Yes you are right, I'm sorry, Windows XP could do it using other software. Windows 9x/Me only use FAT32 and they needed PCI cards to use drives larger than 137 GB.
That was an actual limit, not an arbitrary one aimed at getting people to move to NTFS. Why Microsoft waited until SP2 to remove it in the XP disk driver is difficult to imagine (they also removed it in Win2K - perhaps in SP3, certainly by SP4).
Windows 9x/Me only use FAT32 and they needed PCI cards to use drives larger than 137 GB.
Given the necessary BIOS support (required early in the boot process to access any area of the disk beyond 137 GB - if you kept the required boot code below that point on the disk you could boot and then it became only a software issue to access the rest, as long as you avoided trying to do so in the native DOS environment) all Win98/SE/ME lacked was driver support as well. Third-party drivers were available to allow these systems to use the larger disks (Intel provided one, IIRC in its 'application accelerator' package; others were created by individuals, as described at msfn.org).
Win9x still had partition SIZE limits beyond which some of its disk utilities (such as Scandisk and fdisk) wouldn't work right, so using any single partition larger than 64 GB could be problematic.
With Advanced Format technology we can remove
Sync/DAM blocks, inter-sector gaps and 8 separate
blocks of ECC, improving error rate within the same
available capacity. Advanced Format Technology also
provides an increase in data integrity by providing a
more powerful error correction scheme using longer
ECC code words.
Many disk drive interface technologies already allow
for increased sector sizes. However, the 512-byte
sector has been the standard for over 30 years. As a
result, many points in a computer system (systems like
personal computers, servers, DVRs, PSPs, and cell
phones) have become inflexible and only work using
512-byte sectors. To maintain compatibility with these
devices, Advanced Format media emulates a 512-byte
device by maintaining a 512-byte sector at the drive
interface.
My initial read on the statements was that either the jumper or the software would enable the emulated mode. Reading back, it looks like the drive may run in emulated mode continuously.
What are your thoughts?
Also, if it does run in emulated mode, would you know of a way to disable the emulation to improve performance with systems that support 4KB sectors natively.
The statements in WD's first paragraph reflect many of the good reasons for moving to larger disk sectors. They all apply REGARDLESS of whether the drive interface emulates 512-byte sectors (as I believe all Advanced Format drives currently do).
My initial read on the statements was that either the jumper or the software would enable the emulated mode. Reading back, it looks like the drive may run in emulated mode continuously.
if it does run in emulated mode, would you know of a way to disable the emulation to improve performance with systems that support 4KB sectors natively.
I don't know that any such mechanism exists on today's Advanced Format drives (though probably eventually one will, if only for drives larger than 2 TB to circumvent legacy hardware or software that can address no more than 2^32 sectors on a drive), and I don't know of any reason why it would improve performance in any noticeable way if it did.
EDIT:
Under Windows 9x/Me, I remember only using drives over 120 GB through a PCI card because software could fail and corrupt data, it was just way to risky to rely on software because of potential for data corruption.
ALL software can 'fail and corrupt data', and I know of absolutely no reason to believe that the relatively straight-forward driver software from Intel or from competent individual creators required to support larger drives on Win9x is any more prone to such failure than the rest of the Win9x OS software (in fact, I suspect on average it's less so, and quite possibly less so than the firmware on an add-on PCI card - though my experience with them has also been good).
I don't want to detract from the more relevant issues Xaotic raised but I've been looking for years for software that would allow 160 GB or higher hard drives work on Windows 9x/Me without using a PCI card. I know there was pay-for software which may or may not load when Windows starts which is what scared me about it - but I never heard of software drivers that would allow use of 160 GB + hard drives, did you ever here of any?
WD Align is intended to be run on your Western Digital Advance Format drive WITHOUT a jumper installed across pins 7-8.
WD Align is only necessary for users who have:
* Installed Windows XP to a WD AFD
* Cloned a source bootable hard drive with any Windows OS to a target AFD using a cloning tool other than True Image WD Edition
* Created single or multiple partitions on an AFD using Windows XP
So by implication if the drive is "initialized" and formatted under Windows 7 - no pin is necessary as -bill was saying?
[As long as you take off the pin, boot into Windows 7, create a single volume, format, then never mess with partitioning/formatting of it in Windows XP.]
I don't want to detract from the more relevant issues Xaotic raised but I've been looking for years for software that would allow 160 GB or higher hard drives work on Windows 9x/Me without using a PCI card. I know there was pay-for software which may or may not load when Windows starts which is what scared me about it - but I never heard of software drivers that would allow use of 160 GB + hard drives, did you ever here of any?
I used a very slightly earlier version of http://www.mdgx.com/files/ATADRV98.EXE (examine with 7-zip or equivalent to read its .txt readme file) successfully on Win98SE for years (that .txt file does not claim it works on other Win9x systems). It's also available in http://www.mdgx.com/files/BHDD31.ZIP along with some other enhanced Win9x disk utilities. mdgx.com and mfsn.org are incredibly useful sites for people using older versions of Windows.
Thank you for that link - I have been looking for something like that for years.
And just to clarify if WD Adv Format jumper enabled the emulated mode - this is why it might have been questionable if using non-pin methods would have worked.
My initial read on the statements was that either the jumper or the software would enable the emulated mode. Reading back, it looks like the drive may run in emulated mode continuously.
If the option of Windows 7 is there, I understand now that avoiding the pin and booting into Windows 7 to do the partitioning and/or formatting is the way to go as long as you don't mess with partitioning/formatting inside Windows XP.
Can you not install something like Paragon Hard Disk Manager from 2011 or O&O partition manager from 2011 under Windows XP and partition/format without a pin to achieve the same effect as if you did so by booting into Windows 7?
I understand that if you don't have Windows 7, then you can still avoid the pin and boot with a Linux USB and use parted from Linux to set the drive up for use under Windows XP without the pin.
Can you not install something like Paragon Hard Disk Manager from 2011 or O&O partition manager from 2011 under Windows XP and partition/format without a pin to achieve the same effect as if you did so by booting into Windows 7?
Try it and see. If I had written a new partition manager that would run in both environments I quite likely would have made it default to using the partitioning approach that was native to the system under which it was running and offering the other alternative as an option (it should probably just ask directly which to use if running stand-alone from a CD, when it doesn't have a running system to give it a hint, and certainly could just always ask even if running under Win7 or XP).
I understand that if you don't have Windows 7, then you can still avoid the pin and boot with a Linux USB and use parted from Linux to set the drive up for use under Windows XP without the pin.
I'd try that and see as well: just because Microsoft has decreed that the new approach is what people should always use doesn't mean that the (g)parted developers necessarily agree with that (even if Linux is better able to cope with a mix of approaches on the same disk than Windows can).
I'd hope that if a disk is ALREADY partitioned then any partition manager operating on it that understood the existing format would attempt to follow it.
All right so take off the pin, boot into Windows XP and use the above mentioned programs to format the drive into one partition.
Then taking the same screen shot by EASEUS Partition Master as earlier would tell you if those utilities can be used under Windows XP to achieve what booting into Windows 7 could. Is that right?
All right so take off the pin, boot into Windows XP and use the above mentioned programs to format the drive into one partition.
Then taking the same screen shot by EASEUS Partition Master as earlier would tell you if those utilities can be used under Windows XP to achieve what booting into Windows 7 could. Is that right?
It'll certainly show you what they do by default, anyway. Poke around the options and you should be able to see whether any alternative behavior is available too (and if so you can try it too and see what it does).
Since we have established that for a dual boot scenario, simply boot into Win7 then partition and format in Win7 - remembering to never partition or format in Windows XP after that... since we established that to be a simple solution for dual boot systems, I was offering to do the same if Win7/Vista is not available.
I can do it by booting with ubuntu Linux, I can do it by installing the above mentioned modern software for Windows XP, I can do it using freeware gparted... but I need to know exactly what will tell me that I accomplished what I can accomplish by booting into Win7.
Of course the pin is available for WinXP but since the assumption that 'jumper or the software enable the emulated mode' is incorrect - then if we don't need the pin for WinXP - certainly the equivalent of partitioning and formatting in Win7 would be preferable. But if there is no Win7 then we need to know that "modern" software can do it when installed under Windows XP.
So tell me how exactly to establish that because I'm afraid I may do all the work and end up not offering sufficient proof.
Since we have established that for a dual boot scenario, simply boot into Win7 then partition and format in Win7 - remembering to never partition or format in Windows XP after that... since we established that to be a simple solution for dual boot systems, I was offering to do the same if Win7/Vista is not available.
I can do it by booting with ubuntu Linux, I can do it by installing the above mentioned modern software for Windows XP, I can do it using freeware gparted... but I need to know exactly what will tell me that I accomplished what I can accomplish by booting into Win7.
Of course the pin is available for WinXP but since the assumption that 'jumper or the software enable the emulated mode' is incorrect - then if we don't need the pin for WinXP - certainly the equivalent of partitioning and formatting in Win7 would be preferable. But if there is no Win7 then we need to know that "modern" software can do it when installed under Windows XP.
So tell me how exactly to establish that because I'm afraid I may do all the work and end up not offering sufficient proof.
Most owners of dual boots on the forum now have 17 words which answer the thread topic, they do not need to worry or be confused or know anything else about Advanced Format Drives. If they don't want to - they do not need to know anything about FAT or old IBM Thinkpads... 17 words:
Partition and/or format in Win7 remembering to never partition or format in Windows XP after that.
I am trying to do the same for most forum owners of single boot Windows XP. Of course they can just use the pin but whatever is tested to work on WD will hopefully also work for other brands using Advanced Format.
These drives are inexpensive and energy-efficient, run coolly and quietly, and are only marginally slower in random access (and comparable in sequential-access) performance than all but the fastest current 7200 rpm drives (WD blacks: I have one and am still interested in the new Samsung Advanced Format drive as a higher-capacity replacement for it). They'll make more than adequate system drives for almost anyone.
As I continue to ponder (there's no immediate rush) whether to replace my WD Caviar Black 640 GB (which significantly out-performs its specifications) with the new Samsung Advanced Format 2 TB drive it occurred to me that the above might have been more carefully worded.
The Samsung Advanced Format drive's performance is downright impressive for a 5400 rpm drive. The Seagate's (at 5900 rpm) isn't bad either, but even as a former Seagate loyalist I've become somewhat wary of their 3.5" drives over the past few years. The 5400 rpm WD drive's performance seems considerably less impressive and I'd be a lot more cautious about recommending it for use as a system drive, at least to anyone who's sensitive to system performance.
2. The WP article wasn't bad, though seemed ignorant of the fact that large-sector disks have been available for more than three decades. A bit strange, since it DID mention 520-byte sectors, which are usually found only on SCSI disks used in super-mini and mainframe systems - so its focus was only partially PC-centric (where, indeed, anything other than 512-byte sectors has until recently been very unusual).
3. The hothardware article was brief but excellent. It (and other benchmark results) make it difficult not to conclude that the WD Advanced Format drives' implementation of misaligned writes is seriously brain-damaged. Any sane implementation will read the two misaligned sectors of a small random write request (takes one seek plus on average 1/2 disk rotation to reach the sectors of interest), merge in the update, and then on the very next rotation (without having moved the disk head) update those two sectors (plus those in between if the write was somewhat larger than 4 KB), thus taking only a single additional disk rotation in addition to the half-rotation plus seek that would have been required even with optimal alignment (or on a 512-byte sector drive). Thus even if the head were already on-cylinder (no seek required) the absolute worst that could happen would be an average tripling of write latency, and with any seek at all the average degradation should be no more than about 50% even if the disk is short-stroked.
Can WD possibly be doing the R/M/W sequence only on the first updated sector and then performing another one serially on the second? Or allowing the disk head to seek elsewhere between the reads and the writes (if some other request is already present - though this would almost always be a very bad choice to make)? Could it have something to do with Windows' proclivity for issuing medium-sized writes from its system cache in 64 KB chunks and how WD chooses to order their execution on the disk when they're misaligned (though this would not explain some of the horrendous Linux benchmark results elsewhere)? Who knows.
4. The IBM article was characteristically competent and thorough. I was amused to see it use the same characterization of the WD pin 'solution' ("quick-and-dirty") that I had and suggest the same approach that I've been suggesting for mixing use of the same Advanced Format drive by both newer and older systems (using traditional, rather than new, partitioning mechanisms to align partitions appropriately - though it failed to address the idiosyncrasies caused by the presence of EPBRs with logical partitions). It also failed to note the possibility that older partitioning utilities (Partition Magic being a known offender) might encourage you to 'fix' and instead trash your partitions if you instead attempted to achieve optimal alignment by manipulating partition start-points directly (or indirectly by creating a non-traditional 'cylinder' size) - same comment applies to the formortals.com article - but since that was not in the context of sharing the disk with an older Windows environment it can't really be faulted for that. It does seem to state that when the article was written 9 months ago by default even the newer Linux utilities used traditional partition alignment (at least on MBR-partitioned disks), though that may have been in the process of starting to change.
5. You were indeed wise to preface the above with 'FWIW', since the ironically-named "The Facts" article contains some glaring errors - whether because the author misrepresented what he was told or because some of the information that WD provided was garbage and he didn't know enough to recognize it as such.
a) "Western Digital claims that this process (the required R/M/W sequence when writes are misaligned) doesn't impact performance" - hogwash: it may not (and indeed should not, though the benchmarks above say otherwise) TYPICALLY impact performance SIGNIFICANTLY, but beyond any shadow of a doubt it impacts it.
b) "Western Digital also explained that cache sizes on a hard drive will always remain relatively small versus SSD capacities because on sudden power loss the drive needs enough time to dump the cache to the disk’s platter before it starts to slow down and lose synchronisation." Hogwash again: there's no way in hell that a disk can write back many megabytes of cached 'dirty' data, often to hundreds of different locations on the drive, when power fails - it pretty much guarantees that it will finish writing the current sector being written, and may even try to complete the rest of the write request if it's not too large and resides contiguously on the platter, but that's all, so this has nothing whatsoever to do with limiting the cache size (the cache is mostly used for read-caching and for accumulating writes that can be written back at the disk's leisure with no guarantee of completion if power fails).
c) "Windows XP and other older OSes work by accessing specific disk sectors, not by using 'atomic writes' that work on byte-level data placements that we see in Windows Vista, 7, Server 2008, MacOS 10.4, 10.5 and 10.6 (Tiger, Leopard and Snow Leopard) and recent Linux kernels from 2.6.31 onwards." Hogwash: ALL modern systems 'work by accessing specific (whole) disk sectors' - that's the disk interface that the disk presents, after all. As described above the disk at least attempts to make the write of any individual sector 'atomic' (i.e., if it begins it will be completed), and higher layers in the system CAN make multi-sector and/or non-sector-aligned updates 'atomic' by using transactional mechanisms (which, of course are still BASED on whole-sector disk updates) - though no Windows or Linux system does, or ever has, done so by default (NTFS has since the days of NT used a transaction log to log metadata updates, as do ext3, ext4, and ReiserFS, but none log user data updates by default to make them 'atomic'), and I seriously doubt that Apple's do either.
d) "WD has created an application to run in the background that will avoid the slow-downs that people have been reporting." The WD alignment software is not something which dynamically adjusts disk requests to better alignment "in the background": it is run to relocate existing partitions to better-aligned positions on the disk, and woe betide anyone who thinks they can be accessing the disk while the WD software is running (whether 'in the background' or not).
e) "Certain types of hard disk software needs to support the 4K Advanced Format" - well, no, it doesn't: it will work just fine save for any potential small-random-write performance degradation. There's a significant difference between 'needs to' and 'ideally should'.
f) "We asked about JBOD setups, and Western Digital claimed these only require the jumper workaround, as the whole drive is used as a single partition." This will likely come as news to a good many JBOD users, who partition their JBOD drives just as they would any others.
Given that they so obviously had no clue what they might be looking for (or at), and seem to have screwed up at least some of their benchmark tests as well, the fact that their 'conclusions' were laughable comes as little surprise.
I guess, though, that one should be thankful that only one of the five (edit: whoops, six) hits that you came up with was so abjectly incompetent: that's far better than Sturgeon's Law would have predicted. Thanks for the others - they were good reading.
Western Digital Advanced Format technology hard drives use sectors with 4,096 bytes of user data. They are not optimized to be formatted under Windows XP or earlier operating systems. If possible, partition and/or format WD Advanced Format drives under Windows 7/Vista before using them under Windows XP. (Remember not to partition or format them under Windows XP after that.)
If you cannot partition/format them under Windows 7/Vista before using them under Windows XP:
• For a single partition: you have the option of placing a jumper on pins 7-8 which then allow these drives to be optimally used by Windows XP. [Once again, this is unnecessary if the drive was previously partitioned/formatted under Windows 7/Vista.]
• For multiple partitions under Windows XP: use WD Advanced Format Hard Drive Utility http://www.wdc.com/global/products/features/?id=7&language=1 [Once again, this is unnecessary if the drive was previously partitioned/formatted under Windows 7/Vista.]
Drives made by other manufacturers using Advanced Format technology (such as Seagate SmartAlign drives) do not use pins or software. If possible, they should also be partitioned and/or formatted under Windows 7/Vista for later use on Windows XP.
For single partition Advanced Format WD drives under Windows XP, if you place a jumper on pins 7-8 and boot into Windows XP > Control Panel > Administrative Tools > Computer Management > Storage > Disk Management >
A Wizard will pop-up in Windows XP to initialize the disk. CHECK the disk to be initialized but UNCHECK the disk when asked to convert it to a dynamic disk. Then right click on the Disk > New Partition... > Next > Next > Next > Next > UNCHECK: Perform a quick format > Next > Finish
So if you get a drive that appears to be working, make sure it doesn't fail the extended test like this:
Other manufacturers have their own utilities which also have extended tests. They usually need to run overnight since they take many hours to complete.
If you have no access to Windows 7/Vista for partitioning/formatting, unfortunately, even modern partition software installed under Windows XP may not be able to correctly partition/format Advanced Format Drives.
However, GParted can. Go to http://sourceforge.net/projects/gparted/
and get the GParted iso. Use it to create a boot CD. Try to only have the Advanced Format Drive connected to the system when you boot with it.
- a drive that came back from the manufacturer could possibly have a file system on it already
- if so delete that partition ... highlight partition ... partition > delete > edit > apply
- if it is new or has no partitions then you will need to create a partition table
- CRPTN - select the unallocated portion of the disk (all of it)
- CRPTN - partition > new ... a dialog box appears
- CRPTN - create your partition(s) as NTFS... assign the partition name
- CRPTN - if you create only 1 partition the above will be the only 2 modifications you would need to make
- CRPTN - notice that "align to" is set to MiB ... there's your alignment
- CRPTN - click the add button > edit > apply
- FMT - highlight the partition you just created
- FMT - partition > format to > NTFS > edit > apply
- CHK - to check that all is well from within GParted ...
- CHK - highlight the partition you created ... right click and choose information
- CHK - If you see that "first sector" is 2048 (or any number divisible by 8), then the partition is aligned.
- CHK - If the first sector is not divisible by 8 then something has gone wrong and such unaligned partitions on Advanced Format drives will perform at reduced speeds.
[Note: If you used a jumper for pins 7-8 on the Western Digital drive, then a subsequent examination of the partition by GParted would show the first sector of the partition as 63. But actually the first sector would be 64 (because of the jumper) and would be aligned.]
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
