Newsgroups: comp.periphs.scsi
Path: lth.se!sunic!mcsun!Germany.EU.net!mpifr-bonn.mpg.de!speckled.mpifr-bonn.mpg.de!mlelstv
From: mlelstv@speckled.mpifr-bonn.mpg.de (Michael van Elst)
Subject: Re: EXB-8200 transfer rate
Message-ID: <1992Aug25.183926.13763@mpifr-bonn.mpg.de>
Sender: news@mpifr-bonn.mpg.de
Nntp-Posting-Host: speckled
Organization: Max-Planck-Institut f"ur Radioastronomie
References: <1992Aug19.212240.6537@scf.loral.com> <1992Aug24.182613.2016@kentrox.uucp>
Date: Tue, 25 Aug 1992 18:39:26 GMT
Lines: 27

In <1992Aug24.182613.2016@kentrox.uucp> ronc@kentrox.uucp (Ron Christian) writes:
>(1)  Write speed suffers dramatically due to increased amount of head
>repositioning.  (This may be what you're seeing.)  In general write speed
>to any streaming device will suffer with small (512 byte) blocks, but on
>the 8mm the slow-down is incredible.

Well, no. The head positioning comes from a too low transfer rate. The
Exabyte (not 8mm technology per se) needs several milliseconds per
command compared to some hundred microsecond to transfer a 512byte block
(64 micros at 4MB/s synchronous). This means that less than 1% of SCSI
bandwidth is used for data transfer :-(

>(3)  If the info you put on the tape (a label, for instance) is less than
>512 bytes (only one block is written between file marks), subsequent tape
>operations (a Unix mt command, for instance) may or may not find the
>file on the tape.

Hmm. Seems more like a failure of the UNIX driver. I'm using direct SCSI
commands for the Exabyte (on some special hardware) and have not seen
such a problem.

Regards,
-- 
Michael van Elst
UUCP:     universe!local-cluster!milky-way!sol!earth!uunet!unido!mpirbn!p554mve
Internet: p554mve@mpirbn.mpifr-bonn.mpg.de
                                "A potential Snark may lurk in every tree."
