rsharpe at richardsharpe.com
Thu Jul 10 17:01:37 GMT 2003
On Thu, 10 Jul 2003, Nicholas M. Kirsch wrote:
> > Ummm, you claim to be getting more than is possible with a single GigE
> > link, and you still want more. Or, did you simply misplace a decimal
> > point?
> My apologies for the confusion, that is 250 _megabits_ per second.
In experiments I have done using cifs_bm (cifs-load-gen) and watching
normal traces of netbench etc, I observe the following:
1. I can get 100-114MB/s from Samba 3.0.0beta1 or whatever when I use
cifs_bm with a stream of read&x requests. 
2. Windows never seems to achieve those rates because it inserts all sorts
of additional commands between reads.
I would be interested in:
1. Are the files cached?
2. Were you simply using the Windows exploder to copy the files and do you
have a trace?
3. What happens if you use a series of ReadFile calls in a simple C
BTW, there is an industry standard. Mb/s is megabits/sec, while MB/s is
 That was with the files cached and a server with 4GB of memory. The
files were of the order of 20MB to 50MB. If the server is reading off
disk, the performance slows down. OS is FreeBSD 4.6.2.
Richard Sharpe, rsharpe[at]ns.aus.com, rsharpe[at]samba.org,
More information about the samba-technical