Home . Home .This is an email I sent a friend after he suggested I pick up a buslogic scsi controller from computer geeks discount outlet for $118.00 plus tax, shipping and handling. He made this suggestion when he saw one on their page after hearing of my grief getting an adaptec 2940uw to run reliably under Linux.
Date: Thu, 4 Feb 1999 12:38:22 -0800 (PST) From: Michael Spann <mikes@spann.com> To: brendan miller < > Subject: BT 958
Thanks for the info on computer geeks. I bought one of those controllers. It appears to be, as they said, a system pull. It came in a plain anti-static bag with no manual or software.
I compiled a 2.0.35 kernel with stock BT driver and added 958 tests to my existing bonnie summary table.
As you can see, the performance of the BT958 is comparable to 2940UW on a couple of narrow scsi-2 drives for about half the price. It is supported in the kernel on any linux distribution with a kernel of 2.0 or newer. I had to install the distribution and apply AIC patch using an Adaptec 1522 as the scsi controller because the AIC driver in 2.0.30 kernel on Slackware 3.3 would not boot with the boot device on a 2940uw.
So... my conclusion is: BusLogic controllers are a better bet for Linux systems than Adaptec controllers. They are more compatible with linux kernels, they are less expensive and they perform just as well on narrow drives. I will update you when I add ultra-wide devices to my system.
mike
SCSI controller performance analysis.
Linux 2.0.35 kernel upgrade to slackware 3.3 distribution.
AIC7xxx driver patch 5.1.2
Tulip driver version 0.88 (should have no affect, but is not stock).
quantum empire 540 drive (basic 3 1/2 inch scsi drive)
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
486/100 100 685 97.3 1350 89.4 661 82.7 600 92.9 1403 84.1 26.2 4.1
p120/1522 100 1314 97.0 2078 86.9 939 63.7 1094 89.5 1884 84.4 33.7 2.9
p120/2940 100 2505 93.8 3133 18.9 1413 16.2 2131 85.1 3356 16.4 37.3 2.1
p120/958 100 2524 94.4 3070 18.5 1360 15.9 2125 84.9 3308 15.7 43.2 2.6
p129/958' 100 2474 93.0 3064 18.6 1381 16.1 2195 87.4 3297 16.6 41.6 2.4
486/100 was only run with 1522 as 2940 was unstable in that machine.
958 was run with controller set to scsi-1 setting (5 mbyte/sec).
958' was run with controller set to ultra (20 mbyte/sec). It auto-configured
to fast (10 mbyte/sec). I see little noticeable difference.
seagate ST43400N drive (full size 5 1/4)
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
486/100 100 679 89.8 1349 74.8 641 75.1 576 98.5 1423 98.5 35.8 5.7
p120/2940 100 2495 93.5 2743 16.4 1383 15.6 2157 85.7 3548 18.3 49.9 2.8
p120/958 100 2380 89.0 2581 15.5 1341 15.5 2151 85.7 3318 16.1 50.3 2.8
seagate ST32155W drive (5400 rpm wide)
** using DOS file system may have skewed numbers **
-------Sequential Output-------- ---Sequential Input-- --Random--
-Per Char- --Block--- -Rewrite-- -Per Char- --Block--- --Seeks---
Machine MB K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU K/sec %CPU /sec %CPU
p120/958 100 2220 93.1 2980 28.2 1330 19.1 1735 77.5 3516 38.7 44.5 9.2
/proc/scsi/scsi reports:
Attached devices:
Host: scsi0 Channel: 00 Id: 00 Lun: 00
Vendor: QUANTUM Model: EMPIRE_540S Rev: 110N
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 02 Lun: 00
Vendor: SEAGATE Model: ST43400N Rev: 1022
Type: Direct-Access ANSI SCSI revision: 02
Host: scsi0 Channel: 00 Id: 05 Lun: 00
Vendor: TOSHIBA Model: CD-ROM XM-4101TA Rev: 2483
Type: CD-ROM ANSI SCSI revision: 02
Copyright © 2001-2016 by Fractal
Hosting by ziandra.net