From mboxrd@z Thu Jan 1 00:00:00 1970 From: Manfred Spraul Subject: Re: Early SPECWeb99 results on 2.5.33 with TSO on e1000 Date: Tue, 10 Sep 2002 18:55:35 +0200 Sender: netdev-bounce@oss.sgi.com Message-ID: <3D7E2407.6060209@colorfullife.com> References: <3D78F55C.4020207@colorfullife.com> <20020906.113829.65591342.davem@redhat.com> <3D790499.8020501@colorfullife.com> <20020906.123428.28085660.davem@redhat.com> <15741.57164.402952.136812@robur.slu.se> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Cc: "David S. Miller" , haveblue@us.ibm.com, hadi@cyberus.ca, netdev@oss.sgi.com, linux-kernel@vger.kernel.org Return-path: To: Robert Olsson Errors-to: netdev-bounce@oss.sgi.com List-Id: netdev.vger.kernel.org Robert Olsson wrote: > > Anyway. A tulip NAPI variant added mitigation when we reached "some > load" to avoid the static interrupt delay. (Still keeping things > pretty simple): > > Load "Mode" > ------------------- > Lo 1) RxIntDelay=0 > Mid 2) RxIntDelay=fix (When we had X pkts on the RX ring) > Hi 3) Consecutive polling. No RX interrupts. > Sounds good. The difficult part is when to go from Lo to Mid. Unfortunately my tulip card is braindead (LC82C168), but I'll try to find something usable for benchmarking In my tests with the winbond card, I've switched at a fixed packet rate: < 2000 packets/sec: no delay > 2000 packets/sec: poll rx at 0.5 ms -- Manfred