|
|
From what i have read on ParallelKnoppix and openMPI, all i need to do is compile applications use the wrapper, and it will 'work', aka, utilize the full power of the nodes in the cluster.
is this correct? is there any special intrinsic an application needs to work correctly? or can i just compile the source of any application and have the load spread across all the nodes?
PS-
I plan on using parallelknoppix inside my debian installation as a head, and ill be using around 40 or 50 HP pc's as my nodes. :D
However vmware didn't like my debian installation.. >.>
|