Coffeehouse Thread

16 posts

Forum Read Only

This forum has been made read only by the site admins. No new threads or comments can be added.

Distributed Computing

Back to Forum: Coffeehouse
  • User profile image
    the-​laughing-man

    I first heard about SETI@ home about 5 years ago but never really got into it for some reason. A year ago I decided to take a deeper look into it and found Grid.org and Universal Devices.

    I saw that the BBC has jumped on the band wagon today with a distuributed system to look into Climate Control, another worthy cause.

    I was wondering if any of you guys are running any of these systems like Folding@ Home, SETI@ home, UniversalDevices and if so how long have you been crunching? Wink

  • User profile image
    Cybermagell​an

    the-laughing-man wrote:
    I was wondering if any of you guys are running any of these systems like Folding@ Home, SETI@ home, UniversalDevices and if so how long have you been crunching?


    I've ran Folding and Seti....I quit after awhile...

  • User profile image
    the-​laughing-man

    Why'd you quit? :O

  • User profile image
    W3bbo

    the-laughing-man wrote:
    Why'd you quit?


    I ran Grid.org instead of SETI@Home since I felt it a more ethical use of processing time.

    I ran it for a few months on all of my computers, I reached rank #77,000-something after donating a few weeks worth of processor time.

    I stopped using it after I formatted my workstation and uninstalled it from my LAN's server because I learned that power-consumption goes through the roof.

  • User profile image
    the-​laughing-man

    I'm running Grid.org instead of SETI for the same reason, w3bbo Big Smile

    I figured I may aswell do some help while my PC is on Big Smile Power consumption isn't much of a worry, right now for instance I have no heating on, and just my PC running in this room, not even got lights on, the nice warm glow of a monitor is enough for me Wink

  • User profile image
    AIM48

    Well I once wrote my own.

    Won't mention any company by name - but I was once brought in to solve a sticky problem for a comapny.

    The company stored all of its credit card numbers as MD5 hashes and VISA came out with a new requirement that numbers have to be stored as SHA-1. There is no way to cnvert from md5 to sha-1 (it's one way encryption). 

    We had a list popular bins (first 6 numbers of the credit card numberof  I wrote a distribution server and a client screensaver that distributed fragments of the credit card number space to about 20 workstations. The screensavers kicked in and did the crucnching. They took the keyspaces they were givin and ran thru a bunch of numbers MD5ing them all - if it matched one off the outstanding MD5 then it was sha-1ed and stored.

    To my great suprise it worked! after about a month we found about 98% percent of all of the creditcared numbers.

    That was a fun project.

  • User profile image
    Maurits

    AIM48 wrote:
    To my great suprise it worked! after about a month we found about 98% percent of all of the creditcared numbers.


    A word to the wise... salt your hashes.

  • User profile image
    AIM48

    Maurits wrote:
    AIM48 wrote: To my great suprise it worked! after about a month we found about 98% percent of all of the creditcared numbers.


    A word to the wise... salt your hashes.


    No - They were getting the encrypted hashes from third parties.

  • User profile image
    Mike Dimmick

    the-laughing-man wrote:
    I saw that the BBC has jumped on the band wagon today with a distuributed system to look into Climate Control, another worthy cause.


    Climate change will happen a lot faster if everyone leaves their computer on and running a distributed system, rather letting the system idle or switching off. Modern desktop and even server processors do go into power-saving states when idle, which saves a little power, but disks and fans keep running. The original ADSL specification doesn't make any allowance for power saving states, although the newer ADSL2+ does.
     
    Saving power by switching off equipment you're not using is much more important. I always switch off my home PC when I'm not actively using it although that's at least partly due to how noisy the fans are. I leave my work PC on if I'm expecting updates, but otherwise turn it off at the end of the day.

  • User profile image
    sbc

    Mike Dimmick wrote:
    the-laughing-man wrote:I saw that the BBC has jumped on the band wagon today with a distuributed system to look into Climate Control, another worthy cause.


    Climate change will happen a lot faster if everyone leaves their computer on and running a distributed system, rather letting the system idle or switching off. Modern desktop and even server processors do go into power-saving states when idle, which saves a little power, but disks and fans keep running. The original ADSL specification doesn't make any allowance for power saving states, although the newer ADSL2+ does.
     
    Saving power by switching off equipment you're not using is much more important. I always switch off my home PC when I'm not actively using it although that's at least partly due to how noisy the fans are. I leave my work PC on if I'm expecting updates, but otherwise turn it off at the end of the day.

    It is quite ironic leaving a computer on to research climate control.

    Although I wonder how much damage it actually does compared to the benefits it may bring (if your computer is going to be on anyway, it could be used for something useful).

    Perhaps this is best done on laptops - unplug from mains and leave it running till the battery runs out (if you are someone who runs down a battery to prolong its life).

  • User profile image
    ScanIAm

    Before anyone get's too concerned about how much power is used by a PC, take a close look at common kitchen appliances.  Or an Iron.  Or a space heater.

    If you don't run a webserver on your home machine, then by all means feel free to turn it off out of principle, but turning your thermostat down in winter/up in summer will do a much better job of 'saving the planet'.

  • User profile image
    the-​laughing-man

    an idle computer doesn't use much at all however w3bbo was right, when you begin using a program like any of the ones mentioned here your computer is far from idle and will be using a lot more energy.

    I figure we'd do better spending our time trying to cure disease with our PCs while we can, cause let's face it sooner or later America will eat up all the resources. They seem to want no part in Kyoto

  • User profile image
    rhm

    I'd like to see some hard figures on that. My personal feeling is the PC power supplies are so inefficient that the power drain from the CPU is a small part of the total power consumption. Oh, and you know that ATX PSUs draw power even when the machine is switched off? I hope you switch all your machines off at the wall if you're that bothered about it.

  • User profile image
    rhm

    Oh, FWIW I've run Folding@home for years because it seems that they are doing basic science that not only has some useful application, but will be valid for all time rather than just solving some specific problem.

  • User profile image
    msemack

    rhm wrote:
    I'd like to see some hard figures on that. My personal feeling is the PC power supplies are so inefficient that the power drain from the CPU is a small part of the total power consumption. Oh, and you know that ATX PSUs draw power even when the machine is switched off? I hope you switch all your machines off at the wall if you're that bothered about it.


    On the x86 Single-board computers we design, we usually see an power consumption increase of 40-50% when the CPU goes from idle to full load.  Note that this is just a measurement of the "motherboard power".  To get an accurate picture of how this affects your total system power consumption, you have to weight it against the power requirements of your disk drives, monitor, etc.

    Regarding power supply efficiency:

    Your average ATX power supply has an efficiency between 60 and 80%.
    http://www.tomshardware.com/2005/02/28/strong_showing/page38.html

    The ATX12V specification requires a power supply to be at least 65% efficient.
    http://www.formfactors.org/developer%5Cspecs%5CATX12V_PSDG_2_2_public_br2.pdf

    As for how much power is used in S5/soft-off:

    The amount is usually negligible (well under 1 watt at the motherboard).  In S5, you only need enough power to keep a few circuits in the southbridge active.

  • User profile image
    krgreenlee

    FYI.  If anyone is interested in playing around with a computational grid using your own code I welcome you to request a Developer Edition of the Digipede Network from http://www.digipede.net/faster">www.digipede.net/faster and use the Promotion Code: CHANNEL9.  It's free, although we are currently tracking contact information.

     

    The Digipede Network is built on .NET and has both.NET and COM APIs available.  I do work for Digipede so...there is my disclaimer.  But I’ve had a lot of opportunity to play around with what the Digipede Framework enables and so far I’ve been able to distribute computations for console, Windows Forms, Excel, and web service type applications. I have more projects that I think will be fun to try but I haven’t had the time.  If anyone does give this a go, I would love to hear about what you built. 

     

    Thanks,

    Kim Greenlee
    Digipede Technologies
    http://www.digipede.net/">http://www.digipede.net
    http://krgreenlee.blogspot.com/">http://krgreenlee.blogspot.com

Conversation locked

This conversation has been locked by the site admins. No new comments can be made.