Clustering in LINUX - The Real Power Of Penguin

Discussion in 'Open Source' started by s4saurabh, Oct 28, 2004.

Thread Status:
Not open for further replies.
  1. s4saurabh

    s4saurabh New Member

    Joined:
    Sep 17, 2004
    Messages:
    8
    Likes Received:
    0
    Trophy Points:
    0
    Location:
    India
    Recently i developed a two node beowulf cluster in LINUX when i realized the power of clustering in LINUX for parallel and supercomputing. Now anyone tell me where does Mr. $ of Microsof t offer such a power to normal user and administrators
     
  2. GNUrag

    GNUrag FooBar Guy

    Joined:
    Jun 22, 2004
    Messages:
    1,246
    Likes Received:
    5
    Trophy Points:
    0
    Location:
    Interwebs
    Great achievement..... Its only with UNIX based solutions can you achieve such goals.....

    by the way, what framework did you use for that ? Oscar or OpenMosix ?
     
  3. pradeep_chauhan

    pradeep_chauhan New Member

    Joined:
    Aug 16, 2004
    Messages:
    1,150
    Likes Received:
    0
    Trophy Points:
    0
    Location:
    New Delhi
    what only 2nodes come on min 10 nodes then see the fun involve all friends neighbours enemies relatives .......
     
  4. GNUrag

    GNUrag FooBar Guy

    Joined:
    Jun 22, 2004
    Messages:
    1,246
    Likes Received:
    5
    Trophy Points:
    0
    Location:
    Interwebs
    Yes... 2 is not fun.... But then if number of nodes is high... then all of them should be running processors of same clock speed.... i have read that in OpenMosix, for example.... the collective computing power of the whole cluster cripples down to the clock speed of the slowest not in the cluster...
     
  5. OP
    OP
    s4saurabh

    s4saurabh New Member

    Joined:
    Sep 17, 2004
    Messages:
    8
    Likes Received:
    0
    Trophy Points:
    0
    Location:
    India
    Yups i used Oscar
     
  6. pradeep_chauhan

    pradeep_chauhan New Member

    Joined:
    Aug 16, 2004
    Messages:
    1,150
    Likes Received:
    0
    Trophy Points:
    0
    Location:
    New Delhi
    See Building the cluster is the easy task putting it to use or writing an application that uses the distributed setup is the big challenge. you can enjoy having the benchmarks or doing the calculations of performance for some time then the reality begins to sink in there are no distributed applications using the setup there where the hitch is and one should work on.
     
Thread Status:
Not open for further replies.

Share This Page