netdev
[Top] [All Lists]

Re: 2.2 performance on high network load much much better than 2.4 (fwd)

To: jamal <hadi@xxxxxxxxxx>
Subject: Re: 2.2 performance on high network load much much better than 2.4 (fwd)
From: Martin Josefsson <gandalf@xxxxxxxxxxxxxx>
Date: Tue, 9 Oct 2001 14:07:05 +0200 (CEST)
Cc: Santiago Garcia Mantinan <manty@xxxxxxxxx>, netdev@xxxxxxxxxxx
In-reply-to: <Pine.GSO.4.30.0110090737320.7036-100000@xxxxxxxxxxxxxxxx>
Sender: owner-netdev@xxxxxxxxxxx
On Tue, 9 Oct 2001, jamal wrote:

> On Tue, 9 Oct 2001, Martin Josefsson wrote:
> 
> > The recieving end was a 2 x pIII 600 server with a D-Link DFE570-TX NIC
> > (quad tulip) running 2.4.8-ac12 + tulip-ss010402-polling driver.
> >
> > The sending machine was my workstation here, a pIII 700 with eepro100 NIC
> > running 2.4.9-ac5. This machine is attached to eth1 on the server via a
> > crossover cable.
> >
> > The reciever has a few iptables modules loaded, ip_conntrack was one of
> > them. (the sender doesn't have any iptables modules loaded)
> 
> 
> Its been proven that contrack does slow down things ..

Yes I know that's why I mentioned it.

> [Good test results deleted]
> 
> Could you repeat the tests with NAPI? The interesting bit is when you
> start sending on the other ethernet ports at really high rates
> output of cat /proc/interupts and /proc/net/softnet_stat also
> /proc/net/drivers/* output

I'll see if I can rerun the tests again today with the NAPI-patched kernel
I have here (2.4.9-ac18 + 2.4.10-poll.pat + tulip-NAPI-ss011004
(mislabeled as 011010 on robur.slu.se) 
I might have to wait with these tests untill tomorrow, boring school work
that needs to be done :(

/Martin

Never argue with an idiot. They drag you down to their level, then beat you 
with experience.


<Prev in Thread] Current Thread [Next in Thread>