4. Laboratory experiments prove that a doubling of CO2 can raise the temp about 1 degree C. That's a far cry from the doomsday stuff that we hear.
Blog Doing Math in FPGAs Tom Burke 23 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...