For performance testing, scp just isn't right. SCP encrypts/decrypts, so it uses a lot of CPU... You don't wanna bench the CPU ? To test raw bandwith and latency I would suggest: TTCP: That you can get at http://www.mentortech.com/learn/tools/tools.shtml even better: NetPerf: http://www.netperf.org/ These tools just dump data on network interfaces and mesure raw bandwith. If you don't want to install that, try FTPing from one machine to another, not SCPing. btw: I'm very interested by the results you can get with these cards... please post! cheers, Nick. On Thu, Aug 09, 2001 at 08:46:02PM -0700, David Radclyfe said: > 2) Before trying to run any real code, I have tried > some basic performance tests -- using scp and just > blasting small (~256 byte) UDP packets from one > machine to the other and seeing how long it takes to > send 100k packets. In both cases, the gigabit cards > have only marginal performance gains over the 100Mbit > card -- a 150MB file take 13s rather than 15s to send > using scp. Maybe this is just some other bottleneck, > but, what do you guys do to measure bandwidth? I want > to convince myself I haven't missed some other setup > to make the cards work for gigabit and they are > running properly. I expected better performance. Again > I have tried both the Becker and Intel modules. > > Thanks in advance! > > David > > __________________________________________________ > Do You Yahoo!? > Make international calls for as low as $.04/minute with Yahoo! Messenger > http://phonecard.yahoo.com/ > - > : send the line "unsubscribe linux-net" in > the body of a message to majordomo@vger.kernel.org > More majordomo info at http://vger.kernel.org/majordomo-info.html -- Nicolas Sayer <Nick@loplop.net> +33 6 63.44.02.27 - : send the line "unsubscribe linux-net" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
