Comments on: Facebook Useful? Who Knew? http://www.triplepundit.com/2009/08/facebook-useful-who-knew/ Business, Better. Since 2005 Fri, 10 Feb 2012 04:15:00 +0000 hourly 1 http://wordpress.org/?v=3.3.1 By: Jim Witkinhttp://www.triplepundit.com/2009/08/facebook-useful-who-knew/comment-page-1/#comment-15022 Jim Witkin Thu, 06 Aug 2009 23:50:07 +0000 http://www.triplepundit.com/?p=9718#comment-15022 William, you make a good point that there is still energy required to utilize the idle processor resources. But that same energy would be used by a group like SETI@home anyway to run their computations on their own computers. (SETI@Home is another GridRepublic project, btw.) I think the big savings here is on the computational hardware. By fully utilizing a microprocessor that has already been manufactured, you avoid the manufacturer of additional processors and the consequent resource and environmental impact. The distributed computing model also allows non-profits and pure research organizations to operate at a much lower cost. If they had to spend their funding on buying banks of computers, they have that much less available to run their operations and solve the big problems. So I agree, you still have to power the processors for all these additional cycles, but I think there are significant efficiencies elsewhere. William, you make a good point that there is still energy required to utilize the idle processor resources. But that same energy would be used by a group like SETI@home anyway to run their computations on their own computers. (SETI@Home is another GridRepublic project, btw.) I think the big savings here is on the computational hardware. By fully utilizing a microprocessor that has already been manufactured, you avoid the manufacturer of additional processors and the consequent resource and environmental impact. The distributed computing model also allows non-profits and pure research organizations to operate at a much lower cost. If they had to spend their funding on buying banks of computers, they have that much less available to run their operations and solve the big problems. So I agree, you still have to power the processors for all these additional cycles, but I think there are significant efficiencies elsewhere.

]]> By: William Hertlinghttp://www.triplepundit.com/2009/08/facebook-useful-who-knew/comment-page-1/#comment-15016 William Hertling Thu, 06 Aug 2009 19:15:02 +0000 http://www.triplepundit.com/?p=9718#comment-15016 I think donating processor time to these efforts are worthwhile, but it should be understood that it comes with a cost."Idle processor cycles" is a bit of a misnomer. When a processor is idle, it slows down to consume less electricity and generate less heat. When one of these processor intensive applications is run, the processor is heavily utilized, and runs at full speed.Although I haven't measured this particular application, I have measured the difference in energy consumption on other similar, distributed computation projects:A computer with a low energy consumption VIA processor normally consumed 20 watts while on and idle. While running SETI@Home, it consumed 35 watts.A computer with a more standard Intel processor that consumed 40 watts while idle, consumed 70 watts while running a computationally intensive application.Laptops will experience the biggest difference in power consumption, because they are typically designed to have an extremely low power mode. Desktops will have a smaller difference in consumption because those processors typically have a smaller variance in power consumption between idle and busy modes, and because other components of the computer consume more energy all the time.If a million users contribute "idle" cycles, that could be as much as 20 Megawatts of energy consumption. I think donating processor time to these efforts are worthwhile, but it should be understood that it comes with a cost.

“Idle processor cycles” is a bit of a misnomer. When a processor is idle, it slows down to consume less electricity and generate less heat. When one of these processor intensive applications is run, the processor is heavily utilized, and runs at full speed.

Although I haven’t measured this particular application, I have measured the difference in energy consumption on other similar, distributed computation projects:

A computer with a low energy consumption VIA processor normally consumed 20 watts while on and idle. While running SETI@Home, it consumed 35 watts.

A computer with a more standard Intel processor that consumed 40 watts while idle, consumed 70 watts while running a computationally intensive application.

Laptops will experience the biggest difference in power consumption, because they are typically designed to have an extremely low power mode. Desktops will have a smaller difference in consumption because those processors typically have a smaller variance in power consumption between idle and busy modes, and because other components of the computer consume more energy all the time.

If a million users contribute “idle” cycles, that could be as much as 20 Megawatts of energy consumption.

]]>