Gus Wirth wrote:
> I had a chance to get some cheap gigabit ethernet cards and a switch the
> other day with the thought that moving video files around would get a
> lot quicker. Instead I find that I'm only getting maybe a 10% increase
> in speed. My server box is using a D-Link DGE-530T with kernel module
> skge.ko and my laptop uses the Intel e1000.ko for the built-in ethernet.
> The switch is an Airlink 101 8-port gigabit ethernet switch. I get all
> the correct link lights showing I have gigabit ethernet connections at
> all ends.
> 
> If I do an NFS mount and copy a large file from the server to /dev/null
> on the client or if I scp a file between the server and client I get
> pretty much the same results, about 11MB (that's bytes) per second. This
> is only about 10% better than 100baseT ethernet. I know from previous
> experiments that my hard drives and general system throughput can handle
> about 25MB/sec.
> 
> So any hints on how to get this to work faster, or am I seeing the
> results of a marketing scam?
> 


Have you done any ttcp testing to measure raw networking throughput?
I guess there are other test programs but ttcp is one I have used.

on the receiver end
 ttcp -rs
on the transmitter end
 ttcp -ts -v receiverhostnameorip

The 's' forces 'sink'-mode with internally-generated output and
discarded input, so that there is no disk or filesystem overhead
involved at all!

You can vary the buffer size and count via -b and -n, and there are
other options you may wish to consider, too, eg, -u for UDP.

Anyway, ttcp is a nice tool, in case you haven't come across it before.


..jim


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to