Greg Lindahl wrote:
time on all your nodes, which should improve performance when you have
big clusters and a lot of synchronization in your code.

Provided it scales well and it's integrated in the kernel. I guess You could also think about revisiting the gang scheduling ideas with a better synchronization.

Patrick
_______________________________________________
Beowulf mailing list, [email protected]
To change your subscription (digest mode or unsubscribe) visit 
http://www.beowulf.org/mailman/listinfo/beowulf

Reply via email to