Xref: utzoo comp.dcom.modems:3094 comp.unix.xenix:4277 Path: utzoo!attcan!uunet!lll-winken!lll-lcc!ames!mailrus!tut.cis.ohio-state.edu!rutgers!psuvax1!vu-vlsi!snark!eric From: eric@snark.UUCP (Eric S. Raymond) Newsgroups: comp.dcom.modems,comp.unix.xenix Subject: Re: Slow Throughput with TB+ Modem Message-ID: Date: 25 Dec 88 05:45:01 GMT References: <112@avatar.uucp> <398@eda.com> <1439@neoucom.uucp> <672@vector.uucp> Organization: First Zenarchist International Lines: 23 In article <672@vector.uucp>, chip@vector.UUCP (Chip Rosenthal) writes: > In article <1439@neoucom.UUCP> wtm@neoucom.UUCP (Bill Mayhew) writes: > >I've been surprised at the gernally crummy serial port performance > >of most IBM/Intel clone architecture machines exhibit. Even AT > >class machines with the 16450s don't do very well at speeds greater > >than 4800 bps. > > Correction -- even 386 class machines can't do it. > When I was using a standard, dumb serial port on the XENIX box, I could > never run faster than 4800bps. Even then it would drop characters if the > system was heavily loaded. (I looked at the PC BIOS, and it looks like > hardware flow control is supported, but I only have a 3-wire connection. > XON/XOFF flow control is not supported.) I run a 9600bps line to a 'blazer off a 16MHz 386 box using a generic dumb serial card. It works like a champ. This is under stock Microport 3.0e, using RTS/CTS flow control. I conjecture that either a) XENIX's serial-port driver is brain-damaged (thresholds set wrong maybe, or too little c-list space) , or b) your serial card really sucked. The latter seems marginally more likely. -- Eric S. Raymond (the mad mastermind of TMN-Netnews) Email: eric@snark.uu.net CompuServe: [72037,2306] Post: 22 S. Warren Avenue, Malvern, PA 19355 Phone: (215)-296-5718