Very Computer

Hiyas.. I got a P5-166 linux box with an ASUS MB, BusLogic 958 UltraWide
SCSi controller, 2 IBM DORS ULTRA WIDE SCSi HDs (68pin). And when I boot
up it shows the HD's set to do 20MB/sec sync... These HDs are suppose to
do 40MB/sec. Am I missing something? I set the bios for the Buslogic to
40mb/sync on both HDs id1 and id2. Is there something I have tweak with in
teh driver to get it to do so ??

Both HDs are the same scsi card and bus... Can anyone help or have
suggestions??

  Hiyas.. I got a P5-166 linux box with an ASUS MB, BusLogic 958 UltraWide
  SCSi controller, 2 IBM DORS ULTRA WIDE SCSi HDs (68pin). And when I boot
  up it shows the HD's set to do 20MB/sec sync... These HDs are suppose to
  do 40MB/sec. Am I missing something? I set the bios for the Buslogic to
  40mb/sync on both HDs id1 and id2. Is there something I have tweak with in
  teh driver to get it to do so ??

  Both HDs are the same scsi card and bus... Can anyone help or have
  suggestions??

If what you are seeing is:

scsi0:   Target 0: Synchronous at 20.0 mega-transfers/second, offset 15

then you are getting the full performance unless you've explicitly disabled
wide transfers.  Note that the messages says "mega-transfers/second", not
"megabytes/second".  That's because the driver can only determine the
transfer period, but not whether wide negotiation has taken place.  For 16
bit transfers, that's two bytes/transfer, or 40.0 megabytes/second.

                Leonard

:   Hiyas.. I got a P5-166 linux box with an ASUS MB, BusLogic 958 UltraWide
:   SCSi controller, 2 IBM DORS ULTRA WIDE SCSi HDs (68pin). And when I boot
:   up it shows the HD's set to do 20MB/sec sync... These HDs are suppose to
:   do 40MB/sec. Am I missing something? I set the bios for the Buslogic to
:   40mb/sync on both HDs id1 and id2. Is there something I have tweak with in
:   teh driver to get it to do so ??

:   Both HDs are the same scsi card and bus... Can anyone help or have
:   suggestions??

: If what you are seeing is:

: scsi0:   Target 0: Synchronous at 20.0 mega-transfers/second, offset 15

: then you are getting the full performance unless you've explicitly disabled
: wide transfers.

Also note that the "40" figure you're quoting is the bus channel MAX
BANDWIDTH, _not_ the actualy data transfer speed (excluding bursts).

I seriously doubt that the ibm drive (or any current 'consumer level' drive)
can reach that kind of speed.  The actual xfer rate is based on the 'internal
xfer rate' NOT the max channel bandwidth.  See the difference?

(not to complain about the buslogic stuff - on the contrary, I'm a BIG
buslogic fan) ;-)

--

Legal Warning: Anyone sending me unsolicited/commercial email WILL be charged
a $100 proof-reading fee.  Do NOT send junk email to me - consider this an
official notice:

"By US Code Title 47, Sec.227(a)(2)(B), a computer/modem/printer meets the
 definition of a telephone fax machine.  By Sec.227(b)(1)(C), it is unlawful
 to send any unsolicited adverti*t to such equipment.  By Sec.227(b)(3)(C),
 a violation of the aforementioned Section is punishable by action to recover
 actual monetary loss, or $500, whichever is greater, for each violation."

: Hiyas.. I got a P5-166 linux box with an ASUS MB, BusLogic 958 UltraWide
: SCSi controller, 2 IBM DORS ULTRA WIDE SCSi HDs (68pin). And when I boot
: up it shows the HD's set to do 20MB/sec sync... These HDs are suppose to
: do 40MB/sec. Am I missing something? I set the bios for the Buslogic to
: 40mb/sync on both HDs id1 and id2. Is there something I have tweak with in
: teh driver to get it to do so ??

: Both HDs are the same scsi card and bus... Can anyone help or have
: suggestions??

  As pointed out by Leonard once, the boot up message will say 20.0
  mega-transfers/second, not 20.0 megabytes/second. When wide transfer
  is negotiated, each transfer happens with 16 bits not 8 bits. So you
  are getting 2 times 20.0 mega-transfers/second _theoretically_.
  I would be interested to know if the actual transfer really hopefully
  exceeds 10MB/s or not with your 5400rpm DORS under P5-166.

  J.

--
Joonwoo Nam
MAGNUMS(MAssachusetts Group for NUMerical analysis of Semiconductors)
Dept. of Electrical and Computer Engineering, UMASS at Amherst

http://khushi.ecs.umass.edu/~nam