2 PCI IDE controller cards in 1 PC

  • Thread starter Thread starter Richard Lowen
  • Start date Start date
Richard said:
The 2 cards are made by SIIG and SYBA. The SYBA tech rep
said that since they have the same chipset, they can't be distinguished
apart. The SIIG tech rep thinks they'd probably be distinguishable if
they had different manufacturers (he may have been thinking "chipsets").
Can you think of some way to make the motherboard BIOS
distinguish them although their on-card BIOS's are identical?

TIA,
Rick Lowen
Click to expand...

FWIW, I've had two SIIG cards in a machine, and they worked ok. But I
wasn't using Windows, and it would be an incredible coincidence if I
was using the same motherboard as you have. So there were both hardware
and software differences between what I did and what you're trying to
do.
 
J. Clarke said:
It just occurred to me though--on many motherboards two PCI slots are
physically wired to the same interrupt--which two you'll have to check the
docs for the motherboard--if you've got the two boards in the two slots
that share an interrupt that could be your problem.
Click to expand...

I think in many boards (certainly those that were around a couple of
years ago) the first and last PCI slots shared the same interrupt.

If I recall correctly, it was with a VIA chipset that most problems
occurred.


Odie
 
CJT said:
Rod Speed wrote
Click to expand...
Some people apparently want to play configuration cop.
Click to expand...

No one is saying anything about 'configuration cop',
just considering whether 2 PCI addon cards is the
best approach when there is clearly some difficulty
with getting that config to work.

Even now that he has specified that he wants 4 IDE
channels, it isnt clear that that is going to be that viable
even if he can get it to work since it may well be limited
by the PCI bus performance even if he can.
I don't.
Click to expand...

Your problem if you cant manage to grasp that what matters
is what is trying to be achieved and what may well be an
easier approach to getting the end result he wants.
 
Richard Lowen said:
Thanks, again. I have a reason for wanting one hard drive per
IDE channel.
Click to expand...

Why be coy about what the reason is for that ?

It may be that even if you can get 4 IDE channels on 2 addon PCI
cards working, it wont be viable if its limited by the PCI bus performance.

Not all configs would be, particularly if you want completely
independent seeking on the 4 channels, and dont actually
need to move a lot of data over the PCI bus, but even if
that is the case, why do the 4 IDE channels have to be on
addon PCI cards ? It would be easier to get a total of 4 IDE
channels with two on the motherboard and two on a single
addon PCI card and the PCI bus would be less of limit.

Certainly the Promise cards are notorious for having a
problem with running two cards, and that appears to be due
to a badly designed card bios which cant handle that properly.
 
Rod Speed said:
You sure you wont just get limited by the PCI bus even if
you do manage to get 4 channels on 2 PCI cards working ?
Click to expand...


There would only be one hard drive per channel, and only 3 or 4 ATA133
hard drives would be running, 2 in use at most at any time, the other hard
drives being idle. No other devices connected to the PCI bus would be
active during that time. Would that load down the PCI bus?

TIA,
Rick Lowen
 
Rod Speed said:
No one is saying anything about 'configuration cop',
just considering whether 2 PCI addon cards is the
best approach when there is clearly some difficulty
with getting that config to work.

Even now that he has specified that he wants 4 IDE
channels, it isnt clear that that is going to be that viable
even if he can get it to work since it may well be limited
by the PCI bus performance even if he can.


Your problem if you cant manage to grasp that what matters
is what is trying to be achieved and what may well be an
easier approach to getting the end result he wants.
Click to expand...


My problem is getting the number of IDE channels. My motherboard
has ATA33 speeds, and my ATAPI devices use those ports. All the
ATA133 ports are provided by PCI add-in cards. The problem
doesn't involve throughput, it involves getting more channels.

T,
Rick Lowen
 
There would only be one hard drive per channel, and only 3 or 4 ATA133 hard
drives would be running, 2 in use at most at any time, the other hard drives
being idle. No other devices connected to the PCI bus would be
active during that time. Would that load down the PCI bus?
Click to expand...

Really depends on exactly what you are
doing on those 2 drives active at one time.

Maybe, maybe not.

It would be better to use the two channels on the motherboard and
have two more on a PCI card, less likely to be limited by the PCI bus.
 
Richard Lowen said:
My problem is getting the number of IDE channels. My motherboard
has ATA33 speeds, and my ATAPI devices use those ports. All the
ATA133 ports are provided by PCI add-in cards. The problem
doesn't involve throughput, it involves getting more channels.
Click to expand...

You keep being coy about why you need more channels tho.
It may be that even a single addon PCI card with say 4 channels
wont actually be that viable because of the limitations imposed
by having everything going thru just one PCI bus. Hard do say
unless you say clearly what you are trying to do that needs
a separate channel per drive. If its just so the seeks on one
drive have no impact on the other drive in use at the same time,
a 4 channel single PCI card may well be fine. If there is a
signficant volume of data moving tho, it may not be that viable.
Particularly as the PCI bus may well not be very high bandwidth
in a motherboard that only has ATA33 speeds on the IDE channels.

Basically it might well be more viable to replace the motherboard
than to buy one of those more expensive 4 channel PCI IDE cards.
 
Pete said:
@newsread2.news.pas.earthlink.net:


My 2 Promise Ultra 100 TX2 cards do. I'd suppose the 133's would, too.
Click to expand...

It seems strange to me that two identical Promise cards would work
together, and a SIIG doesn't work with a SYBA. I get the nagging
feeling that there might be a way to make the SIIG and SYBA cards
work together without major surgery.

T,
Rick Lowen
 
J. Clarke said:
If it's not doing it automatically then there's generally no way to force
it. Some motherboards in the past allowed IRQs to be manually assigned to
specific slots but I haven't seen one of those in years.

It just occurred to me though--on many motherboards two PCI slots are
physically wired to the same interrupt--which two you'll have to check the
docs for the motherboard--if you've got the two boards in the two slots
that share an interrupt that could be your problem.


In a cheap board, Promise and High Point both make boards that work. The
High Point RocketRAID 454 should do what you want for about 90 bucks--while
it supports RAID you don't _have_ to configure a RAID on it.

The Promise TX4000 is a bit more expensive at about $130.

If you want to go upmarket some, the 3Ware 7506-4LP goes for about 200
bucks, or if you need more channels you can go as high as 12 with the
7506-12 for about $550.
Click to expand...


I heard from Highpoint Support. They said they "know of no conflicts"
between their cards and the SIIG and SYBA cards and that they "should
work". I'm tempted now to get the single or dual IDE channel card made
by Highpoint, and I will if I can't find a way to make the SIIG and SYBA
cards work together. The $100 4-channel card by Highpoint would be a
last (expensive) resort.

Thanks for the info.
Rick Lowen
 
It seems strange to me that two identical Promise cards would work together,
Click to expand...

They dont always.
and a SIIG doesn't work with a SYBA. I get the nagging
feeling that there might be a way to make the SIIG and SYBA cards work
together without major surgery.
Click to expand...

Yes, its possible. Its possible that the PCI
slots used matter, particularly with the IRQ use.
 
Richard Lowen said:
I heard from Highpoint Support. They said they "know of no conflicts"
between their cards and the SIIG and SYBA cards and that they "should
work". I'm tempted now to get the single or dual IDE channel card made
by Highpoint, and I will if I can't find a way to make the SIIG and SYBA
cards work together. The $100 4-channel card by Highpoint would be a
last (expensive) resort.
Click to expand...


I just spoke with SIIG support (you can actually *speak* with someone
intelligent at SIIG!). The rep said that he had done an experiment as
a result of my call yesterday, and by wiping out the on-card BIOS of
one card, he could get past the boot conflict - BUT, it rendered the
card without its BIOS inoperable. So..., he suggested going with
either an external HD or getting a 4-channel card (which SIIG doesn't
make). He also laughed at Highpoint support's reply that they didn't
"know of any conflicts" and that their card "should work" with SIIG's card.
He said the lack of negative news is not positive news. <groan>
Of course, he's right.

Rick Lowen
 
Rod Speed said:
They dont always.


Yes, its possible. Its possible that the PCI
slots used matter, particularly with the IRQ use.
Click to expand...


Thanks for suggestion. I'll play Musical Slots when I get the time and
report back to the newsgroup. I hope I don't run into a Windows XP
re-activation issue.

Rick Lowen
 
Thanks for suggestion. I'll play Musical Slots when I get the time and report
back to the newsgroup.
Click to expand...

You may well have some control of the IRQ used by
individual PCI slots with a motherboard that elderly too.

Doesnt look that promising given the SIIG test with the bios
conflicting tho. That doesnt sound like an IRQ problem, sounds
like a poorly designed bios that doesnt allow for more than one.

I'd be replacing the motherboard myself. Lot less hassle.
I hope I don't run into a Windows XP re-activation issue.
Click to expand...

Shouldnt matter even if you do, just explain what you
are trying to do and why its triggered the reactivation.
 
It doesn't have to be cards.
It's also possible that an interrupt is shared with an internal.
I think in many boards (certainly those that were around a couple of
years ago) the first and last PCI slots shared the same interrupt.
Click to expand...

That depends on the number of PCI slots used since there are 4 PCI
interrupt lines and (in theory) they are connected to each slot, all 4 of
them, but the interrupt numbers are rotated on each slot so the same pin
is IntA on one slot, B the next, then C then D and if you have 5 slots, A again.
IntB on the first slot is C on the next, .... etc. etc. IntC on first slot is D on
the next..... etc. etc. You get the picture?
Cards usually use one (of the four) interrupt(s), usually A.
The interrupt number rotation then makes them use different interrupts
indifferent slots.
 
Richard said:
I just spoke with SIIG support (you can actually *speak* with someone
intelligent at SIIG!). The rep said that he had done an experiment as
a result of my call yesterday, and by wiping out the on-card BIOS of
one card, he could get past the boot conflict - BUT, it rendered the
card without its BIOS inoperable. So..., he suggested going with
either an external HD or getting a 4-channel card (which SIIG doesn't
make). He also laughed at Highpoint support's reply that they didn't
"know of any conflicts" and that their card "should work" with SIIG's
card.
He said the lack of negative news is not positive news. <groan>
Of course, he's right.
Click to expand...

Sounds like Plug and Pray fails yet again. There is a special hell for the
people who came up with the idea of implementing that without making manual
overrides a standard part of the system.
 
J. Clarke said:
Sounds like Plug and Pray fails yet again. There is a special hell for the
people who came up with the idea of implementing that without making manual
overrides a standard part of the system.
Click to expand...
Another clueless troll from Johny. PCI PnP works just fine.
If you are stupid enough to buy broken hardware, it doesn't.
 
Another clueless troll from Johny. PCI PnP works just fine.
If you are stupid enough to buy broken hardware, it doesn't.
Click to expand...

Actually you're wrong. Those PnP devices worked fine separately but
they won't work together in the same computer. Multiple PCI IDE card
rarely works together in the same PC anyway.
 
Actually you're wrong.
Nope.

Those PnP devices worked fine separately but
they won't work together in the same computer.
Click to expand...

Not because of PnP, just because of a ****ed approach the bios takes.
Multiple PCI IDE card rarely works together in the same PC anyway.
Click to expand...

Wrong again.
 
Back
Top