home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.3b1
- Path: sparky!uunet!ceilidh!dnichols
- From: dnichols@d-and-d.com (DoN. Nichols)
- Subject: Re: Getting xfer stats from HDB
- Message-ID: <1993Jan23.211312.8970@d-and-d.com>
- Sender: usenet@d-and-d.com (Usenet)
- Nntp-Posting-Host: shindig
- Organization: D and D Data, Vienna VA
- References: <1993Jan22.025514.2843@toybox.raleigh.nc.us>
- Date: Sat, 23 Jan 1993 21:13:12 GMT
- Lines: 37
-
- In article <1993Jan22.025514.2843@toybox.raleigh.nc.us> hruska@toybox.raleigh.nc.us (Bob Hruska) writes:
- >I have been running HDB with a Supra FAXModem (14400bps) for
- >a couple months now, and now I'm curious what kind of transfer rates
- >I'm getting.
- >
- >I have seen a few postings here that quoted output from "uutraf" but
- >I don't know what this is. Is there some sort of accounting that
- >HDB does that I can use? What/where is uutraf?
-
- Uutraf is a public-domain program out there somewhere, perhaps on
- the BSD source - I haven't looked for it yet. However, you can check on the
- transfer speeds by looking in the file "/usr/spool/uucp/.Admin/xferstats".
- Beware that the transfer rates shown for short files are close to
- meaningless, since you have the overhead of establishing the transfer for
- that file, and if your modem is capable of uucp spoofing (like the
- Telebits), and has a large buffer, everything goes into the buffer very
- quickly, and the granularity of the time measurement under SysV is rather
- coarse, so you get strange numbers. A flie larger than 10k is probably a
- reasonable measure of transfer speed. (You don't mention whether the 14.4k
- modem is interfaced at 9600 or 19200 baud.) If the system is doing
- something else which generates a lot of interrupts, such as ethernet
- transfers, you can wind up with slower effective transfer rates with a baud
- rate of 19.2k than at 9.6k, since you have more overruns and failed blocks
- which need to be retried. On a lightly loaded 3B1, 19.2k is probably
- reasonable with standard 'g' protocol with 3 windows. Any attempt to make
- it more efficient with longer packets or more windows seems to be
- counterproductive, based on the general experience on this newsgroup, though
- you will find a few for whom there are no apparent troubles.
-
- Good Luck
- DoN.
-
- --
- Email: <dnichols@d-and-d.com> | ...!uunet!ceilidh!dnichols
- <dnichols@ceilidh.beartrack.com>
- Donald Nichols (DoN.) | Voice (Days): (703) 704-2280 (Eves): (703) 938-4564
- --- Black Holes are where God is dividing by zero ---
-