On Tue, Sep 15, 2009 at 06:28:33PM +0200, frantisek holop wrote: > hmm, on Tue, Sep 15, 2009 at 08:41:22AM +0200, Claudio Jeker said that > > Hah. That's why he did not update his site since 2003. Do you realy think > > that OpenBSD 3.4 and 4.6 are the same? > > nobody is arguing 3.4 and 4.6 is the same. > or that that particular benchmark has any more than historic value.. > > no matter how you put it, both a well done and an incompetent > benchmark succeeds on some level: some things were looked into because of it. >
Yes but it was surely not your merit. > > The reason for C has nothing to do with speed. > > "nothing" is a strong word. but be it. c and speed > shall never be used in the same sentence. > > > And here again comes this style of uninformed dumb rant. Why do you think > > a web server will not do that much without sendfile()? Honestly it is > > exactly the opposide, a web server that never touches the disk for content > > delivery will outperform all others and can server enough data to fill a > > can you show me the magic openbsd webserver that never touches the disk? > Write a cgi or better apache module that outputs a static HTML hello-world page and see for your self. Sure this webserver would not be to useful but that was not requested :) > since you picked on sendfile(), let's turn it around. i dont see your > numbers to prove that sendfile() doesn't help. how is your statement > any better informed than mine? it's just easier to say: nah, it wouldn't > help that much at all, it's overrated anyway. > Sure, send me an implementation for sendfile() on OpenBSD and I will do the benchmarking for you. > > having said all of this, i've been here for some years now. i know > speed is not the priority for this project -- and i am fine with that > otherwise i wouldn't be here. i am simply sick of all this downplaying > of any and all benchmarks (micro and macro), when everyone who writes > programs knows that there are important benchmarks for finding > bottlenecks and improving program/data structure. > I think the developers know how benchmarks work and how to interpret the results, we don't need a huge flamefest on misc@ to make us understand how great benchmarks are. > i would be more than happy to see a "competent" benchmark by someone > who knows openbsd intimately. but exactly because of the prevailing > mindset no one ever bothers. so when it comes to openbsd performance > it's all just legends and hearsay, and if being low, hardware, software > and user gets blamed. > In 95% of the cases that answer is correct. It sucks but it is true. Sure there are parts in OpenBSD that are far from optimal (threading and SMP for example) but I think those are known for people considering to use OpenBSD for a high performance system going into production. > the devs are sitting around smug saying nothing until someone comes > up with this theme again and then they send something like: > > i have a bgp machine forwarding 800MBit/s of real world generic > internet traffic. can handle at least twice that. > > what can i say? i am happy for you :] Yes, not everybody has 800Mbps of upstream traffic. -- :wq Claudio
