running ATA-133 hard drives in older BX motherboard systems.

Joseph <spmok2 RemoveThis @canada.com> wrote in message
news:1112115530.855858.257070@l41g2000cwc.googlegroups.com...

 > With so many systems with BX motherboards around for such great price
 > I am hoping to utilize these systems. Only problem is their IDE controllers
 > usually only run at ATA-33 (way too slow for anything good).

 > I have an idea that if I can plug in a pci-ide ata-133 controller into
 > these systems, then the problem is solved. However through more
 > research I realize this has an inherent problem. Which is that you
 > need to install the Win2k driver so that it will detect these pci
 > cards. so that u can then see the hard drive its connected to.
 > Defeats purpose because you must already be running your
 > Win2k boot OS on the slower interal drive.

Nope, Win2K loads the driver at boot time, so
it runs the card at full speed once its booted.

 > My question is this. Are there any PCI IDE ata133
 > controllers out there that has bios boot capability?

Yes, most of them do.

 > By this I mean being able to at bootup of computer read
 > an installed OS on its hard drive and boot from that instead.

Yes, thats how SCSI works as well.

 > Do you remember back in the old days, Adaptec SCSI
 > controller has this ability to hijack the boot process and
 > force a boot from a scsi drive connected on its daisy chain?
 > I am hoping ot find ATA-133 controller PCI cards that can do this.

Yes, most of them do. Its uncommon to find one that cant.

 > Anyone know anything about it?<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

On Tue, 29 Mar 2005 14:18:48 -0800, "Eric Gisin"
<ericgisin RemoveThis @hotmail.com> wrote:

 >What are you using them for, servers? A pentium-II can do 25 MB/s or less on
 >ethernet, so the UDMA-33 channels are not really a bottleneck.

440BX was also used with P3's

 >The bigger problem with older BX boards is the 32GB BIOS bug.

It should be conquerable between software layer and/or win2k or
above's non-reliance on the BIOS.

With the price of ATA controllers, though, there's absolutely no point
in dealing with ATA33 mode or the potential nuisances of the BIOS
limitations<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

Curious George wrote:

 > On Tue, 29 Mar 2005 14:18:48 -0800, "Eric Gisin"
 > <ericgisin DeleteThis @hotmail.com> wrote:
 >
  >>What are you using them for, servers? A pentium-II can do 25 MB/s or less
  >>on ethernet, so the UDMA-33 channels are not really a bottleneck.
 >
 > 440BX was also used with P3's
 >
  >>The bigger problem with older BX boards is the 32GB BIOS bug.
 >
 > It should be conquerable between software layer and/or win2k or
 > above's non-reliance on the BIOS.
 >
 > With the price of ATA controllers, though, there's absolutely no point
 > in dealing with ATA33 mode or the potential nuisances of the BIOS
 > limitations

Ignore Eric--he's in "full of crap" mode.

I've got 750 gig running on a BX server, and getting 300 Mb/sec out of
Ethernet in the process, on Windows, which is not particularly efficient in
that regard.

His "25 MB/sec" is actually the PCI bottleneck around 400 Mb/sec.

His "32 GB BIOS bug" may be present on some boards but it certainly isn't on
Intel or Supermicro server boards.

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

"J. Clarke" <jclarke.usenet DeleteThis @snet.net.invalid> wrote in message
news:d2dpva024g8@news2.newsguy.com...
 > Curious George wrote:
 >
  > > On Tue, 29 Mar 2005 14:18:48 -0800, "Eric Gisin"
  > > <ericgisin DeleteThis @hotmail.com> wrote:
  > >
   > >>What are you using them for, servers? A pentium-II can do 25 MB/s or less
   > >>on ethernet, so the UDMA-33 channels are not really a bottleneck.
  > >
  > > 440BX was also used with P3's
  > >
Who wants 100MHz FSB? Many were too dumb to realize the i815 was superior.

Even an i810 makes a good server, the IDE ports are not on PCI, so GB
ethernet has more available.

   > >>The bigger problem with older BX boards is the 32GB BIOS bug.
  > >
  > > It should be conquerable between software layer and/or win2k or
  > > above's non-reliance on the BIOS.
  > >
  > > With the price of ATA controllers, though, there's absolutely no point
  > > in dealing with ATA33 mode or the potential nuisances of the BIOS
  > > limitations
 >
 > Ignore Eric--he's in "full of crap" mode.
 >
We see a lot more bullshit coming from you.

 > I've got 750 gig running on a BX server, and getting 300 Mb/sec out of
 > Ethernet in the process, on Windows, which is not particularly efficient in
 > that regard.
 >
40MB/s for ethernet and another 40 for IDE. The PCI bottleneck is more
significant than 2*30MB/s of the PIIX4.

 > His "25 MB/sec" is actually the PCI bottleneck around 400 Mb/sec.
 >
There is no such bottleneck in PCI. Tell us some more bullshit.

 > His "32 GB BIOS bug" may be present on some boards but it certainly isn't
on
 > Intel or Supermicro server boards.
 >
The 440BX's days were 97-99, when there were no drives larger than 30GB.
Intel did not use Award BIOS.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

 > 440BX was also used with P3's
 >
  > >The bigger problem with older BX boards is the 32GB BIOS bug.
 >
 > It should be conquerable between software layer and/or win2k or
 > above's non-reliance on the BIOS.
 >
 > With the price of ATA controllers, though, there's absolutely no point
 > in dealing with ATA33 mode or the potential nuisances of the BIOS
 > limitations

Three or four years ago I dropped a 1.2 gig Celeron into my Dell 400 MHz
440BX box, added a Promise 100TX controller with a couple of DeathStars
(which have always ran perfect for me) and it's still running fine for my
wife. Works fine as a file server to store my video files, she does
PhotoShop stuff on it and my daughter does some limited video editing on it
as well. Not bad for a box almost seven years old.

So to the OP, yes the card will enable a bios boot so if you just need the
extra space, a Promise card and a new drive is a worthwhile upgrade for a
440BX box.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

Previously Curious George <cg.RemoveThis@email.net> wrote:
 > On Tue, 29 Mar 2005 14:18:48 -0800, "Eric Gisin"
 > <ericgisin.RemoveThis@hotmail.com> wrote:

  >>What are you using them for, servers? A pentium-II can do 25 MB/s or less on
  >>ethernet, so the UDMA-33 channels are not really a bottleneck.

A current mainboard with a Gigabit Ethernet card in a PCI slot also is
limited to 250...450 Mbit/s = 30...60MB/s in my experience. I have
seen peak rates of 600Mbit/s = 75MB/s, but only for UDP streaming,
with the right card/mainboard combo and no other device on the
IRQ used by the card.

For more speed a faster bus is needed. I see >900Mbit/s with TCP both
on a 33MHz/64bit PCI bus (dual Athlon 2800+, Intel NIC) as well as on
a PCI-X bus (Dual Opteron, Broadcom NIC).

It is not processing overhead, but the interrupt latency and bus
delays that limit network speed on traditional 33MHz/32bit PCI.

Arno<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

"Joseph" <spmok2.RemoveThis@canada.com> wrote in message
news:1112115530.855858.257070@l41g2000cwc.googlegroups.com...
 > With so many systems with BX motherboards around for such great price I
 > am hoping to utilize these systems. Only problem is their IDE
 > controllers usually only run at ATA-33 (way too slow for anything
 > good).
 >
 > I have an idea that if I can plug in a pci-ide ata-133 controller into
 > these systems, then the problem is solved. However through more
 > research I realize this has an inherent problem. Which is that you
 > need to install the Win2k driver so that it will detect these pci
 > cards. so that u can then see the hard drive its connected to. Defeats
 > purpose because you must already be running your Win2k boot OS on the
 > slower interal drive.
 >
 > My question is this. Are there any PCI IDE ata133 controllers out
 > there that has bios boot capability? By this I mean being able to at
 > bootup of computer read an installed OS on its hard drive and boot from
 > that instead. Do you remember back in the old days, Adaptec SCSI
 > controller has this ability to hijack the boot process and force a boot
 > from a scsi drive connected on its daisy chain? I am hoping ot find
 > ATA-133 controller PCI cards that can do this. Anyone know anything
 > about it?
 >

Pseudo ide controller cards are viewed by the PC as SCSI. So just set the
bios to boot from SCSI first. If you don't have that option, remove all
from the primary ide controller and disable it in the bios. It still may
work.

If you already have a SCSI controller, the one seen first by the PC is the
one the PC will try to boot from. But not both an pseudo ide controller
card and scsi controller card.

Yes, you have to install the driver for the controller card. This technique
is well-documented with Promise controller card additions.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

On Wed, 30 Mar 2005 03:55:24 -0500, "J. Clarke"
<jclarke.usenet.DeleteThis@snet.net.invalid> wrote:

 >Ignore Eric--he's in "full of crap" mode.
OK

 >I've got 750 gig running on a BX server, and getting 300 Mb/sec out of
 >Ethernet in the process, on Windows, which is not particularly efficient in
 >that regard.

Bit apples & oranges. It may be a BX but you're not using the onboard
ATA or onboard ethernet (or equivalent) so it doesn't really deal with
Eric's comment re usefulness of U33 for a server we're responding to.

 >His "25 MB/sec" is actually the PCI bottleneck around 400 Mb/sec.

Total available PCI 32/33 bandwidth is more like 133 MB/sec minus
overhead. 400Mb/sec is more like 50MB/sec, not 25.

 >His "32 GB BIOS bug" may be present on some boards but it certainly isn't on
 >Intel or Supermicro server boards.

Dunno. Never bothered upgrading one that way. More familiar with the
Serverworks boards of that day (which bypass the bandwidth
restrictions being discussed). Probably a better bet for a second
hand cheap server you want to run raid or new storage and gigabit with
decent results.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

Curious George wrote:

 > On Wed, 30 Mar 2005 03:55:24 -0500, "J. Clarke"
 > <jclarke.usenet DeleteThis @snet.net.invalid> wrote:
 >
  >>Ignore Eric--he's in "full of crap" mode.
 > OK
 >
  >>I've got 750 gig running on a BX server, and getting 300 Mb/sec out of
  >>Ethernet in the process, on Windows, which is not particularly efficient
  >>in that regard.
 >
 > Bit apples & oranges. It may be a BX but you're not using the onboard
 > ATA or onboard ethernet (or equivalent) so it doesn't really deal with
 > Eric's comment re usefulness of U33 for a server we're responding to.

Well, actually, that same server until a couple of months ago was having no
trouble filling a 100 Mb/sec pipe using a 120 gig drive connected to the
onboard host adapter and using the native Ethernet. If you don't have a
gigabit board in the machine then you're limited by the bandwidth of the
Ethernet port anyway.

  >>His "25 MB/sec" is actually the PCI bottleneck around 400 Mb/sec.
 >
 > Total available PCI 32/33 bandwidth is more like 133 MB/sec minus
 > overhead.

That's the "total available bandwidth". Which is meaningful only on
machines with a dual bus or on transfers from RAM to LAN or RAM to disk.
In the real world of servers you have to move data from disk to RAM then
from RAM to LAN, which means two PCI transfers and overhead, and in the
real world it is generally accepted that you're going to be limited to
about 400 MB/sec across the wire on a machine with a PCI bus.

 > 400Mb/sec is more like 50MB/sec, not 25.

You're correct--what I get for posting when tired.

  >>His "32 GB BIOS bug" may be present on some boards but it certainly isn't
  >>on Intel or Supermicro server boards.
 >
 > Dunno. Never bothered upgrading one that way. More familiar with the
 > Serverworks boards of that day (which bypass the bandwidth
 > restrictions being discussed). Probably a better bet for a second
 > hand cheap server you want to run raid or new storage and gigabit with
 > decent results.

The Intel server building blocks work just fine. And are relatively cheap
these days. They're GX, not BX, but it's a minor distinction. You're not
going to fill a gigabit pipe with one unless you do everything perfectly,
but they're quite stable and fast enough for any use I can think of where
ebay hardware would be appropriate.

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems. 

On Sat, 02 Apr 2005 10:26:27 -0500, "J. Clarke"
<jclarke.usenet.TakeThisOut@snet.net.invalid> wrote:

 >Curious George wrote:
 >
  >> On Wed, 30 Mar 2005 03:55:24 -0500, "J. Clarke"
  >> <jclarke.usenet.TakeThisOut@snet.net.invalid> wrote:
  >>
   >>>Ignore Eric--he's in "full of crap" mode.
  >> OK
  >>
   >>>I've got 750 gig running on a BX server, and getting 300 Mb/sec out of
   >>>Ethernet in the process, on Windows, which is not particularly efficient
   >>>in that regard.
  >>
  >> Bit apples & oranges. It may be a BX but you're not using the onboard
  >> ATA or onboard ethernet (or equivalent) so it doesn't really deal with
  >> Eric's comment re usefulness of U33 for a server we're responding to.
 >
 >Well, actually, that same server until a couple of months ago was having no
 >trouble filling a 100 Mb/sec pipe using a 120 gig drive connected to the
 >onboard host adapter and using the native Ethernet.

100% efficiency? If that's what you mean I'll need to get a shovel.

 > If you don't have a
 >gigabit board in the machine then you're limited by the bandwidth of the
 >Ethernet port anyway.
 >
   >>>His "25 MB/sec" is actually the PCI bottleneck around 400 Mb/sec.
  >>
  >> Total available PCI 32/33 bandwidth is more like 133 MB/sec minus
  >> overhead.
 >
 >That's the "total available bandwidth". Which is meaningful only on
 >machines with a dual bus or on transfers from RAM to LAN or RAM to disk.
 >In the real world of servers you have to move data from disk to RAM then
 >from RAM to LAN, which means two PCI transfers and overhead, and in the
 >real world it is generally accepted that you're going to be limited to
 >about 400 MB/sec across the wire on a machine with a PCI bus.

You're talking basically file server role where I guess I agree
~50MB/sec is a maximal number. PII's normally choke before these
numbers get reached anyway.

  >> 400Mb/sec is more like 50MB/sec, not 25.
 >
 >You're correct--what I get for posting when tired.
 >
   >>>His "32 GB BIOS bug" may be present on some boards but it certainly isn't
   >>>on Intel or Supermicro server boards.
  >>
  >> Dunno. Never bothered upgrading one that way. More familiar with the
  >> Serverworks boards of that day (which bypass the bandwidth
  >> restrictions being discussed). Probably a better bet for a second
  >> hand cheap server you want to run raid or new storage and gigabit with
  >> decent results.
 >
 >The Intel server building blocks work just fine. And are relatively cheap
 >these days. They're GX, not BX, but it's a minor distinction. You're not
 >going to fill a gigabit pipe with one unless you do everything perfectly,
 >but they're quite stable and fast enough for any use I can think of where
 >ebay hardware would be appropriate.

I don't disagree. Either platforms are fine for a home nas, firewall,
backup media server, etc. with fast ethernet normally being fine also.
I've just seen some PIII systems with 133MHz ECC ram and multiple PCI
busses in a similar price range on eBay. It's probably more important
to make sure its one in good condition and with a few extras for as
little money as you can manage for that kind of purchase.<!-- ~MESSAGE_AFTER~ -->
 >> Stay informed about: running ATA-133 hard drives in older BX motherboard systems.