home *** CD-ROM | disk | FTP | other *** search
- From: Kay Roemer <roemer@informatik.uni-frankfurt.de>
- Posted-Date: Mon, 25 Jul 94 11:19:15 MESZ
- Received-Date: Mon, 25 Jul 94 11:19:15 +0200
- Message-Id: <9407250919.AA19656@hera.rbi.informatik.uni-frankfurt.de>
- Subject: Re: MiNTlib (I think I'm on pl43): Bug in sprintf().
- To: nox@jelal.north.de (Juergen Lock)
- Date: Mon, 25 Jul 94 11:19:15 MESZ
- In-Reply-To: <9407172345.AA00280@jelal.north.de>; from "Juergen Lock" at Jul 18, 94 1:45 am
- Mailer: Elm [revision: 70.85]
-
- > > Well, I'm not taking of 30 characters, but about a screen full of them, ie
- > > about 80x25=2000! When I use modm0dev which polls at 5 times per second
- > > using a separate porcess I get much smaller hunks, about 500.
- >
- > thats on a 68000 right? i guess you just found the `feature' that a
- > RAW tty read is slower copying stuff out of the receiver buffer than
- > data comes in at 19200 bps... sounds unbelievable hmm? ;) well, try this!
- > (relative to 10h6)
-
- Your Bios IO / raw tty write / pipe speedup patches are actually a great deal!
-
- 1) Serial IO is much faster now, data comes in small hunks.
- 2) IO to the console is pretty fast. Even with fasttext or vcon
- scrolling upwards is substantially faster.
- 3) Pty/pipe througput is much higher. Try "rlogin localhost;
- cat /usr/man/cat1/perl.1" with and without the patch and
- you get the difference!
-
- Thanks,
- Kay.
-
-