Return to BSD News archive
Newsgroups: comp.os.386bsd.questions Path: sserve!newshost.anu.edu.au!munnari.oz.au!constellation!osuunx.ucc.okstate.edu!moe.ksu.ksu.edu!vixen.cso.uiuc.edu!howland.reston.ans.net!cs.utexas.edu!uunet!newsflash.concordia.ca!sifon!fox From: fox@cs.mcgill.ca (Colin BRADLEY) Subject: Re: NetBSD - routing and slip Message-ID: <1994Jan12.221740.2979@sifon.cc.mcgill.ca> Sender: news@sifon.cc.mcgill.ca Organization: SOCS, McGill University, Montreal, Canada References: <2gv4lc$l0i@darkstar.UCSC.EDU> <1994Jan12.142552.6187@jupiter.sun.csd.unb.ca> <CJJ0w6.ADo@usenet.ucs.indiana.edu> Date: Wed, 12 Jan 1994 22:17:40 GMT Lines: 35 In article <CJJ0w6.ADo@usenet.ucs.indiana.edu> pitts@mimosa.astro.indiana.edu (Jim Pitts) writes: >In article <1994Jan12.142552.6187@jupiter.sun.csd.unb.ca>, >Peter Howlett <b6ps@jupiter.sun.csd.unb.ca> wrote: >>Wailer at the Gates of Dawn (banshee@cats.ucsc.edu) wrote: >> >>: >}But get yoo many of them at once, and ver soo all network >>: >}(or maybe just all slip-related) tasks fail with "no buffer space available" >>: Use a 16550 card or (even better) one of the hayes ESP boards. We run >>: netbsd 0.9 with all std silo drivers at 57600 with 28.8 modems and get >>: peaks of >40K with text files. 26K is common with binary files and other >>: net traffic. The ESP board has in/out 1024k buffers and emulates a 16550 >>: (albeit with huge buffers). >>I can get up to 3000 Chars/Sec on kermit transferring binary (gzip'd) files. >I find this -very- hard to believe. I have basically the same set up except >that I run FreeBSD, the sio device drivers, have a faster CPU and use zmodem >rather than kermit (I also use ftp with ppp). I am luck to get 1.2 K/s transfer >rate with a 16550A. Either I am doing something drastically wrong or you are >doing something drastically correct. I've got a slip line on NetBSD-0.9, and I get about 1.2 K on good days over my 14.4k 16550 arrangement. One thing I have noticed, is that with small files, especially small text files, I get absurdly high tranfer rates. My own feeling on the matter is that there is some sort of problem with the timing mechanism, that gives rise to these figures... Cheers. Colin. -- Colin Bradley fox@binkley.cs.mcgill.ca Proponent of NetBSD, XFree86. Ain'tcha glad ya'ain't no duck?