clock drift with 2x Promise Ultra133 TX2 (PDC 20269) (again)

clock drift with 2x Promise Ultra133 TX2 (PDC 20269) (again)

[Oskar Liljeblad]

Problem:
I have two Promise Ultra133 TX2 (PDC 20269) cards, each with two
Maxtor UDMA133 hard drives connected. The system also has another
hard drive connected to the motherboard. A total of five drives.

When I concurrently read from two hard drives connected to
different Promise cards, there is heavy software clock drift.
The software clock runs too fast - approx. a few seconds per
minute during this i/o. To reproduce:

  hwclock --hctosys
  hwclock --show
  date
  dd if=/dev/hde of=/dev/null bs=1M count=7200 &
  dd if=/dev/hdi of=/dev/null bs=1M count=7200 &
  wait
  wait
  hwclock --show
  date

This doesn't happen if I read from two drives connected to the
same card, or one drive connected to the motherboard and one
connected to a Promise card.

I've removed all other PCI devices in the system - problem remains.
I tried different motherboards:

  Asus A7V266-C with AMD AthlonXP 2000+ (i686) - problem occurs
  Asus A7V600-X with AMD Sempron 2600+  (i686) - problem occurs
  Abit KV8 Pro  with AMD Athlon64 3000+ (i686) - problem does not occur!

So what makes the Abit motherboard different?

Regards,

Oskar Liljeblad (oskar@osk.mine.nu)

Re: clock drift with 2x Promise Ultra133 TX2 (PDC 20269) (again)

[Drew Winstel]

On Tue, 2005-05-24 at 18:53 +0200, Oskar Liljeblad wrote:

>   Asus A7V266-C with AMD AthlonXP 2000+ (i686) - problem occurs
>   Asus A7V600-X with AMD Sempron 2600+  (i686) - problem occurs
>   Abit KV8 Pro  with AMD Athlon64 3000+ (i686) - problem does not occur!
> 
> So what makes the Abit motherboard different?
> 

Are you running the AMD64 machine on a 32- or 64-bit distribution?  

That was just the first thought that came into my head.

Drew