bonding e1000

Jon Tegner Jon.Tegner at
Sun May 12 11:19:15 PDT 2002

> 2)
> I read a post by Gordon Gere and he said that bonding of GigE adapters 
> shows a negligable performance gain.  Could someone enlighten me as to 
> why this is true?  A url would be nice if it is to time consuming of 
> an explaination.

At least for my applications (Computational fluid dynamics), I doubt 
that one would gain anything (doing bonding). Would be better to try 
using either m-via/mvich or gamma to try to reduce latency.


More information about the Beowulf mailing list