| Title: | Atari ST, TT, & Falcon |
| Notice: | Please read note 1.0 and its replies before posting! |
| Moderator: | FUNYET::ANDERSON |
| Created: | Mon Apr 04 1988 |
| Last Modified: | Tue May 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 1433 |
| Total number of notes: | 10312 |
Hello all,
Please can anyone suggest the fastest method of moving data
to and from the ST to our VAX here.
I use very reliably, Kermit 32 on the VAx and Kermit on the ST,
but its so SLOOOOOOWWWW!.
Our group is very fortunate having 2400 baud error correcting
modems, V22 bis, available. So logically, it would be faster if the
Kermits used longer packets, wouldnt it? I dont know the ratio of
the time spent in protocol handling to the time spent in transmitting
data, what is the best throughput I could hope for at 2400 baud both
ways?
I think Kermit32 on the VAX is set to a very short packet 80 odd, ??.
I dont know of any way of altering it, do you?
Is there any other software that works faster, I have Flash for
the ST, this has an Xmodem implementation that can use a block size
of 1K, Anyone know of a VAx xmodem that can do 1K blocks?
The reason I want speed is to cut the cost of line time down, in
the UK, even in off-peak periods, BT charge an arm and two legs for
connect time, and I am getting into postscript files which even
when Arc'd are HUMUNGUS. >17K for an A4 page!
Cheers
Pat Keane.
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 62.1 | Use WHACK and STRANSF | SKITZD::MESSENGER | An Index of Metals | Tue Apr 26 1988 12:55 | 3 |
Using WHACK and STRANSF will achieve results very close to the
theoretical maximum transfer speed.
- HBM
| |||||
| 62.2 | Some details about Whack/Stransf (see topic 3) | PRNSYS::LOMICKAJ | Jeff Lomicka | Tue Apr 26 1988 13:28 | 30 |
Whack/Stransf is explicitly designed to work well on split-speed error correcting modems. Translated, this means that the STRANSF error recovery is weak, but the protocol is very streamlined. (You did mention that you had error correcting modems.) (Note to current users - you don't NEED error correcting modems. Whack does still detect the errors and retry when they occur, it's just not very good at it. On really bad lines it can hang or get confused.) Using a 2400 baud line, Whack/Stransf consistently achieves 200bytes/second, which is 80% of the theoretical maximum. STRANSF returns data transfer statistics when the transfer is complete, so you can compute this figure yourself. You should ARC your files first. ARC can get 50% out of many text files, like PostScript files. Although it doesn't apply to you, if you have access to a U.S. Robotics 9600 baud modem, WHACK/STRANSF is ideal. This particular modem decides which way is getting more traffic, and assigns a 9600 baud channel to one direction, and a 300 baud channel to the other. KERMIT and XMODEM include ACK messages, which get sent at 300 baud, and as a result reduce the transfer rate to less that what you would get with a 2400 baud true full duplex modem. STRANSF uses XON/XOFF synchronization on an as-needed basis, and does't send any ACK messages, only NAK messages. As a result, the full 9600 baud is dedicated to data transfer, even when the return path is 300 baud. You may have to increase the TYPEAHEAD VMS sysgen parameter when uploading using this configuraiton, because the XOFF codes arrive slowly. Of course, if you want FAST data transfer, you should put a 3� floppy drive on your VAX. | |||||
| 62.3 | Kermit's robust but slow. | BOLT::BAILEY | Steph Bailey | Tue Apr 26 1988 13:59 | 8 |
The maximum size of a Kermit packet is 96 bytes, I believe, so the
80 that the VAX selects doesn't really do much worse than the best
you can do.
For non-VMS, non-DEC internal systems, my recommendation is some
variant of XMODEM, which is a much more streamlined protocol.
Steph
| |||||
| 62.4 | Use Whack/Stransf.... | ASPEN2::BOIKO | Tue Apr 26 1988 16:21 | 3 | |
I would use STRANSF....it is fast!!!
-mike-
| |||||
| 62.5 | RDGENG::KEANE | Wed Apr 27 1988 06:46 | 17 | ||
Many thanks for your valuable suggestions (.1->.4). I will give
stransf another go. When I first tried to use it I had a lot of
trouble. Perhaps it was finger trouble on my part, but the VAx end
kept dropping out of STRANSF back to DCL, which blew sys$input's
mind!
Jeff:- Regarding the suggestion about putting a 3.5 inch drive on
a VAX, thats no problem, the RX33 on a VAxstation 2000 is a standard
TEAC 80T D/S, so if I unplug the RX33, a 3.5 should plug in OK, The
question is what do we use under VMS to read/write to it ??.
Thanks again.
Pat K
| |||||
| 62.6 | Kermit, whats KERMIT? | RDGENG::KEANE | Thu May 05 1988 08:04 | 16 | |
After a great deal of fat finger trouble on my part (and poor modem
documentation), I have got STRANSF working.
It is GREAT!! I can move enormous quantities of data around
at very acceptable rates, approx 200 Bytes/sec on a 2400/2400 link.
Anyone who has trouble:- Check your XON/XOFF Flow control, right
through the link, including modems. SET SSU ENA and Hostsync and
ttsync on. GO FOR IT !!
Cheers
Pat K.
| |||||
| 62.7 | STRANSF is fast... | ASPEN2::BOIKO | Thu May 05 1988 23:34 | 3 | |
Pat is right, it is great...and very fast too...
-mike-
| |||||