Large Hard Drive on Promise Card Doesn't Work

"Ken Dibble" wrote in message
...
snipped

I understand that there remains some disagreement on what will happen
if I attempt to actually use the second partition on this new drive.
Can anyone add anything definitive on that issue?

If both show in Windows and neither have an issue, both are fully usable.


--

Brian A. Sesko { MS MVP_Shell/User }
Conflicts start where information lacks.
http://basconotw.mvps.org/

Suggested posting do's/don'ts: http://www.dts-l.org/goodpost.htm
How to ask a question: http://support.microsoft.com/kb/555375

| The link that Rick
| supplied for the Promise manual was a useful clue, although the manual
| was actually incorrect as to what I needed to do to flash the card
| BIOS.

Glad you got yours working, & I see your question has been answered. What did you do different from what the manual said...?...

BIOS Upgrade

The following instructions detail the steps that need to be taken in order to upgrade the BIOS of your Ultra100 controller card. BIOS (Basic Input/Output System) controls the booting operation of the Ultra100. If you receive information regarding an update for the Ultra100 BIOS (from the Promise Technology website or technical support), download the update and then follow these instructions to install the update.

NOTE: Before updating the Ultra100 BIOS, you can save the old BIOS by choosing option 1 in the main menu (“Backup Current ROM BIOS to File”). Type the path and file name where you wish to save the old BIOS (i.e., “A:\OLDBIOS.BIN”).

1. Create a bootable floppy, and copy the files “PTIFLASH.EXE” and “ULBIOS.BIN” to the floppy.
2. The file “PTIFLASH.EXE” is a command prompt-only program; a bootable floppy must be used to run the program.
3. At the “A:\” prompt, type “PTIFLASH.” The main menu should be generated.
4. In the main menu, choose option 2 - “UpdateBIOS From File.” In the pop-up dialog box that appears, enter the path and name of the BIOS file you wish to update to (i.e., “A:\ULBIOS.BIN”).

NOTE: If, after typing in the path and file name while trying to update the BIOS, you receive the message “This file is not found,” make sure you’ve entered the correct path and file name.

5. The utility will then update the Ultra100 BIOS, and an “Update Success” message will appear.
6. Restart the system.

NOTE: When the Ultra100 BIOS appears, make sure that the BIOS version is the new version.


--
Thanks or Good Luck,
There may be humor in this post, and,
Naturally, you will not sue,
should things get worse after this,
PCR

"Ken Dibble" wrote in message ...
| Thanks very much to all who replied. My problem appears to be solved!
|
| I moved the 250 GB drive to the slave position on the original cable
| on my Promise card--the one that was already supporting a 40 GB drive
| without problems. This single step immediately cleared up all the
| performance problems with the new drive. I could now reliably transfer
| large files to and from the drive as well as modify and run them on
| the new drive.
|
| My third-party partitioner still couldn't see the full 250 GB, though,
| and ScanDisk was still running on boot-up. I followed the links
| supplied by Rick for the Promise site and was able to download a later
| Windows driver version than the one I had installed. (Apparently the
| link I'd followed, from the Promise homepage, to obtain an updated
| Windows driver was incorrect.) I ascertained that the card BIOS update
| that I'd downloaded earlier was the correct one. The link that Rick
| supplied for the Promise manual was a useful clue, although the manual
| was actually incorrect as to what I needed to do to flash the card
| BIOS.
|
| Once I got the BIOS and driver upgrades installed, my Acronis Disk
| Director partitioner could see the full 250 GB, and I created a second
| partition using the remaining free space. However, ScanDisk was still
| running on bootup. I downloaded and installed the Win ME versions of
| ScanDisk and Defrag, but this didn't stop ScanDisk from running.
|
| So I deleted both partitions, rebooted, created two 116 GB partitions,
| rebooted, formatted the partitions, rebooted, and finally transferred
| a large multimedia file to the first new partition and ran it. I
| rebooted again and ScanDisk did not run!
|
| So it appears that all is well.
|
| I understand that there remains some disagreement on what will happen
| if I attempt to actually use the second partition on this new drive.
| Can anyone add anything definitive on that issue?
|
| Thanks very much again. This was extremely helpful!
|
| Ken Dibble

Ken Dibble wrote:
Thanks very much to all who replied. My problem appears to be solved!

I moved the 250 GB drive to the slave position on the original cable
on my Promise card--the one that was already supporting a 40 GB drive
without problems. This single step immediately cleared up all the
performance problems with the new drive. I could now reliably transfer
large files to and from the drive as well as modify and run them on
the new drive.

I like to connect only two hard drives to the controller, and each hard drive is
connected as Master of course, where one master to each of the two ports on the card,
and never connect anything as slave to either - this allows for exceptional speeds
for disk to disk transfers (not having to use the same cable to do 2way data travel).
This way is very very useful lets say if you use DriveImage or TrueImage (partition
imagers - I have them both and lovem) and using their disk to disk (cloning) and is
where I'm always dazzled at the excellent transfer speeds I can get doing it this way
off the Card.
Anyway, maybe you have a defective second cable, or bad second port on your card and
is why it wouldn't work right on that side.


My third-party partitioner still couldn't see the full 250 GB, though,
and ScanDisk was still running on boot-up. I followed the links
supplied by Rick for the Promise site and was able to download a later
Windows driver version than the one I had installed. (Apparently the
link I'd followed, from the Promise homepage, to obtain an updated
Windows driver was incorrect.) I ascertained that the card BIOS update
that I'd downloaded earlier was the correct one. The link that Rick
supplied for the Promise manual was a useful clue, although the manual
was actually incorrect as to what I needed to do to flash the card
BIOS.

Yes the bios flash explanation on that particular pdf is a bit scant. Actually
Promise made that card for some other manufactures too who just put their name on it,
but for one, the Maxtor(Promise) version of the 100 card at Maxtors website gives a
little more flash details if you were interested in saving the webpage as an mht for
reference.
http://www.maxtor.com/portal/site/Ma...&downloadID=28
...and/or you can go to Promise and download their 133 version bios but I only
suggest that Only because they have a little readme in the zip and so you can read
their similar flash instructions. In either case you can ad-lib your own filename
http://www.promise.com/support/downl...=all&os=100 #
..of course do not use that 133 bios ...smile but by all means buy the card


Once I got the BIOS and driver upgrades installed, my Acronis Disk
Director partitioner could see the full 250 GB, and I created a second

Fancy that you use Acronis Disk Director
I hope even moreso that you Also use Acronis TrueImage imaging software which is
imho, beside DriveImage being 2nd, is the finest Partition Imager in the world..
imvho

partition using the remaining free space. However, ScanDisk was still
running on bootup. I downloaded and installed the Win ME versions of
ScanDisk and Defrag, but this didn't stop ScanDisk from running.
So I deleted both partitions, rebooted, created two 116 GB partitions,

For my own preference I like to make all my storage partitions even below 128 too,
and I choose to and do make them 121 GB even. FAT32 primary OS partitions are
specifically made as previously stated.

rebooted, formatted the partitions, rebooted, and finally transferred
a large multimedia file to the first new partition and ran it. I
rebooted again and ScanDisk did not run!

So it appears that all is well.

great then.

I understand that there remains some disagreement on what will happen

I don't think when you get right down to the basic facts that there's really any
disagreement though, this subject is all old news anyway and there is nothing new
about the basics really, its just perhaps at first a little misunderstanding about
getting all the presented facts on the table properly to be understood in full, but
we all here don't really care who is right or wrong to begin with - but only that
ultimately we serve everyone with the best complete and accurate facts that we can,
and that is the only things that we should ever concentrate on - and so we serve to
do so.

if I attempt to actually use the second partition on this new drive.
Can anyone add anything definitive on that issue?

There should be absolutely no problem as long as your Card is bios'd and driver'd
correctly, and the card itself and your HD is working correctly, etc, then by all
means it should be fine... all mine are.

Rick

Thanks very much again. This was extremely helpful!

Ken Dibble

On Mon, 26 Dec 2005 19:47:01 -0500, "Rick Chauvin"
cquirke (MVP Windows shell/user) wrote:


Could you provide a list of what is required for Win98SE to support
48-LBA? Does this work for Win98 and WinME?

Simply a bios setup that supports LBA
Absolutely of course it works for WinME.

No fixes to the OSs? Is DOS mode safe, too?

The easiest way for anyone to get past large hard drive support is to use the right
bios setup - and the easiest way to do that for most people is to use something
simple like a controller card.

Agreed - it may be safer to do that, than to attempt a BIOS update,
and chances are the new HD will benefit from improved UDMA support
that a new card would add over whatever shipped with an older mobo.

I know Chris you don't have your newsreader setup like the rest of us,
and don't use OE, and so don't see your threads all in order

That's not the problem - I have exactly the same problems when I use
OE for the private MS newsgroups.

It's a server post aging thing that hits busy newsgroups in
particular; by the time I re-check news, some posts I couldn't get
last time, are already aged off so that I can't get them this time.

This is obviously worse when I come back to ngs after a period of
absence, as I'm doing now. Even three days without checking for new
posts is enough to disrupt continuity; all the existing headers marked
to retreive bodies find no bodies, and the new headers do not start
where the last ones left off.

"with the proper ingredients"

What proper ingredients?

Not much really, besides the Card, the other ingredients is simply have no partition
over 128 GB. It's FAT32 naturally. It needs to be under 128GB for its stock
Scandisk and Defrag to work properly, otherwise it won't

See, that 128G-per-volume limit suggests the OS is not OK with 48-LBA
after all. The assumption on which all this rests, is that the OS
never addresses the physical HD, but always addresses sectors relative
to the start of the volume, which as successfully been found via BIOS
rather than native OS code. Further, it assumes that the mechanism
whereby these calls are translated to raw disk addressing is
48-LBA-safe, presumably by using BIOS services.

Given that Win9x redirects BIOS HD interrupt service vectors to its
own code (unless in DOS compatibility mode), I'm wondering whether the
OS can in fact safely find the start of volumes over the 137G limit.

Given that Win9x detects HDs and volumes via PnP rather than simply
inheriting these via BIOS, I'd wonder whether BIOS 48-LBA is in
effectwhen these are discovered?

Perhaps the key to this is whether the HD controller card's drivers
impart 48-LBA safety, and this might explain better mileage with
particular-brand add-on controller cards vs. integrated controllers?

As it is, I note DOS mode Scandisk can corrupt volumes that traverse
the 137G limit, BIOS 48-LBA notwithstanding.

For all of these reasons, as well as the failure of SP1 to truly be
48-LBA-safe, I'd really need to be COAB sure before suggesting the use
of large HDs with Win9x. Else it could be a case of "works fine long
enough to build up an irreplaceable data set, then eats the lot".

OTOH, you sound pretty sure of this, and it sounds as if you've
researched and tested it pretty thouroughly. Can you URL me? The
URLs you mentioned above, don't really hit the spot on this. For
example, does the Promise site have a tech FAQ on this?

Win98SE is not capable of addressing beyond "128G" but that this is OK
as it addresses each volume only from the start of that volume as
address 0, so that even if the volume passes the HD's "128G" (137G is
what I usually hear) limit, it doesn't matter.

The regular blurb about that issue is that 128 GB which is the true windows binary
limit - not the 137 decimal GB which many mention but that's only a HD manufacturers
marketing spec number used in decimals by them so that their numbers look/sound
bigger for marketing selling points, but most importantly the answer is that since
Windows and Windows Utilities uses binary calculations, and so that 137 converts to
128 GB only. If someone was to try a 137 figure they would find it would not work and
don't try it with a good setup, only go by binary numbers therefore the 128 GB is the
true number and what should be used when working with Windows Utilities.

OK, I was wondering about that. Often 1G = 1000 x 1M, or 1000 x 1000
k, whereas in binary it's 1024 x 1024 x k. Seeing as there are no HDs
between 120G and 160G, it doen't matter at that level, but kicks in
when you have to select max safe volume sizes. It may also be that
128G is where FAT32 changes cluster size to the nextsize up.

I don't see myself running Win9x on new PCs in GUI mode, but I would
like to use DOS mode Scandisk to maintain FATxx volumes in XP (via
alternate DOS mode boot), as I don't trust ChkDsk / AutoChk.

So far, my mileage has been poor there. You mention WinME's Scandisk,
but does WinME have a DOS mode Scandisk, and if so, is it safe?

The systems I will be using this on, will be...
- 7.99G FAT32 C:
- extended for rest of HD, as...
- 2G FAT16 D:
- laaargeG FAT32 E:
- 2G FAT16 or 7.99G FAT32 F:

That E: will be 128G on 120G HDs, but 128G on 200G HDs.



---------- ----- ---- --- -- - - - -
Don't pay malware vendors - boycott Sony
---------- ----- ---- --- -- - - - -

On Tue, 27 Dec 2005 05:12:41 GMT, Ken Dibble

My third-party partitioner still couldn't see the full 250 GB, though

Which partitioner?

and ScanDisk was still running on boot-up.

ScanDisk will run on bootup if one or both of two flags are set within
the FAT of the volume. One flag is left set if file system operations
are interrupted by a bad exit from Windows, which implies there will
be logical file system corruption to clean up. The other is set
whenever a disk access fails at the physical level, and this will
prompt a surface scan of all volumes on that physical HD.

Because Windows ALWAYS writes to C:, it's safest to perform this
automatic Scandisk before Windows boots. In Win95xx/98xx, this is
done by running DOS mode Scandisk before Windows loads.

In WinME, there is no suitable DOS mode opportunity to do this, so the
Windows Scandisk is run instead - and indeed, is constantly
interrupted by potentially-destructive file writes, which maybe why MS
"fixes" without prompting by default, in the hope that this will
happen fast enough to complete before such interruptions.

What a mess!

...for the Promise site and was able to download a later Windows
driver version than the one I had installed. I ascertained that the
card BIOS update that I'd downloaded earlier was the correct one.

This is pretty messy and fragile-looking as well.

Once I got the BIOS and driver upgrades installed, my Acronis Disk
Director partitioner could see the full 250G

OK - it sounds as if the partitioner relies on BIOS 48-LBA, which the
previous setup was not putting into effect. Ew.

However, ScanDisk was still running on bootup. I downloaded
and installed the Win ME versions of ScanDisk and Defrag, but
this didn't stop ScanDisk from running.

It wouldn't, if that was the Windows ScandiskW.exe, because that is
not what Win95xx or Win98xx uses in this circumstance. Does WinME
have an updated DOS mode Scandisk.exe as well?

So I deleted both partitions, rebooted, created two 116G partitions,
rebooted, formatted the partitions, rebooted, and finally transferred
a large multimedia file to the first new partition and ran it. I
rebooted again and ScanDisk did not run!

That's a sledgehammer was of resetting the flags in the FAT that were
prompting Scandisk to run. I suspect if those flags ever get set
again, it will be back to square one... and if DOS-based disk editing
tools like Diskedit are not 48-LBA-OK, you may not be able to cleanly
clear the flag in FAT in that way.

So it appears that all is well.

Until the next bad exit, that is.

I suspect the situation is this:
- BIOS is 48-LBA-OK
- Windows HD controller drivers are 48-LBA-OK
- Windows is not 48-LBA-OK (max volume size must be 128G)
- DOS mode and tools (e.g. Scandisk) are not 48-LBA-OK

So if Windows sets a flag to prompt DOS mode Scandisk, DOS mode
Scandisk may corrupt the volume and/or be unable to clear that flag.
Hence perpetual DOS mode Scandisk on every boot, forcing you to either
suppress the automatic Scandisk via C:\WINBOOT.INI or C:\MSDOS.SYS,
"just" wipe and re-create the affected volume, or find some
48-LBA-safe tool to clear the flag.

This suggests the only way to maintain volumes over the 128G limit is
to do so from within Windows, using WinME's ScandiskW.exe (assuming
this is safe for 48-LBA). Perhaps this will clear the relevant flags
in FAT so that the automatic DOS mode Scandisk goes away.

This in turn suggests you must ensure Windows does not automatically
write to these volumes, so that it is safe to start Windows and then
Scandisk them from there. No Windows OS install on such volumes, and
betware indexers and other underfootware trash.

Thanks very much again. This was extremely helpful!

We could do with a web page on 128G in Win9x; it seems possible, but
looks like a bit of a minefield. I don't like dangling my data's
health off a 3rd-party BIOS or driver version point revision.



---------- ----- ---- --- -- - - - -
Don't pay malware vendors - boycott Sony
---------- ----- ---- --- -- - - - -

"Rick Chauvin" wrote in message ...
| Ken Dibble wrote:
|
....snip
| My third-party partitioner still couldn't see the full 250 GB, though,
| and ScanDisk was still running on boot-up. I followed the links
| supplied by Rick for the Promise site and was able to download a later
| Windows driver version than the one I had installed. (Apparently the
| link I'd followed, from the Promise homepage, to obtain an updated
| Windows driver was incorrect.) I ascertained that the card BIOS update
| that I'd downloaded earlier was the correct one. The link that Rick
| supplied for the Promise manual was a useful clue, although the manual
| was actually incorrect as to what I needed to do to flash the card
| BIOS.
|
| Yes the bios flash explanation on that particular pdf is a bit scant. Actually
| Promise made that card for some other manufactures too who just put their name on it,
| but for one, the Maxtor(Promise) version of the 100 card at Maxtors website gives a
| little more flash details if you were interested in saving the webpage as an mht for
| reference.
| http://www.maxtor.com/portal/site/Ma...&downloadID=28
....snip

This one does it without making the flash floppy itself bootable. (I would have deleted all the files from a Startup Diskette & put the two Promise files there.)Also, it says to unplug the drives at the controller end first. That's all that's different, that I see.

WELL, naturally, Maxtor has changed the name of the BIN from Promise's B14.BIN to 267B43.BIN. Can't use the Maxtor download. It is only... BIOS Version 2.01.0 (Build 43), where the Promise is... v2.20.0.14, you said.

.......Quote............
After downloading U100B43.EXE insert a new formatted floppy into the A:\ drive. Double click on the U100B43.EXE icon, the files will automatically load on to the floppy. The two files that load to the floppy are PTIFLASH.EXE, 267B43.BIN. The BIOS flash does not run in a Windows environment and is not bootable, you must first boot the system to a Windows 95/98/ME startup disk. Power the system off, and disconnect all Hard Disk data cables from the ultra controller. Insert the Windows 95/98/ME startup disk into the A:\ drive and power the system on. After the system boots to the A: drive, remove the startup disk and insert the disk with the U100B43 BIOS into the A:\ drive. Follow the instructions below to perform the update.As a precaution It is recommended to disconnect any drives that are connected to the Ultra ATA 100 PCI Adapter Card before performing the BIOS flash.

Step-by-Step Procedure

After the system boots to the A:\ prompt and you have inserted the floppy with the U100B43 BIOS into the A:\ drive.

From the A:\ prompt type PTIFLASH and press [Enter] this will execute the Flash Memory Utility.

The Old BIOS needs to be backed up first. Press number 1 on the keyboard this will display the Backup BIOS screen. Press the [Backspace] key to move the flashing cursor to the left side of the field, then type A:\OLDB27.BIN and press the [Enter] key. After a moment the Backup Success box will display with the name of the saved file. Press [Enter] to continue. From the flashing cursor after Please Select: press number 2 on your keyboard. The Update BIOS screen will appear. Press the Backspace key to move the cursor to the far left of the field and type 267B43.BIN and press [Enter].

After the Update Success screen appears, remove the floppy from the A:\ drive and power the system off.

Power the system on to check the BIOS revision of the Ultra ATA PCI Adapter Card. When the card is detected on the system, the new BIOS revision will show as Ultra100â„¢ BIOS Version 2.01.0 (Build 43).

The BIOS has flashed successfully and you can power the system off, reconnect cabling to the Ultra controller and restart the system. After the system boots to the Windows desktop, replace the existing Ultra Host driver with supporting 48-Bit version 2.0.0050.42.
.......EOQ..............

cquirke (MVP Windows shell/user) wrote:
On Mon, 26 Dec 2005 19:47:01 -0500, "Rick Chauvin"
cquirke (MVP Windows shell/user) wrote:


Could you provide a list of what is required for Win98SE to support
48-LBA? Does this work for Win98 and WinME?

Simply a bios setup that supports LBA
Absolutely of course it works for WinME.

No fixes to the OSs? Is DOS mode safe, too?


It installs very easily, and you just install the drivers for it and your off and
running after you change hd connections. I prefer that All cd/dvd/rw croms, zips,
etc stays connected to the motherboard controller which all that stays the same but
without a hard driver to contend with you have an extra master position to run a key
...rom as master instead of slave.


The easiest way for anyone to get past large hard drive support is to
use the right bios setup - and the easiest way to do that for most
people is to use something simple like a controller card.

Agreed - it may be safer to do that, than to attempt a BIOS update,
and chances are the new HD will benefit from improved UDMA support
that a new card would add over whatever shipped with an older mobo.

I know Chris you don't have your newsreader setup like the rest of us,
and don't use OE, and so don't see your threads all in order

That's not the problem - I have exactly the same problems when I use
OE for the private MS newsgroups.

It's a server post aging thing that hits busy newsgroups in
particular; by the time I re-check news, some posts I couldn't get
last time, are already aged off so that I can't get them this time.

This is obviously worse when I come back to ngs after a period of
absence, as I'm doing now. Even three days without checking for new
posts is enough to disrupt continuity; all the existing headers marked
to retreive bodies find no bodies, and the new headers do not start
where the last ones left off.


Okay I see, and I'm sorry you have to contend with that.


"with the proper ingredients"

What proper ingredients?

Not much really, besides the Card, the other ingredients is simply have
no partition over 128 GB. It's FAT32 naturally. It needs to be under
128GB for its stock Scandisk and Defrag to work properly, otherwise it
won't

See, that 128G-per-volume limit suggests the OS is not OK with 48-LBA
after all. The assumption on which all this rests, is that the OS
never addresses the physical HD, but always addresses sectors relative
to the start of the volume, which as successfully been found via BIOS
rather than native OS code. Further, it assumes that the mechanism
whereby these calls are translated to raw disk addressing is
48-LBA-safe, presumably by using BIOS services.


Well I will leave that up to you with your ability to critique the details.


Given that Win9x redirects BIOS HD interrupt service vectors to its
own code (unless in DOS compatibility mode), I'm wondering whether the
OS can in fact safely find the start of volumes over the 137G limit.

Given that Win9x detects HDs and volumes via PnP rather than simply
inheriting these via BIOS, I'd wonder whether BIOS 48-LBA is in
effectwhen these are discovered?

Perhaps the key to this is whether the HD controller card's drivers
impart 48-LBA safety, and this might explain better mileage with
particular-brand add-on controller cards vs. integrated controllers?

As it is, I note DOS mode Scandisk can corrupt volumes that traverse
the 137G limit, BIOS 48-LBA notwithstanding.

For all of these reasons, as well as the failure of SP1 to truly be
48-LBA-safe, I'd really need to be COAB sure before suggesting the use
of large HDs with Win9x. Else it could be a case of "works fine long
enough to build up an irreplaceable data set, then eats the lot".

Well, I don't feel that way and have use controllers since day1
Actually Chris, I would not run any OS without one becasue of the remarkable
performance advantages above and beyond what it does for LBA support, but honestly
this must been seen with your own eyes to believe.

OTOH, you sound pretty sure of this, and it sounds as if you've
researched and tested it pretty thouroughly.

I always use them and would not be without one really. It's not that I have full
intellectual knowledge to impart all the specs about it, I just have everyday hands
on experience with what they can do. Now don't think this is just for W9x because
the across the board performance is for W2K & WXP too. As you know I triple boot to
each W98SE/W2Kpro/WXPro and the benefits are with all of them. Rather if it's with
an old P2 with and an off brand Biostar MB, whatever, or with the latest Intel D865
PERLL MB with it's excellent specifications, the promise controller card in each
situation will accordingly to each particular setup, percentagewise, DrAmAtiCally
boost the speed of overall real-time processing
of data across the board for system performance of each setup. Now the individual
RAM on each system is a separate issue, but obviously the older machines MB may not
be able to accommodate larger RAM but it dosent matter, but if it can then all the
better, but RAM is a separate issue and has it's own benefits to add to the mix wehre
of course the more RAM you have then the more performance gains that gives can you
independently, but my point is speaking to the controller cards advantages itself.

In each and every case I have used a controller card rather it's on a typical 5 year
old setup, or a brand new modern built hot setup, it's not the fact if it's an under
1 GHz processor or a 3.4 GHz processor, but proportionally I can conservatively say
that it increases overall system data transfer, including significantly accelerating
the load time of the OS - everything, and it's typically at least a minimum of 50 %
times faster over just the standard MB controller. I'll stop at saying 50% here but
personally in many situations I have seen much greater! I have tested the results on
many different setups and the increase has always been the same proportionally each
relative to hardware specifications of each setup, but the card will give
proportional advance to each, iow, the controller card does not depend on the hottest
setup to give benefits, it's backward compatible and will even help the 5 year old
setups to find new meaning in the term performance. Typically for a real world
example, if let's say you wanted to copy a bunch of large 1 GB files, or much smaller
it doesn't matter, or if you took a bunch of large files to zip - the time it will
take 'with, verses, without' using the controller card, will be anywhere from 50 -
100%+ ..faster to do that process or any process that crosses it's path. This
increase is realized with smaller MHz/GHz and even moreso with 2 GHz + machines.
Actually once testing 'before and after' the cards installation on a P3 500MHz, it
was 3x's (300%) faster afterwards, but with a 3 GHz setup it was 5x's faster.

..You're going to want to know what does it I know you Chris, and the technical
process of what happens when using a controller which causes this increase is similar
to what the Intel Application Accelerator software does, however comparatively is so
much better doing it with controller hardware verses IAA software - it's a no contest
in my book; Application Accelerator software applications are the pits anyway and
have proprietary pitfalls.

Simply said, with the controller card in the system and that process greatly reduces
the storage sub-system bottleneck that's always talked about, which enables faster
delivery of data from the hard drive to the processor as well as other system level
hardware. That's why W98 is such a dog when installed on a modern MB with 3 GHz + for
a processor, but with a CC installed, W9x never flew so fast in its life........!

Can you URL me? The URLs you mentioned above, don't really
hit the spot on this. For example, does the Promise site have
a tech FAQ on this?

Well no not that I have on hand actually nor have I ever really looked honestly, and
I never needed to know what I already know that I get since I always use them, but
looking at google just now here's one place for instance ...scroll down to where it
says "Conclusion" http://www.phoronix.com/scan.php?pag...item=107&num=2

..or you can look up all kinds of stuff on google it if you want:
http://www.google.com/search?hl=en&q...troller+review

...and then you can tell me the technical details since that is your 'cup of tea' to
be able to do.

I just know what I have always achieved with them not only on W9x, but on W2K and WXP
as well, and it's that which I can speak to.


Win98SE is not capable of addressing beyond "128G" but that this is OK
as it addresses each volume only from the start of that volume as
address 0, so that even if the volume passes the HD's "128G" (137G is
what I usually hear) limit, it doesn't matter.

The regular blurb about that issue is that 128 GB which is the true
windows binary limit - not the 137 decimal GB which many mention but
that's only a HD manufacturers marketing spec number used in decimals by
them so that their numbers look/sound bigger for marketing selling
points, but most importantly the answer is that since Windows and
Windows Utilities uses binary calculations, and so that 137 converts to
128 GB only. If someone was to try a 137 figure they would find it would
not work and don't try it with a good setup, only go by binary numbers
therefore the 128 GB is the true number and what should be used when
working with Windows Utilities.

OK, I was wondering about that. Often 1G = 1000 x 1M, or 1000 x 1000
k, whereas in binary it's 1024 x 1024 x k. Seeing as there are no HDs
between 120G and 160G, it doen't matter at that level, but kicks in
when you have to select max safe volume sizes. It may also be that
128G is where FAT32 changes cluster size to the nextsize up.


Ah, I know you like those specific technical details. You have a very good brain.
Here is a quote what your own MVP Ron Martell said about that and I will paste in his
exact quote:

Windows and Windows utilities report hard drive capacity in binary
gigabytes because binary values (1 kilobyte = 1024 bytes [2^10], 1
megabyte = 1024 kilobytes [2^20 bytes], and 1 gigabyte = 1024
megabytes [2^30 bytes] have always been used for both memory and
storage related items since the earliest versions of DOS.

The problem you are encountering arises because Scandisk and Defrag
cannot cope with drives that have more than 4.1 million total clusters
[2^22] and with a maximum cluster size of 32 kilobytes [2^15 bytes]
this equates to a maximum partition size of 128 binary gigabytes [2^37
bytes].


I don't see myself running Win9x on new PCs in GUI mode, but I would
like to use DOS mode Scandisk to maintain FATxx volumes in XP (via
alternate DOS mode boot), as I don't trust ChkDsk / AutoChk.


Yep I do that too, I do my work on my 2k & xp partitons from 9x ...of all things


So far, my mileage has been poor there. You mention WinME's Scandisk,
but does WinME have a DOS mode Scandisk, and if so, is it safe?


I have WME, but rarely use it.
I could boot it and tell you but really I'm not the one to ask about WME stuff though
since I rarely use it.


The systems I will be using this on, will be...
- 7.99G FAT32 C:

Ahh, me too, I make my OS partitions at exactly 7,185.3 MB which will show up as
exactly 7 GB when looking in My Computer (it's my lucky number) I do my under 128
GB at exactly 121GB which is exactly 124,597.9 MB ..heh who says I'm not fussy.
Actually snapping a quick screenshot my W98SE My Computer folder looks like this:
http://img423.imageshack.us/img423/8099/mcinse3ui.png ..and booting to my WXP the
same screen there looks like this:
http://img423.imageshack.us/img423/8594/mcinwxp0pr.png
Shown is I only have one 250 GB plugged in at the moment, but the other 250GB HD sits
right next to it and I will occasionally plug it in and do a disk to disk using
DriveImage, and then unplug it, since there's no sense in running both in my book.

Rick


- extended for rest of HD, as...
- 2G FAT16 D:
- laaargeG FAT32 E:
- 2G FAT16 or 7.99G FAT32 F:

That E: will be 128G on 120G HDs, but 128G on 200G HDs.



---------- ----- ---- --- -- - - - -
Don't pay malware vendors - boycott Sony
---------- ----- ---- --- -- - - - -

Hi PCR,

Yes Maxtor does have their filenames of thhings a bit different, and that's why I
mentioned before he could just adlib when it came to the filenames or wording they
put in their for their stuff, but otherwise it's all the same process.

have a good night

Rick



PCR wrote:
"Rick Chauvin" wrote in message
...
| Ken Dibble wrote:
|
...snip
| My third-party partitioner still couldn't see the full 250 GB, though,
| and ScanDisk was still running on boot-up. I followed the links
| supplied by Rick for the Promise site and was able to download a later
| Windows driver version than the one I had installed. (Apparently the
| link I'd followed, from the Promise homepage, to obtain an updated
| Windows driver was incorrect.) I ascertained that the card BIOS update
| that I'd downloaded earlier was the correct one. The link that Rick
| supplied for the Promise manual was a useful clue, although the manual
| was actually incorrect as to what I needed to do to flash the card
| BIOS.
|
| Yes the bios flash explanation on that particular pdf is a bit scant.
| Actually Promise made that card for some other manufactures too who
| just put their name on it, but for one, the Maxtor(Promise) version of
| the 100 card at Maxtors website gives a little more flash details if
| you were interested in saving the webpage as an mht for reference.
|
http://www.maxtor.com/portal/site/Ma...&downloadID=28
...snip

This one does it without making the flash floppy itself bootable. (I
would have deleted all the files from a Startup Diskette & put the two
Promise files there.)Also, it says to unplug the drives at the
controller end first. That's all that's different, that I see.

WELL, naturally, Maxtor has changed the name of the BIN from Promise's
B14.BIN to 267B43.BIN. Can't use the Maxtor download. It is only... BIOS
Version 2.01.0 (Build 43), where the Promise is... v2.20.0.14, you said.

......Quote............
After downloading U100B43.EXE insert a new formatted floppy into the A:\
drive. Double click on the U100B43.EXE icon, the files will
automatically load on to the floppy. The two files that load to the
floppy are PTIFLASH.EXE, 267B43.BIN. The BIOS flash does not run in a
Windows environment and is not bootable, you must first boot the system
to a Windows 95/98/ME startup disk. Power the system off, and disconnect
all Hard Disk data cables from the ultra controller. Insert the Windows
95/98/ME startup disk into the A:\ drive and power the system on. After
the system boots to the A: drive, remove the startup disk and insert the
disk with the U100B43 BIOS into the A:\ drive. Follow the instructions
below to perform the update.As a precaution It is recommended to
disconnect any drives that are connected to the Ultra ATA 100 PCI
Adapter Card before performing the BIOS flash.

Step-by-Step Procedure

After the system boots to the A:\ prompt and you have inserted the
floppy with the U100B43 BIOS into the A:\ drive.

From the A:\ prompt type PTIFLASH and press [Enter] this will execute
the Flash Memory Utility.

The Old BIOS needs to be backed up first. Press number 1 on the keyboard
this will display the Backup BIOS screen. Press the [Backspace] key to
move the flashing cursor to the left side of the field, then type
A:\OLDB27.BIN and press the [Enter] key. After a moment the Backup
Success box will display with the name of the saved file. Press [Enter]
to continue. From the flashing cursor after Please Select: press number
2 on your keyboard. The Update BIOS screen will appear. Press the
Backspace key to move the cursor to the far left of the field and type
267B43.BIN and press [Enter].

After the Update Success screen appears, remove the floppy from the A:\
drive and power the system off.

Power the system on to check the BIOS revision of the Ultra ATA PCI
Adapter Card. When the card is detected on the system, the new BIOS
revision will show as Ultra100â„¢ BIOS Version 2.01.0 (Build 43).

The BIOS has flashed successfully and you can power the system off,
reconnect cabling to the Ultra controller and restart the system. After
the system boots to the Windows desktop, replace the existing Ultra Host
driver with supporting 48-Bit version 2.0.0050.42.
......EOQ..............

"cquirke (MVP Windows shell/user)" wrote in message ...
| On Tue, 27 Dec 2005 05:12:41 GMT, Ken Dibble
|
| My third-party partitioner still couldn't see the full 250 GB, though
|
| Which partitioner?
|
| and ScanDisk was still running on boot-up.
|
| ScanDisk will run on bootup if one or both of two flags are set within
| the FAT of the volume. One flag is left set if file system operations
| are interrupted by a bad exit from Windows, which implies there will
| be logical file system corruption to clean up. The other is set
| whenever a disk access fails at the physical level, and this will
| prompt a surface scan of all volumes on that physical HD.
|
| Because Windows ALWAYS writes to C:, it's safest to perform this
| automatic Scandisk before Windows boots. In Win95xx/98xx, this is
| done by running DOS mode Scandisk before Windows loads.
|
| In WinME, there is no suitable DOS mode opportunity to do this, so the
| Windows Scandisk is run instead

That seems to be the case. So, why do you ask about it below?

| - and indeed, is constantly
| interrupted by potentially-destructive file writes, which maybe why MS
| "fixes" without prompting by default, in the hope that this will
| happen fast enough to complete before such interruptions.

There might be a Registry setting in WinME to prevent that. (I know I've seen one for... Defrag, at least... somewhere. Not in 98SE, though.)

|
| What a mess!

It is horrible to see those restarts! (I THINK I really need only disable McAfee to prevent most of mine. BUT I do a full cleanish boot, anyway, before a Scandisk, Thorough, or a Defrag.)

|
| ...for the Promise site and was able to download a later Windows
| driver version than the one I had installed. I ascertained that the
| card BIOS update that I'd downloaded earlier was the correct one.
|
| This is pretty messy and fragile-looking as well.
|
| Once I got the BIOS and driver upgrades installed, my Acronis Disk
| Director partitioner could see the full 250G
|
| OK - it sounds as if the partitioner relies on BIOS 48-LBA, which the
| previous setup was not putting into effect. Ew.

Right. The BIOS upgrade docs say it is for that purpose...
http://www.promise.com/support/downl...=bios&os=10 0
......Quote.............
Description - Supports 48-bit LBA for drives larger than 137 GB
......EOQ...............

|
| However, ScanDisk was still running on bootup. I downloaded
| and installed the Win ME versions of ScanDisk and Defrag, but
| this didn't stop ScanDisk from running.
|
| It wouldn't,

I think WinME's Scandskw.exe supposedly does it. See below.

| if that was the Windows ScandiskW.exe, because that is
| not what Win95xx or Win98xx uses in this circumstance. Does WinME
| have an updated DOS mode Scandisk.exe as well?

I think it does NOT, by the following article...

http://support.microsoft.com/kb/186365/EN-US/
Description of ScanDisk for Windows (Scandskw.exe) in Windows 98/Me

.......Quote..............
MORE INFORMATION
When you use the ScanDisk for Windows (Scandskw.exe) disk-checking and repair tool, you can check the integrity of your media (which includes hard disks and floppy disks), and repair most problems that may occur.

NOTE: Windows 98 and Windows 98 Second Edition start ScanDisk for MS-DOS (Scandisk.exe) automatically when the operating system is shut down improperly or your disk contains a critical error. Windows Me starts ScanDisk for Windows (Scandskw.exe) automatically when the operating system is shutdown mproperly or your disk contains a critical error.
.......EOQ................

WinME seems to expect it's Scandskw.exe will do it. HOWEVER, I'm sure, if you just throw that into Win98, it STILL will be SE's Scandisk.exe that will auto-run. And, if that one is disabled, I much doubt anything will run on an automatic basis! Something more would need to be cannibalized from WinME-- something to note those two flags you mentioned & direct Windows to run Scandskw.exe, IF it uses those same flags.

|
| So I deleted both partitions, rebooted, created two 116G partitions,
| rebooted, formatted the partitions, rebooted, and finally transferred
| a large multimedia file to the first new partition and ran it. I
| rebooted again and ScanDisk did not run!
|
| That's a sledgehammer was of resetting the flags in the FAT that were
| prompting Scandisk to run. I suspect if those flags ever get set
| again, it will be back to square one... and if DOS-based disk editing
| tools like Diskedit are not 48-LBA-OK, you may not be able to cleanly
| clear the flag in FAT in that way.

I speculate, as he had a non-48-LBA-OK BIOS when the large partition was first created, there was nothing any Scandisk could do to fix it. Now that the BIOS has been upgraded, it could be Scandisk all along was using BIOS routines-- but the wrong ones. Also, as you, yourself, speculated elsewhere...

.......Quote...........
Perhaps the key to this is whether the HD controller card's drivers
impart 48-LBA safety, and this might explain better mileage with
particular-brand add-on controller cards vs. integrated controllers?
.......EOQ.............

I do know it will replace MS's SMARTVSD.VXD with it's own. It also will add PU66VSD.VXD & ULTRA.MPD (which also descibes itself as a driver). Can one of these be used in DOS as well as in Windows?

|
| So it appears that all is well.
|
| Until the next bad exit, that is.

I don't know. That could have been a false alarm, having created a big one before the Promise upgrade.

|
| I suspect the situation is this:
| - BIOS is 48-LBA-OK
| - Windows HD controller drivers are 48-LBA-OK
| - Windows is not 48-LBA-OK (max volume size must be 128G)
| - DOS mode and tools (e.g. Scandisk) are not 48-LBA-OK

Could be the BIOS flash or added drivers makes it OK.

|
| So if Windows sets a flag to prompt DOS mode Scandisk, DOS mode
| Scandisk may corrupt the volume and/or be unable to clear that flag.
| Hence perpetual DOS mode Scandisk on every boot, forcing you to either
| suppress the automatic Scandisk via C:\WINBOOT.INI or C:\MSDOS.SYS,
| "just" wipe and re-create the affected volume, or find some
| 48-LBA-safe tool to clear the flag.
|
| This suggests the only way to maintain volumes over the 128G limit is
| to do so from within Windows, using WinME's ScandiskW.exe (assuming
| this is safe for 48-LBA). Perhaps this will clear the relevant flags
| in FAT so that the automatic DOS mode Scandisk goes away.

Well, he said (somewhere) ME's ScandskW.exe didn't fix it. Could have been unfixable at the time. Or did he think it would auto-run?

|
| This in turn suggests you must ensure Windows does not automatically
| write to these volumes, so that it is safe to start Windows and then
| Scandisk them from there. No Windows OS install on such volumes, and
| betware indexers and other underfootware trash.
|
| Thanks very much again. This was extremely helpful!
|
| We could do with a web page on 128G in Win9x; it seems possible, but
| looks like a bit of a minefield. I don't like dangling my data's
| health off a 3rd-party BIOS or driver version point revision.

Maybe. But who has suffered the problem after upgrading his Promise BIOS. Also, one must rule out the normal causes for such things, such as...

http://support.microsoft.com/kb/273017/EN-US/
ScanDisk Runs Even Though Windows Shut Down Correctly
........Quote.........
CAUSE
During the shutdown process, the contents of the virtual cache are written to the destination device. If the destination device is an Integrated Drive Electronics (IDE) hard disk, data may be written to the hard disk's onboard cache but not to the disk itself. When this occurs, the data is lost from the cache when the computer turns off.
........EOQ.............

|
|
|
| ---------- ----- ---- --- -- - - - -
| Don't pay malware vendors - boycott Sony
| ---------- ----- ---- --- -- - - - -

OK. But I'm not ready to sleep. OK, good night.

Wait! One thing... does auto-Scandisk work well & not loop on your systems that have a Promise card installed, provided you don't go over the 128 GB partition limit?


--
Thanks or Good Luck,
There may be humor in this post, and,
Naturally, you will not sue,
should things get worse after this,
PCR

"Rick Chauvin" wrote in message ...
| Hi PCR,
|
| Yes Maxtor does have their filenames of thhings a bit different, and that's why I
| mentioned before he could just adlib when it came to the filenames or wording they
| put in their for their stuff, but otherwise it's all the same process.
|
| have a good night
|
| Rick
|
|
|
| PCR wrote:
| "Rick Chauvin" wrote in message
| ...
| | Ken Dibble wrote:
| |
| ...snip
| | My third-party partitioner still couldn't see the full 250 GB, though,
| | and ScanDisk was still running on boot-up. I followed the links
| | supplied by Rick for the Promise site and was able to download a later
| | Windows driver version than the one I had installed. (Apparently the
| | link I'd followed, from the Promise homepage, to obtain an updated
| | Windows driver was incorrect.) I ascertained that the card BIOS update
| | that I'd downloaded earlier was the correct one. The link that Rick
| | supplied for the Promise manual was a useful clue, although the manual
| | was actually incorrect as to what I needed to do to flash the card
| | BIOS.
| |
| | Yes the bios flash explanation on that particular pdf is a bit scant.
| | Actually Promise made that card for some other manufactures too who
| | just put their name on it, but for one, the Maxtor(Promise) version of
| | the 100 card at Maxtors website gives a little more flash details if
| | you were interested in saving the webpage as an mht for reference.
| |
| http://www.maxtor.com/portal/site/Ma...&downloadID=28
| ...snip
|
| This one does it without making the flash floppy itself bootable. (I
| would have deleted all the files from a Startup Diskette & put the two
| Promise files there.)Also, it says to unplug the drives at the
| controller end first. That's all that's different, that I see.
|
| WELL, naturally, Maxtor has changed the name of the BIN from Promise's
| B14.BIN to 267B43.BIN. Can't use the Maxtor download. It is only... BIOS
| Version 2.01.0 (Build 43), where the Promise is... v2.20.0.14, you said.
|
| ......Quote............
| After downloading U100B43.EXE insert a new formatted floppy into the A:\
| drive. Double click on the U100B43.EXE icon, the files will
| automatically load on to the floppy. The two files that load to the
| floppy are PTIFLASH.EXE, 267B43.BIN. The BIOS flash does not run in a
| Windows environment and is not bootable, you must first boot the system
| to a Windows 95/98/ME startup disk. Power the system off, and disconnect
| all Hard Disk data cables from the ultra controller. Insert the Windows
| 95/98/ME startup disk into the A:\ drive and power the system on. After
| the system boots to the A: drive, remove the startup disk and insert the
| disk with the U100B43 BIOS into the A:\ drive. Follow the instructions
| below to perform the update.As a precaution It is recommended to
| disconnect any drives that are connected to the Ultra ATA 100 PCI
| Adapter Card before performing the BIOS flash.
|
| Step-by-Step Procedure
|
| After the system boots to the A:\ prompt and you have inserted the
| floppy with the U100B43 BIOS into the A:\ drive.
|
| From the A:\ prompt type PTIFLASH and press [Enter] this will execute
| the Flash Memory Utility.
|
| The Old BIOS needs to be backed up first. Press number 1 on the keyboard
| this will display the Backup BIOS screen. Press the [Backspace] key to
| move the flashing cursor to the left side of the field, then type
| A:\OLDB27.BIN and press the [Enter] key. After a moment the Backup
| Success box will display with the name of the saved file. Press [Enter]
| to continue. From the flashing cursor after Please Select: press number
| 2 on your keyboard. The Update BIOS screen will appear. Press the
| Backspace key to move the cursor to the far left of the field and type
| 267B43.BIN and press [Enter].
|
| After the Update Success screen appears, remove the floppy from the A:\
| drive and power the system off.
|
| Power the system on to check the BIOS revision of the Ultra ATA PCI
| Adapter Card. When the card is detected on the system, the new BIOS
| revision will show as Ultra100â„¢ BIOS Version 2.01.0 (Build 43).
|
| The BIOS has flashed successfully and you can power the system off,
| reconnect cabling to the Ultra controller and restart the system. After
| the system boots to the Windows desktop, replace the existing Ultra Host
| driver with supporting 48-Bit version 2.0.0050.42.
| ......EOQ..............
|
|
|