Hi,
Any better way or suggestion to test through put on various network
cards and architecture to find one somewhat meaning full numbers for
kpps other then doing timed flooding pings?
I am trying to tests a bunch of different network cards, on different
architecture with different loads for I386 and AMD64 on AMD servers to
get something not that scientific, but somewhat meaning full and valid
to compare things.
Any suggestion on how to proceed to get more valuable numbers and that
can be somewhat more comparable between different servers type.
Also, looking at ping to the loop back interface and comparing to the
network one gateway and remote, etc.
Just trying to find a somewhat valid way to do it and compare it and
isolated if possible what's the architecture limitation of the
processor, etc compare to the network card and driver itself only.
Many not be possible, but getting close to somewhat comparable number is
fine.
Input very much appreciated?
Daniel
