home *** CD-ROM | disk | FTP | other *** search
- Path: news.gate.net!not-for-mail
- From: dhaire@gate.net (doug haire)
- Newsgroups: comp.dcom.modems
- Subject: Re: Is USR going to support 42bis+ on future courier upgrades?
- Date: 29 Mar 1996 01:59:32 -0500
- Organization: CyberGate, Inc.
- Message-ID: <4jg1ok$1mti@seminole.gate.net>
- References: <4j2fv1$8kf@nnrp1.news.primenet.com> <4j2iun$a3t@newsbf02.news.aol.com> <4j3r1f$1tc4@seminole.gate.net> <4j468j$gg1@drencrom.insync.net> <4j68or$nug@navajo.gate.net> <4j7iai$qcc@drencrom.insync.net> <4jbasq$102o@navajo.gate.net> <4jd6e0$kbo@dr
- NNTP-Posting-Host: seminole.gate.net
- encrom.insync.net>:
- X-Newsreader: TIN [UNIX 1.3 950824BETA PL0]
-
- Greg Bretting (bretting@insync.net) wrote:
- : On 27 Mar 1996 07:04:42 -0500, dhaire@gate.net (doug haire) wrote:
- :
- : [snip]
- : >: What does _that_ have to do with anything? You said above that you were
- : >: using a null modem cable, and if that's the case then my LapLink example is
- : >: perfectly valid because it's essentially the same setup except for
- : >: different application software being used.
- : >
- : >Bingo! The same thing EXCEPT for the software involved. You are using
- : >software designed to dedicate all of the CPU to the task. That's why I
- : >said apples and oranges.
- :
- : Well, let's see - LapLink is designed to transfer data to and from a serial
- : connection, perform error checking, and doing disk I/O. How does this
- : substantially differ from the comm software you were using?
-
- Because the comm program is not dedicating all processor time to the
- transfer.
-
- : >If you don't understand the difference between using LapLink and running
- : >a comm program protocol on a null modem cable at full speed then you
- : >aren't grasping the concept.
- :
- : See above...
-
- Like I said, you aren't grasping the concept.
-
- : >: >: >When the common computer software platform is capable of handling 115200
- : >: >: >properly perhaps we can then consider the 230k UART speed.
- : >: >:
- : >: >: Well, DOS 6.2 is pretty common, and I know that I've been able to (almost)
- : >: >: saturate the port at 115200 without any errors. Here's a log from one such
- : >: >: test using QModem Pro for DOS, a Courier V.34 external, and plain 16550
- : >: >: UART:
- : >: >:
- : >: >: 22:40:52 09-14-95 Online Timer Started
- : >: >: 22:42:00 09-14-95 Download File(s). Protocol : Zmodem
- : >: >: 22:42:01 09-14-95 ++ File 1MEGTEST.RUN
- : >: >: 22:43:34 09-14-95 ++ End of file
- : >: >: 22:43:34 09-14-95 ++ Chars Per Second : 11272
- : >: >: 22:43:34 09-14-95 ++ Effective Percent : 0%
- : >: >: 22:43:40 09-14-95 Elapsed Online 00:02:48
- : >: >
- : >: >Sure and I have also. In fact, I posted several articles showing this on
- : >: >transfers between computers over phone lines and modems. That's not, of
- : >: >course, what I was talking about here and it has little to do with my point.
- :
- : But the above _does_ illustrate a case where a common DOS platform
- : functions well at a 115200bps DTE rate, which is a possibility you seemd to
- : take issue with in your "When the common computer software platform is
- : capable of handling 115200 properly perhaps we can then consider the 230k
- : UART speed." comment.
-
- The port speed is not truly relevent because the modem is controlling the
- flow to the port. The modem is effectively controlling the port, the CPU
- is not.
-
- : >: Then I obviously am missing your point; I interpreted it to be that you
- : >: felt 115200 DTE rates were unworkable on DOS platforms, let alone 230,400,
- : >: and that discussion of DTE rates > 115200 were pointless since 115200
- : >: didn't work very well on most platforms. I then provide two examples, one
- : >: using a null modem connection and another a dial-up session with an
- : >: external modem, that seem to contradict what you are saying.
- : >
- : >No, I offered that 115200 DTE rates were more than adequate for current
- : >operating system platforms in the real world. That having a port set to
- : >230k (and a modem that would accept that) is unnecessary and simply
- : >advertising hype.
- : >
- : >You offered a link using a specialized piece of software as a counter to
- : >this. Different game.
- :
- : To begin with, I wasn't arguing that 230.4Kbps DTE rates made sense - I was
- : addressing your apparent (to me, anyway) claim that DOS-based platforms
- : weren't up to the task of supporting 115.2Kbps port speeds. Let's also not
- : forget the example I offered involving a fairly common DOS based comm app
- : (Qmodem Pro, in case you forgot).
-
- And I stand by what I said. I would like to see someone with a Hayes to
- Hayes connection at 28800/28800 with ports open at 230k at each end send
- a file and report the results so we can see what happens.
-
- : >: Not only that, but I know for a fact that I can connect two modems to _one_
- : >: DOS machine and pass data full-duplex (simultaneous send and receive of the
- : >: same file) between the ports at 115200 without errors. This is normally
- : >: how I run throughput testing on modems, using SoftArt's HowFast v1.65
- : >: testing software running under DOS, on a wide variety of machines.
- : >
- : >How do you connect two modems and pass data at 115200 between them?
- : >Answer: you don't. You pass data from a port to a modem on one end and
- : >from a modem to a port on the other at that speed; between the modems,
- : >you are limited to the DCE rate of the modems.
- :
- : What, did we forget entirely about the subject of this thread - namely,
- : compression? The DCE-DCE rate isn't the important part, it's the rate at
- : which the modem presents data _after compression_ to the DTE - which,
- : depending on the file, can be substantially higher than the DCE rate and
- : can under certain circumstances approach the limit of the port speed.
- : Regardless of whether we are talking about a NMC or modem call with data
- : compression enabled, the examples that have been discussed involve data
- : rates being presented to the DTE at or near it's limit (as configured) of
- : 115.2Kbps. Unless I'm missing something (always a possibility I'm willing
- : to accept), there isn't much difference between a NMC connection that
- : presents data to the DTE at 115200 and a modem with data compression
- : enabled operating at a sufficient speed and passing highly compressible
- : data to the DTE at speeds very closely approaching the limit of the port.
-
- The rate the data is sent to the port is limited by the modem's ability
- to efficiently decompress the data brought in from the line. In essence,
- you have the modem processor controlling the port of the receiver.
-
- : Not only that, but keep in mind that in the above example, only _one_ CPU
- : and DOS environment is servicing the interrupt load for _both_ ports.
-
- Only one port: the receiving unit. And, even there, it is subject to the
- processing power of the modem to send to the port. The port is not fully
- controlled by the CPU on the computer in that situation. Remove the
- modems and you have one CPU (the receiver) trying to control both ports
- and it does a poor job.
-
- : >: Doug, I didn't just start doing this stuff yesterday - I know for a fact
- : >: that MS-DOS is perfectly capable of supporting 115200 DTE rates, and I've
- : >: demonstrated that on dozens of machines and literally hundreds of modems
- : >: during the course of testing these things, using all sorts of software, and
- : >: I know that it works.
- : >
- : >Good, then you can simply ignore what I said and hit Enter. Or you can
- : >discuss this without bringing in specialized software designed to
- : >overcome DOS's limitations.
- :
- : Well, okay then, let's deal with the examples I've posed so far that aren't
- : "specialized" software - namely, the QModem Pro log I posted previously and
- : the Procomm Plus/Win 2.11 screenshot I uploaded to alt.binaries.misc (which
- : I posted yesterday, have you seen it?)- both of which demonstrate
- : throughput on a DOS platform in excess of 11,000 cps. Whatever the
- : limitations of DOS may be (and I'm not saying there aren't any), it doesn't
- : take exotic programming to overcome them - it would appear to me that just
- : about any competently written and properly configured comm app will do just
- : fine.
-
- Just do what I did. Link two computers up via a null modem cable and run
- the comm program of your choice in each computer with the ports locked at
- 115200 and then come back with the results, ok?
-
- That's all I ask.
-
-