11 Responses

  1. The Death and Rebirth of a Windows Home Server « MS Windows Home Server
  2. Dustin Wyatt
    Dustin Wyatt December 10, 2008 at 9:15 am |

    And, if you’re going to have a PC running 24/7/365, it’s better for the planet to donate what that system is doing when it is idle.

    This isn’t necessarily true. Folding causes your computer to use much more power than when it’s just sitting there idle.

    Reply
  3. Christopher Price
    Christopher Price December 10, 2008 at 4:44 pm |

    The power increase is very negligible. If every home server was doing Folding or other scientific grid research, we would greatly accelerate the research that will find cures to all sorts of diseases.

    And, curing things like cancer is certainly good for the environment as well, it reduces the environmental impact of people dramatically each time we cure a major disease.

    Granted, some projects like SETI@Home may never produce a result, which is generally why I promote grid-based research projects that have a reasonable return on investment horizon.

    Reply
  4. Dustin Wyatt
    Dustin Wyatt December 10, 2008 at 7:26 pm |

    I didn’t say it wasn’t always true, I said it wasn’t necessarily true. Under many circumstances, loading the GPU/CPU increases power draw significantly

    For example, Toms Hardware shows systems doubling their power usage when idle vs under load.

    Anandtech shows many CPU’s and GPU’s doubling (or more) their power requirements when under load.

    They again show complete systems increasing their power draw greatly whilst under load.

    There is all kinds of factors that can contribute to how much of an increase in load you’re going to be looking at. Additionally, how many diseases have been cured via Folding? Not that it isn’t a great thing to donate your cpu/gpu time to, but don’t kid yourself that it’s always going to be greener to do so.

    Statements such as these…

    Donating idle computer resources to research is a lot more “green” than falsely-titled “green” PCs (which often cut back on performance dramatically, in order to save pennies on your power bill each year… literally). Put your power to good use, and keep that in mind when buying your (next) Home Server.

    …are overblown.

    Reply
  5. Christopher Price
    Christopher Price December 10, 2008 at 9:13 pm |

    I realized the distinction between “not necessarily true” and “not true”. However, marginal analysis has to come into play… both on that point, and on the environmental aspect.

    Even if the power consumption doubles (as I will explain below, not unreasonable either in perspective), that’s still much more efficient than having to construct a supercomputer grid in a lab, to do the same analysis.

    Furthermore, power analysis that says that consumption doubles, is short sighted. If you redesign a CPU to have added idle parameters (much like Intel has on Penryn and Nahalem cores), sure, you’re going to see a higher percentage increase from folding in terms of watts used.

    That said, if your computer runs on the same power as a light bulb, and you double it… you’re donating necessary scientific research at the cost of running a light bulb. I’m sure most would agree that’s not a significant price to pay in terms of donating efforts.

    Now, the era of GPU computing changes things quite a bit in terms of power. Specifically, the amount of research contributed from that “cost of a light bulb” is now multiples more than before.

    As to the aspect of “number of diseases cured”… that kind of logic is impossible to assert. There never will be a single key to curing diseases that require intervention at the protein level. It will take a the symphony of humanity’s research and development resources to come up with a cure. Grid-based research is just one of those aspects, and you’re free to see the list of scientific advances the Folding@Home team has come up with, independent of any other grid project.

    Grid computing is the future of research, and thankfully, it’s great for the environment. I stand by my original statement as being perfectly correct. And, I encourage every Home Server user to donate their unused resources (either to Folding@Home, or any other grid project that they are interested in).

    Reply
  6. Dustin Wyatt
    Dustin Wyatt December 11, 2008 at 7:51 am |

    Even if the power consumption doubles (as I will explain below, not unreasonable either in perspective), that’s still much more efficient than having to construct a supercomputer grid in a lab, to do the same analysis.

    Not necessarily correct. The power distribution network to hundreds or thousands of computers spread across the world is not going to always be more efficient than the power distribution to one point of presence.

    Furthermore, power analysis that says that consumption doubles, is short sighted. If you redesign a CPU to have added idle parameters (much like Intel has on Penryn and Nahalem cores), sure, you’re going to see a higher percentage increase from folding in terms of watts used.

    Not a valid point. The only thing this proves is that you have even more of an incentive to idle your CPU. That very same CPU is going to be either idling or maxed out if you run a F@H client. You run it maxed out you double the power usage.

    I’m not sure what the idle power usage of older CPU’s even has to do with this. The Dell you say you’re using now, is using one of these new CPU cores! You max out your very own CPU, and you double it’s power consumption. Same with your GPU.

    That said, if your computer runs on the same power as a light bulb, and you double it… you’re donating necessary scientific research at the cost of running a light bulb. I’m sure most would agree that’s not a significant price to pay in terms of donating efforts.

    But that’s not what you originally claimed! Your claim was that it was [b]more[/b] green to donate idle time. I have no problem with the opinion that Folding and similar projects are worth the extra power costs.

    As to the aspect of “number of diseases cured”… that kind of logic is impossible to assert. There never will be a single key to curing diseases that require intervention at the protein level.

    “Number of diseases cured” isn’t a line of reasoning unto itself and thus isn’t a “kind of logic”, it’s a piece of evidence contributing to the line of reasoning that demonstrates that running a distributed client isn’t necessarily greener than not.

    Grid computing is the future of research, and thankfully, it’s great for the environment.

    You’ve not presented any evidence of this.

    And, I encourage every Home Server user to donate their unused resources (either to Folding@Home, or any other grid project that they are interested in).

    I do too, but people shouldn’t do it thinking that they’re being green.

    Reply
  7. Christopher Price
    Christopher Price December 11, 2008 at 7:04 pm |

    So you really believe that using existing hardware, existing PCs, to do supercomputer work, is not as green as building (and powering) an independent supercomputer facility to run the research? Please!

    The CO2 used to manufacturer a supercomputer lab alone justifies grid computing at home, as a green technology. I don’t even have to get into a power debate on that level… but I still fully maintain that using one lightbulb of power (in terms of idle resources) is better than using two (to power one blade in a supercomputer lab). So, it is green on both levels.

    You just don’t like the evidence I’ve presented… but don’t create a straw man and say I’m making a baseless argument. It undermines your own argument.

    Reply
  8. Dustin Wyatt
    Dustin Wyatt December 12, 2008 at 12:53 pm |

    So you really believe that using existing hardware, existing PCs, to do supercomputer work, is not as green as building (and powering) an independent supercomputer facility to run the research? Please!

    I claimed no such thing. I do claim that if a research team already has the equipment, or has the ability to farm the job to already existing supercomputers, that is the greener approach.

    Since your assumptions about my point are incorrect, I have constructed no strawman.

    You seem unwilling to accept that your matter-of-fact statements about the “greeness” of running distributed computing are not a matter-of-fact but needs evaluated case-by-case.

    It’s not that I just don’t like the evidence you’ve presented, you’ve presented incorrect evidence and have not addressed any point I’ve raised.

    Reply
  9. Christopher Price
    Christopher Price December 12, 2008 at 6:23 pm |

    If you can find a real-world supercomputer lab that can crunch numbers as fast as what Folding@Home can do… I’d love to see it. The fact of the matter is that this type of research (faster than realtime computing) would require a grid of supercomputer labs to accomplish what folks are donating at home.

    I’m a realist, I’m not going to rationalize that if Stanford managed to acquire 15 existing supercomputing facilities, that it might be slightly greener than using home resources. I don’t accept that’s true, and I doubt it could be accomplished.

    On the flip side of the coin, being a realist, I refuse to consider unrealistic scenarios that wouldn’t even require grid work. If the task doesn’t even warrant grid computing in the first place, of course it would be stupid environmentally to grid it out.

    If we have to consider unrealistic scenarios such as those, in every argument, it doesn’t make sense to ever recommend anything, to anyone. We’ll just have to agree to disagree.

    Reply
  10. Windows Home Server Trial Now in Downloadable Fashion | CentralGadget.com
  11. Drk
    Drk July 25, 2010 at 4:11 pm |

    Hey folding@home has produced a number of results for Stanford. I love my WHS and I am using that folding@home addin available here http://www.homeserverland.com/blogs/b/hslblog/archive/2010/06/09/folding-at-home-on-windows-home-server.aspx

    Reply

Leave a Reply to Christopher Price Click here to cancel reply.