[534] in linux-scsi channel archive

home help back first fref pref prev next nref lref last post

No subject found in mail header

daemon@ATHENA.MIT.EDU (Leonard N. Zubkoff)
Sun Aug 20 08:20:38 1995

Date: Sat, 19 Aug 1995 23:14:44 -0700
From: "Leonard N. Zubkoff" <lnz@dandelion.com>
To: redd@pat.mdc.com
Cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <199508192337.SAA13317@pat.mdc.com> (message from Jarrett Redd on Sat, 19 Aug 1995 18:37:21 -0500)

  Date: Sat, 19 Aug 1995 18:37:21 -0500
  From: Jarrett Redd <redd@pat.mdc.com>

  Original poster (J. Redd) responds:

  Thanks to everyone for your help.  I just have one more question about all
  this.  Suppose I forget about this "pseudo-EISA" stuff and buy an ISA SCSI
  controller?  Will the HiNT chipset still affect it?  And if not and it works
  with Linux, what kind of performance hit am I likely to take?  Finally, I've
  read some good stuff about the BusLogic controllers.  What about getting a
  caching, ISA SCSI-II BusLogic card?  I already have some available memory.

BusLogic does not make any caching SCSI controllers.  In any event, since the
bottleneck would be the ISA bus, you won't be gaining anything by putting a
cache on the other side of a slow ISA link.

  I came up with this as a substitute for replacing my entire motherboard.
  What does everyone think?  Thanks again for your help!

I have to admit I've forgotten your complete system configuration, but I think
an ISA SCSI card is a bad idea.  For example, in my testing a 545C or 540CF
only achieved less than 40% of the throughput that a 445C VLB card was capable
of.  The *only* difference in the two configurations was the SCSI controller
swap.  Better to get a decent 486 VLB motherboard and a 445C, or a proper EISA
motherboard.

		Leonard

home help back first fref pref prev next nref lref last post