home *** CD-ROM | disk | FTP | other *** search
- Path: pathway1.pathcom.com!ts2l16
- From: insystem@pathcom.com (Geoffrey Welsh)
- Newsgroups: comp.dcom.modems
- Subject: Re: UART 16650 Serial Card - Not 16550!
- Date: 11 Feb 1996 20:10:24 GMT
- Organization: InSystems Technologies Inc.
- Message-ID: <4flifg$mv5@pathway1.pathcom.com>
- References: <4fakhv$dcq@tst.hk.super.net> <31196797.71426027@uchinews.uchicago.edu> <4fe0dg$i8g@hopi.gate.net> <9602101718.AA26031@pathcom.com> <Pine.A32.3.91.960210121902.92128A-100000@hopi.gate.net>
- NNTP-Posting-Host: ts2l16.pathcom.com
- X-Newsreader: News Xpress Version 1.0 Beta #4
-
- In article <Pine.A32.3.91.960210121902.92128A-100000@hopi.gate.net>,
- doug haire <dhaire@gate.net> wrote:
- >On Sat, 10 Feb 1996, Geoffrey Welsh wrote:
- >> Heck, lock the carrier at 1200 (with the serial port to 115200) and send a
- >> million or so zero bytes...
- >
- >There is a limit to the amount of compression and no need to go that low.
-
- Yes, there's a limit to compression... but 115200:1200 (actually, about 80:1
- when the reduced overhead of the synchronous link is taken int account) isn't
- beyond that limit for a large file of null bytes... unless the manufacturer
- has intentionally limited the string length in their compression
- implementation (which, unfortunately, is too common). I don't know whether
- V.42bis implements any form of RLE and, if so, what effect the BTLZ string
- length limit might have... in any case, MNP5's RLE is almost sure to be the
- most effective compressor of a file of nulls.
-
- Given no artificial string length limits _or_ RLE compression, there is in
- fact a need to go _lower_ (or higher than 115200 bps at the other end) to
- demonstrate the limits of data compression.
-
- >I also wonder if lowering the connect rate too far would have its own
- >effect on the throughput rate.
-
- It's own effect? No... obviously, if you're sending uncompressible data
- you'll get less than 145 CPS; if you're sending the most compressible data
- possible, you should get near serial port speed.
-
- In fact, I chose 1200 bps carrier speed specifically because I believe that
- it's the lowest synchronous carier speed and thus the lowest which will permit
- LAP-M or MNP3 (and thus V.42bis or MNP5 on top of it). It also gives a
- very high serial port to connect speed ratio (96:1), all the better to
- show up those who mindlessly quote 4:1 as the "limit" for V.42bis.
-
- Geoffrey Welsh, Developer, InSystems Technologies Inc.: insystem@pathcom.com
- At home: geoff@zswamp.uucp, [xenitec.on.ca|m2xenix.psg.com]!zswamp!geoff
-