Quoting Mark Hahn <hahn@coffee.psychology.mcmaster.ca>:
> note that I did NOT reply to the list.
>
> > > > We're running the channel at 160, and are getting the following (over
> > > > and over again):
> > >
> > > not to be obtuse, but if running a cable at 80 MHz doesn't work,
> > > reporting data corruption, but 40 MHz does, then shouldn't you
> > > suspect that the cable is too long, has stubs, wrong impedence, etc?
> >
> > We're running LVD, and LVD should manage 20 meters. And the terminator
> > is also a LVD terminator etc...
>
> stub? etc.
This is a brand new LVD kable with brand new disk and a brand new
terminator, with a brand new SCSI card...
We payed ~ $40US for the cable alone (not that the price is the main
issue, but one could think that a cable that expensive should be ok).
> the point is that you're assuming, for some reason, that the driver
> is spuriously reporting a parity error. given adaptec's history of
> asinine hardware requiring bizarre workarounds, this is conceivable,
> but certainly not a sure thing.
I'm not assuming anything, I just find it strange that brand new
hardware should be at fault here. And this is not the first time I'm
getting problem with SCSI under Linux.
> in any case, there is no possible performance difference, since no
> shipping disk sustains better than around 40 MB/s.
We have 4 IBM 36LZX 9Gb 10k rpm/sec, and doing one 'dd if=/dev/zero
of=file count=1000000' per disk. This file took 14 sec to complete on
each disk (running simultaneous) which gives ~ 62Mb/sec... Getting the
controller up to 160 MHz should double the speed (more or less any
way).
-- South Africa cryptographic Legion of Doom AK-47 SDI [Hello to all my fans in domestic surveillance] NORAD Nazi class struggle tritium counter-intelligence terrorist domestic disruption genetic ammonium- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed May 31 2000 - 21:00:19 EST