From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ben Greear Subject: Re: 2.6.7 tulip performance (with NAPI) Date: Wed, 06 Oct 2004 14:37:11 -0700 Sender: netdev-bounce@oss.sgi.com Message-ID: <41646587.7070401@candelatech.com> References: <41633174.7070805@candelatech.com> <16740.17875.574967.11417@robur.slu.se> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Cc: "'netdev@oss.sgi.com'" Return-path: To: Robert Olsson In-Reply-To: <16740.17875.574967.11417@robur.slu.se> Errors-to: netdev-bounce@oss.sgi.com List-Id: netdev.vger.kernel.org Robert Olsson wrote: > Ben Greear writes: > > > I tried setting the NICE level of pktgen to -10 and softirq to -18. I > > still see way more packets transmitted than received. > > If you have "max_before_softirq" in your version pktgen you can try it > to balance your load. HZ=1000 in 2.6 can make scheduling different. Yes, I was able to get it to smooth out by doing something like that. I also got better desktop performance by simply decreasing the niceness of the pktgen thread to 0 instead of -10. On a related note, I am now working on a way to use a hook in the netif_wake_queue callback to wake up pktgen. This should allow me to have a pktgen that does not need to spin in a tight loop like it does now. So far, I was able to saturate two GigE ports using about 3% of the CPU (as reported by top), using 1514 byte pkts. Still tweaking to fix some corner cases... Thanks, Ben -- Ben Greear Candela Technologies Inc http://www.candelatech.com