[Tip] How to Break the 2.2TB Capacity Barrier for XP/W2k/W2k3

Forward Compatibility - 16TB+ MBR GPT BIOS UEFI PCIe 5.0 DDR5 PAE 128GB Sapphire Rapids Discussions 2000 XP 2003 2009 7nm hexadecacore dotriacontacore




Happy St. Patrick’s Day!

I haven’t been around much due to my hectic schedule and insufficient time and have been working on something I’ve code named The XP Yeager Project: 16TB. It is the culmination of my research while working on another project.

Now this experiment was successfully done a week ago but I had to compile all my research together and just show you the aftermath of how it looks in Windows XP. I don’t believe anyone else has achieved this yet so it will be a first anywhere.

For awhile Microsoft had led us to believe you simply couldn’t access more than 2.2TB in Windows XP and we had to switch to a newer OS that supported GPT which could alleviate it.

Being a die hard user of XP and consumer of mass storage capacities I simply said FU I’m not upgrading without a fight.

For many years I lived with 2.0 TB drives out of convenience but over the years I migrated to 4TB and 8TB on Windows XP but I found the capacities still too limiting for my taste… :wink: 16TB was the next threshold and I wanted to achieve it.

After much experimenting and planning I devised a plan to get ahead of the tired waiting game from manufacturers to truly test the limits of XP before anyone else could and not wait 5 or 10 years from now to do it.

Most people have probably given up on using XP a long time ago by 2014 or earlier and one reason could have been this 2TB capacity limit. Some may find 2TB plenty of space today but I’m a power user so when there’s a will there’s a way.

Now there are large capacity SSDs such as the Samsung PM1633a which cost the price of a Mercedes Benz S Class when it first came out. Today it can be had for $11,000. I’m not Bill Gate’s kid or this would be chump change right? :smiley:

https://www.cdw.com/product/Samsung-PM16…-12Gb-s/4079174

But this would only get you to 15.36TB and shy of 16TB.
MZILS15THMLS-00003

Back in late 2008 the earliest inkling of the release of the very first 3.5" 2TB drive was found in this pdf file from December 2008.

http://web.archive.org/web/2009020611211…dwd6400aacs.pdf

WD20EADS was the Western Digital’s breakthrough capacity smasher and by then XP had become 7 years old and on the verge of feeling its age with Vista breathing down its neck for 2 years with no luck and 6 months before Microsoft made another attempt with Windows 7 to shut down XP once and for all.

Windows XP is now 16 years old so it’s fitting that I now bring you the first glimpses of what 16TB actually looks like working and remember this is highly experimental but I can confirm XP handles it.

Meanwhile I have another upcoming project breaking the 16TiB Barrier which comes to around 17.59TB and searching for ways to use larger capacity drives than this theoretical limit in XP.


[Chuck Yeager]
https://en.wikipedia.org/wiki/Chuck_Yeager

Some may wonder who or what is this Yeager? Chuck Yeager is a ground breaker and an American icon who was the first man to break the speed of sound at Mach 1. I’m here to try and break the capacity barrier in XP. It’s all for fun really and it just might give XP one more foot out of the grave. However this might not be so fun for Microsoft and drive manufacturers who really have made great efforts to stop it from happening and forcing people to believe 2.0 TB was all you could ever use on XP.

https://support.microsoft.com/en-us/help…arger-than-2-tb

Some of what I’ve done I can’t reveal at the moment as they will find a way to possibly disable it and thus wasted my time and effort and my plan would have failed. I’m hoping by the time Cannon Lake is released I can do another confirmation to test my technique hasn’t been broken. Intel and Microsoft has a habit of breaking things that worked fine before when releasing a new chipset. An example is Intel killing of eHCI USB 2.0 ports making installing XP, Vista, and Windows 7 more difficult. I might wait one more generation after Cannon Lake has been released before revealing how it’s done where by then it will be also much simpler and cheaper to attempt.

The longest wait will be the release of Tiger Lake or Sapphire Rapids which should be at the 7nm die shrink. By then there will be plenty of supply out in the public that they can’t stop my technique from happening and people can do it affordably.

But how many people will still be using XP in 2020-2022? Currently about 7% in 2018. How common and affordable will 16TB+ capacity drives be in 2022?

Early estimates it looks like 2020 when the 12th Generation Sapphire Rapids will arrive and 7nm with possibly 8C/16T CPUs should be common place. I’m betting it may be around 4 years time in 2022 if given their past estimates are always delayed before the secrets of The XP Yeager Project is fully revealed assuming it still works. Hopefully by then it won’t matter as the capacity limits have been reached for XP through my experiments and the significance of XP won’t be a concern for them to waste resources on disabling or preventing my methods from working and probably either Windows 7 will have fended off Windows 10 from overtaking it. But who wouldn’t like to use 16TB+ in Windows XP and using it on octacore CPUs?

Other future projects I’m working on is getting Windows XP to work on Intel Coffee Lake, Cannon Lake and beyond including AMD Bristol Ridge and Ryzen.

Now on to the DOUBLE money shot !!!





Edit by Fernando: Thread title simplified/shortened

@XPLives ,

We use XP in the office and because certain hardware I own only works with XP thanks to a culture in the healthcare industry not supporting their hardware that litterally costs 1000’s of dollars they would rather have you spend another five grand. In anycase to now use XP in a healthcare setting is a health privacy violation issue punishable by the state at another tens of thousands of dollars thanks to Microsoft no longer supporting it anymore and not patching security flaws. So after spending a little time I found there is a community based Win XP service pack 4. Not sure if you are familiar with it but if your using XP you should check it out.

The bigger issue was that most motherboard BIOS at the time that XP came out had a 2.2TB limit due to 32-bit addressing.

So even if your OS allowed larger drives you were still limited by what the BIOS could "see".

How you broke the 2.2tb limit?

Whom do you mean?
If you want to be sure, that the related person gets notified about your question, you should use the direct addressing way (= @nickname1 ).

This appears to be a 16TB Hard Drive connected thru a 4K Sector translating USB Enclosure or Adapter.

I have been working on Hard Drive enhancements for over 20 years, mostly for DOS and Windows 9x.
My current product supports 512TiB Hard Drives with either.

@rloew :
Welcome to the Win-RAID Forum and thanks for your contribution!
Regards
Dieter (alias Fernando)



@rloew

Close but no cigar.

I did not have any 16TB hard drive. As far as I know it did not exist at the time when I conducted my experiment. But if you knew of a 16TB hard drive for sale then in March 2018 let us know the Brand and Model to make this assumption. The earliest evidence of a 16TB drive was released in June 2019 from Seagate that diderius6 mentioned which was over a year after my experiment.

@nofearek9 , @diderius6 , @infuscomus



Hmm hard drive manufacturers took long enough to reach 16TB drives.

When I conducted my own test in March 2018 they had no 16TB drive available on the market and I beat them to the punch. It looks like Seagate might have released the first 16TB on June 2019. So I beat them by over a year and 3 months to see if XP could handle it in MBR. We can round down to 1 year since maybe it takes them 3 months to mass produce real stable drives and marketing it. If they can release a 17TB model, 17.5TB model, and 18TB model it would be interesting to test these. Or maybe an 18TB 2.5" SSD that you can modify firmware to only see 17.59TB capacity would be a more precise experiment.

Also I agree to avoid Seagate 3.5" drives. Choose Western Digital as they are more reliable and cheaper when you pry the drive out of external USB enclosure than buying bare drive. For 2.5" drives choose Seagate as they have bare 2.5" drives in them whereas WD does not and uses fused USB adapters so no way to hook them internally as SATA for data recovery.

So far no external 16TB enclosure available with a single hard drive. Only 10TB exist in large quantities.

12TB can be bought bare bones.
https://www.walmart.com/ip/Purple-WD121P…54d70ec77d6db14

14TB here:
https://www.amazon.com/Western-Digital-1…0/dp/B07KPL474H

16TB here:
Date First Available June 4, 2019
https://www.amazon.com/Seagate-IronWolf-…ctronics&sr=1-1


The exact limit should be 17.59TB for MBR but possibly 35TB is possible in XP without modding. 70TB could be done on XP with software modding the OS for 4K drives. Other OS like Vista, Windows 7+ cannot understand large capacity MBR drives without some sort of modding so will be capped at 17.59TB so will not see these 35TB MBR drives that XP can. So if they made a 17TB hard drive model I would buy that over the 16TB to maximize the capacity limit by 1 extra TB of space. Hopefully SSDs will come in 17TB capacity if HDD makers skip this nice round number to gain an extra 6.25% of capacity for legacy MBR usage.

If switching to Server 2003 32-Bit it can understand GPT which allows > 17TB drives to work while still maintaining XP compatibility and larger than 3.2->3.5GB memory access. So this is a better solution going forward as you will not have to patch each individual XP installation to access the drive data which if you have 20 XP computers this is a real headache or if you go to another location where they have XP but cannot install a GPT program onto it without permission the data is inaccessible.

Since Server 2003 have both MBR and GPT support you will not have to worry any further. 2003 will not understand 35TB MBR drives just like Vista+ OS. So it will suffer the same 17.59TB limit for 4K drives. So your 3TB drive is best used on 2003 Server unless you have only 1 copy of XP installed and only plan on using that drive then try GPT loader as an alternative.

From what I understand GPT Loader is only for internal drives > 2.2TB.

I seem to recall external USB drives of 3TB GPT are not supported? Maybe diderius6 or infuscomus has knowledge of this test?

That was the only inherent weakness I can remember about GPT Loader which is why 2003 Server is a better solution as it can access both internal and external GPT drives larger than 17.59TB or if you prefer 16.0TiB which hard drive manufacturers will never use for marketing it.


You can check my links here:
How to install Windows XP OS on AMD (AM3+ 990FX and AM4 X370 Ryzen 7)
2003 Server can see 16GB of memory no problem with a single 16GB DDR4 memory stick.
64GB will work on most motherboards if you have 4 memory sticks installed.

16TB MBR capacity will detect in XP as demonstrated by me here:
The XP Yeager Project: 16TB – Breaking the 2.2TB Capacity Barrier

But in order to really break the 17.59TB MBR limit on 4K drives will probably require a team of people with assembly and programming windows driver skills that to make a universal driver to do large sector translation so XP can understand and boot from 4K drives (internal and external) and reverse engineering hardware for a SATA adapter that would allow any XP computer to read large drives without modifying any XP files on the target machine so it will be seamless to the user.

Unfortunately no people are willing to assist in this project and do this kind of work for free so it is unlikely to happen as with everything it requires sacrificing time & money.



@rloew

Close but no cigar.

I did not have any 16TB hard drive. As far as I know it did not exist at the time when I conducted my experiment. But if you knew of a 16TB hard drive for sale then in March 2018 let us know the Brand and Model to make this assumption. The earliest evidence of a 16TB drive was released in June 2019 from Seagate that diderius6 mentioned which was over a year after my experiment.




You mentioned SSD Drives larger than 16TB in your original post. You did not specify how you did the experiment.
You could have either obtained such a Drive or faked it. I gave you the benefit of the doubt.
I have plenty of experience faking larger Drives to test code. I created a 128TiB Dos Partition almost 10 years ago.

@rloew



Don’t mean to correct you here but where exactly did I mention SSD drives larger than 16TB in my original post? I only mentioned a 15.36TB SSD - Samsung PM1633a which is way too expensive and not quite the 16TB which I shown in the picture. This was obvious this was not the SSD shown as it would have said 13.92TB in Disk Management not 14.5TB as you see it. Aside from that I mentioned only Bill Gate’s son could afford something like this and not someone like I or the average consumer. As it turns out the specs say it’s a SAS SSD drive which would have been incompatible with my experiment.

Obtaining such a 16TB 3.5" SATA hard drive that did not exist when I posted it would have required either access to such a prototype drive if it existed then which I do not know of or work in the hard drive manufacturing industry to get access to one. But the faking would be your software method of doing it which I don’t participate in. I used real hardware to achieve the results and I mentioned if you read the entire original posting from beginning to end that the reveal of how it was done may come when Sapphire Rapids is released or I’ll wait till when 5nm is first released. That would ensure the longest longevity of using my method with XP when storage capacity prices have dropped tremendously by then including 16TB drives for $130 or less. It’s also possible SSDs prices will drop dramatically by 2023-2024 that getting a true 16TB SSD would cost significantly less unlike the 15.36TB SSD from Samsung which costs a fortune.

Revealing my techniques sooner would be unlikely to prevent Intel and Microsoft from jeopardizing my method to work even on a later OS should they choose to employ their compatibility breaking methods. They have a habit of killing legacy OS advancements. I plan on having at least 5nm CPUs with 512GB of RAM using high capacity storage drives under XP or 2003. By then I’m sure Windows 10 will be solidly stable that they will not care to add in some OS patch to destroy backward compatibility for a twenty some year old OS.

Since I’ve been using DOS for nearly 40 years at the moment a 128TiB DOS partition would be kind of pointless even today or even a decade ago. Since NTFS read/write access in DOS would be its biggest hindrance. If you’ve come up with your own NTFS DOS and 9X Windows version capable of reading and writing to up to 256TiB NTFS DOS partitions that I would find uniquely useful for cross retro OS platform usage.

But at the moment also needing to patch every system would be a nightmare in disguise or if the partition got corrupted data recovery would be another obstacle when dealing with proprietary partitioning schemes. Since it appears you are in the business of making money via selling software solutions it would be unlikely I would choose this method to increase capacity with the risk of incompatibility or data loss. Due to the way legacy OS versions of DOS would be unable to access these larger partitions it would again be incompatible with my needs as well if I chose this path. You can’t force very old DOS programs to see these partitions you’ve created which I believe you will find out sooner or later why most people stick to much smaller DOS partition sizes for compatibility reasons.

However if ever consider offering to release something like this to the public freely I would support it and help assist in testing it. It’s possible had the Millennium Edition been successful, one more follow up Windows on DOS version may have included true NTFS read/write like Windows 2000, USB 3.0 drivers, 4GB+ memory support, multicore CPU support, and more modern graphics cards drivers support. But for the moment Windows Server 2003 32-Bit is the best substitute OS for XP.

There are two other possibilities.
A sparsely configured drive, which is another form of faking.
Or a JBOD Array which looks like a 16TB Hard Drive.
Since you refuse to explain your method, I can’t make any useful comments about it.

Incidentally, “faking it” is a perfectly acceptable method to test code. I debugged my MBR Extensions years before I was able to get a hard drive larger than 2TiB. I only got a 4Kn Drive a couple of years ago. It worked perfectly in DOS and 9x, having “faked it” several years earlier.

Selling software does not make it incompatible or less reliable. Remember Microsoft is in the business of selling software too. You are just cheap.
I’m not sure what legacy OS versions of DOS you are referring to. I modified Dos 7.10. Older versions have severe size limits.

@rloew



Respectfully, your individual speculations / insights are interesting and I wouldn’t say aren’t useful because it got you thinking of other possibilities than your first assumption and I’m certain you have others lingering in your consciousness. Since the original goal was to test 16TB in MBR in XP and no hard drive manufacturer had released a 16TB drive yet as far as I knew to confirm it would work and it was also intended to goad them into getting their act together as they had been promising these would have come out as early as 2016. Now in 2019 16TB drives are out and in the coming years we will see much larger capacities even 32TB could be on the horizon if doubling capacities can continue in years rather than decades.

As I already specified in the original post when I would possibly reveal the method used assuming it still works in the future. Just because you’re impatient to wait till Sapphire Rapids doesn’t mean I’m going to to reveal it sooner to fulfill your selfish curiosity. Now if you had your own 16TB storage device to test and prove go right ahead. Try writing/storing 16TB of data onto it and see if you can still access every single file on it. Or are you just using a small capacity drive and constantly overwriting the data?

Simulations are one thing but actually doing it is another. Imagine if all you did was simulations of going to the Moon thousands of times but never actually physically going there vs landing on the Moon itself. Which offers more proof and deserves more respect? I’m pretty sure simulations will not be 100% and when actually doing it you will find you missed or didn’t anticipate something.

I only politely corrected you and you go off on a tangent without admitting your error as I never said anything about larger than 16TB SSDs in the original post.

As for Microsoft I have every Windows OS running on my setup including DOS all the way to Windows 10. So you’re definition of “cheap” doesn’t apply when it comes to purchasing software and I have several dozen sealed copies of DOS and other rare items that you probably don’t have. I also have a huge collection of old MFM drives that cost a fortune when they first released. I’m currently using an i9-9900K octacore and what CPU are you using now that’s not “cheap” in comparison?

But if you think it’s wise to spend tens of thousands on a SSD drive is being not “cheap” then you’re probably loaded like Scrooge McDuck and this isn’t the forum for you to load your piggy bank. You have to be practical as well. Most people here don’t have an abundance of wealth to continually upgrade their systems or their OS. I just happen to prefer XP above all other OS but doesn’t mean I don’t use all other operating systems on the same computer. I would probably say XP and Windows 7 are the top two OS I prefer today if I had to choose but 2000 and Vista are probably the best favorites had all dependencies of XP and Windows 7 been patched in for compatibility.

We are just spending our own money and time to pursue a goal and trying to offer what information they can freely to continue using the same OS as long as possible on more modern systems. I’ve spent thousands on hardware and countless hours to the pursuit of XP working on modern systems even at the cost of a relative getting an illness because I was too focused on it. So I’ve had to divide my time recently this year to take care of them. What have you sacrificed and what makes you not “cheap”? If you could afford to spending tens of thousands on such a SSD drive then you certainly have a lot of expendable cash to call others cheap because they’re not interested in your software.

I never said anyone selling software makes it incompatible and I’m not sure where you logically concluded this. I buy software often and I said certain DOS software will not work with your proposed partitioning method. Even some software I tested corrupted a Windows 98SE partition when it was installing the program. Going to a much older DOS version didn’t have this problem. That’s 6.22 and earlier at least down to 5.0 and in some rare cases 2.1 and when the DOS partition is too large the installation program may not even proceed because it misidentifies the storage capacity. There are countless other reasons I won’t get into as this has nothing to do with the point of this thread.

Do you have thousands of DOS software games and applications that you’ve gone through and checked for compatibility from installation to running it to make that assessment?

Maybe you are here to just troll or trying to take advantage of people trying to sell something to someone they don’t want because you don’t agree with their opinion based on their own user experience of 40 years with DOS and Windows? If that’s the case then you deserve a troll award.

There is nothing selfish about curiosity with regards to your statement that cannot be verified because you refuse to explain how it was done.
I see no connection between 16TB Drives in XP and “Sapphire Rapids” CPUs.

I said you were cheap because you wouldn’t even consider buying my software. Yet you would be happy to use them if they were free.
I see people all the time who will spend plenty on their computers but won’t spend $20 for one of my Patches, preferring to do without, or spending a lot of time trying to find a workaround.

Obviously a simulation isn’t going to let me write 16TB of files. It would be intended to verify that the OS can handle the size.
I know it is not perfect. It would not have found the bug at 8TiB that would have occurred with Windows 98.
You never mentioned if you had ever written 16TB of files to that Partition. The screenshot shows it to be essentially empty.

What do you know about my partitioning scheme?

I may not have thousands of games to test, but my customers probably do.

I’m not trying to sell you anything you don’t want. I do mostly 9x not XP. I only joined this thread to provide a possible explanation for what you did.
I already dropped the subject of the 16TB SSD despite there being an announcement of a 30TB SSD Drive before your post.
I initially assumed that an affordable 16TB Drive had become available.
Despite what Microsoft says, what you did is not remarkable. You just tested closer to the ultimate limit.

Only 40 years?
I have 50 years of Computer experience.
Incidentally, DOS and Windows haven’t been around that long, so you don’t have that much experience in them.

@XPLives

I’ve heard CSM support is being dropped by intel next year, so if Sapphire Rapids is UEFI only it might become impossible to boot XP

@rloew



They say don’t feed the troll. Unfortunately you either have a comprehension skills deficiency or still didn’t read fully what I initially wrote. My delay in revealing my method has more to do with preserving the method than say Intel or Windows doing something to make it incompatible. If they don’t know what I did they cannot prevent it. And given each generation is typically 2 years on average and 4 years between a significant chipset jump to a more efficient CPU. There is still more enhancements that Intel/AMD has to do in a future chipset before I’m willing to lay it out there.

>>There is nothing selfish about curiosity with regards to your statement that cannot be verified because you refuse to explain how it was done.
I see no connection between 16TB Drives in XP and “Sapphire Rapids” CPUs.<<

Sapphire Rapids will be around 2023-2024 at earliest. By then hopefully memory limits will have hit 256GB-512GB or possibly 1TB with DDR5. I’m not so urgent to see USB v4.0 as when you have more cores you need more memory to go along with it.

Sandy Bridge MBs had hit 32GB with 4 memory slots. Stagnant growth since 2011 with 4 cores for the longest time until Coffee Lake thanks to AMD’s ass kicking with an 8 core. Now that Coffee Lake dropped their 8 Core we are still stuck with 64GB and DDR4. Although possible to get 128GB on DDR4 now. But 256GB-512MB would be the bare minimum I’d want the consumer desktop chipset to hit before revealing my method and by then 16TB-32TB hard drives and maybe SSDs would drop down to $130 price range by then. Now imagine 32 Cores running at 65 TDP, 1TB of DDR5 memory, 32TB SSDs for $100 a piece. That’s the time to reveal the method as by then Intel nor Microsoft can stop all the previous generation CPUs that already can hopefully still work with my method continue to work.

Even if MS or Intel decided to do something about adding some Windows 10 code or force BIOS manufacturers to do some disable trick to kill MBR from working on large capacity drives I would already be content with my 32 Core power efficient power house machine.

The other reason I will not disclose since your own interests are not compatible with mine so it wouldn’t matter.

>>I said you were cheap because you wouldn’t even consider buying my software. Yet you would be happy to use them if they were free.
I see people all the time who will spend plenty on their computers but won’t spend $20 for one of my Patches, preferring to do without, or spending a lot of time trying to find a workaround.<<

When did you discuss buying your software? You must be making giant leaps here. I said I would contribute my time to help create a better Windows ME assuming you had the necessary skills to do it. That costs me my own time and testing it for no money. So if my time is not worth money to you then you yourself are cheap. Offering my own time freely for the betterment of the OS and for others to use is the only way an OS can have more users which you don’t understand. The more users you have for a dead/dying OS the better pool of people you could potentially make money off of with other utilities.

You shouldn’t make a prejudicial statement like that. Those people who spend thousands on their computer probably aren’t looking to run Windows 98 on it and thus won’t be your target client. And while it’s true some may prefer to use a work around solution than pay for something that does the same thing that’s only human nature is it not? If you had to pay to take an uber ride to the store a half mile away or not pay and simply walk 15 minutes and get some exercise while you’re at it could you blame them?

Let’s consider in the future after it was completed. Say I did assist in testing your modified patch that would make Windows ME run natively multicore, use over 4GB of Ram, read/write to NTFS 16TB partitions, works with Intel USB 3.0, et cetera. Then you’ll have a new pool of potential customers using an old OS and they would be interested in other programs or tools possibly for it. If you knew how to create a game making engine for that OS, or make a universal graphics driver that works will all graphics cards then that would be another way to open the flood gates for new interest in Windows 98/ME usage. Since more users would be using their newer systems to run an old OS it’s likely the user pool could grow being these old operating systems didn’t require activation either.

Now let’s say you don’t do this and continue on your path of 50 years+. And let’s assume your pool of old OS users is currently 1000. It will only continue to shrink over time. People wouldn’t want to run an old OS stuck on a single core when they have 8 or 16 Core CPUs with 1TB of memory. So that leaves them with the choice of buying older hardware that works on such an old OS without needing your patch. So that does not in any way benefit you either. In the end you’ll just keep getting bitter that no one is buying your patches and blaming it on them being cheap. Your only new source of income would be from new computer systems running an old OS and if you can’t see how to bridge those two together it’s game over and the limited pool of users will eventually dry up.

Meanwhile XP on the other hand still has lots of potential in comparison. Thanks to multicore support, >4GB memory usage, large capacity MBR drives, and probably the largest software library to date continues to work even on the internet today.

If I were to buy any patches from anyone or any company I would have to have a use for it in my situation. It doesn’t matter if it’s $20 or $1 or $100. Even if it’s free it doesn’t mean I’ll use it if it’s inferior or I got something better. If your own patches start at $20 that’s your own pricing and prices should adjust to the market value of its worth. I never asked about any of your patches or whatever you’re trying to offer. I have no use for any 9X related patches unless it addresses something I truly need. I have so many older systems that work fine without patching and newer ones that work as well with my own technique. Currently it’s a dead OS for the time being which was why I outlined the limitations and what could be done to improve and modernize it along with even offering my own time to test. It was meant as an olive branch but I withdraw it now as I have less time now due to taking care of an ill relative.

>>Obviously a simulation isn’t going to let me write 16TB of files. It would be intended to verify that the OS can handle the size.
I know it is not perfect. It would not have found the bug at 8TiB that would have occurred with Windows 98.
You never mentioned if you had ever written 16TB of files to that Partition. The screenshot shows it to be essentially empty.<<

Without native NTFS read/write under Windows 98 why would 8TiB matter? At most 2TiB with FAT32 would be the most likely usage of large capacity partitions being useful in that OS if you want to maintain cross OS compatibility with Windows 98 and say 2000+.

Obviously I’ve written to the 16TB partition in XP or I wouldn’t classify it as being successful but if it had failed then I would have reported it or wasted my time taking snapshots. I had used extremely large video files stored on it for playback. This is why simulations cannot fully foresee every problem but are a cheap way to see what could happen. The only potential bug if I were to give you a clue is breaking the 35TB barrier will be harder to do with MBR compatibility still intact with Vista and later without patching. But it would still be possible strictly sticking to Windows 2000 and possibly as far as XP 2009 to get close to 70TB using MBR without OS patching with certain caveats. But the easier solution is to use GPT for drives larger than 35TB and use XP with a GPT Loader or 2003.

>>What do you know about my partitioning scheme?

I may not have thousands of games to test, but my customers probably do.

I’m not trying to sell you anything you don’t want. I do mostly 9x not XP. I only joined this thread to provide a possible explanation for what you did.
I already dropped the subject of the 16TB SSD despite there being an announcement of a 30TB SSD Drive before your post.
I initially assumed that an affordable 16TB Drive had become available.
Despite what Microsoft says, what you did is not remarkable. You just tested closer to the ultimate limit.<<


Your second and third sentences of your very 1st post gives the impression you are trying to sell something or market yourself. Your first sentence does not and addresses your assumption.

>>2nd - “I have been working on Hard Drive enhancements for over 20 years, mostly for DOS and Windows 9x.

3rd - My current product supports 512TiB Hard Drives with either.”<<


Care to offer any other people who’ve successfully used 16TB MBR partitions in XP in 2017? I’ve already tested beyond this limit before I posted the 16TB photo but never released those photos but briefly mentioned it in a log but did plan on posting it one day but never got around to doing it. I already know where the limits are for XP without needing to patch it to death. I value compatibility. I could have posted a 70TB photo that I created then in 2017 but that would be going closer to your route of faking it but with hardware instead of software which would still be unremarkable to me since I couldn’t actually write the full capacity but it looks like it might be possible soon. But my method would still be closer to how the actual drive would look for such large capacities beyond 16.0TiB/17.59TB.

But even your 20 years of HD enhancements would be unremarkable and let’s say we push this number to 25 years since you said over and that would put you starting around 1995 as an overestimate. Much earlier in the 80s we were already dealing with a park disk program made to park MFM drives before they were shipped and a custom BIOS for the IBM XT clone machine done. There were already other fine expensive software tools we paid for like Spinrite at the time. What were you doing then at the time to contribute to the software market?

>>Only 40 years?
I have 50 years of Computer experience.
Incidentally, DOS and Windows haven’t been around that long, so you don’t have that much experience in them.<<

No kidding Sir Holmes that’s why I stated the 40 years since 1981 began with the IBM PC-DOS 1.0. But I started with the Apple ][ DOS which was earlier in 1978. The most relevant experience is using DOS since 1.0 for the IBM PC side as I never truly stayed with Apple when they went the Mac line. Early Windows was not particularly useful to me until Windows 3.1 with multimedia as computers then due to their clock speed and memory limits ran too sluggish compared to DOS. When CPU processing powers finally caught up and Windows 98SE came around did MS make its final useful Windows on DOS version. Being a beta tester for software companies, Windows 95 was too damn buggy constantly crashing that it wasn’t until Windows 2000 did Windows truly shine as a stable enough OS.

Although I haven’t looked into it yet but today Windows ME could probably work with Intel USB 3.0 ports if it can use the Intel USB 2.0 drivers from XP. It’s obvious that Windows didn’t exist for 40 years yet but Windows will continue to age if you’ve used it since Windows 1.0 till Windows 10 and beyond. I’m still using DOS since 1978 and to this day on real and virtual systems and in about 10 years from now it will be half a century. So your 50 years will be even more meaningless again when you claim 60 years. Seeing how Windows versions have generally downgraded itself since Windows 2000, Windows 10 may be the end of the road unless a Windows XP 128-Bit version with full backwards compatibility to XP->W10 is released from Bill Gates returning to Microsoft from his charity work.

Now I don’t consider anything computer related prior to these years of any importance in my life. If you had said you had 60 or 100 years of computer experience it would still be as meaningless to me since anything prior to the Apple ][ would have no significance aside from the Apple 1 computer only for a trophy museum piece to play around with. Your 50 years could imply you got into computers in 1969 then maybe you had to wait a few years for Pong or Spacewar to have come around for any gaming. For you it may have a deeper meaning getting into computers that early on since you are not a gamer nor have a library of software to test as vast as mine nor the hardware. Also running a computer store longer than you’ve been in your business and testing hardware and software since the Apple ][ I doubt any of your own customers even have a software library comparable to mine or a hardware collection of vintage computers, hardware, and software that could only be obtained living through those years and having a massive warehouse to store them.

But I give you props for being into computers much earlier with EE experience and most likely a non-gamer focused on coding which would at least make you adept at Assembly which would be niche useful today. If you truly got into computers that early in 1969 then that would give you an uber nerd rating as I couldn’t have gotten into them at that time as they were uninteresting then. Had games not played a part in computers I think it likely that I would have gotten into low-level programming or program design to create my own given I had the patience to do it. I would probably have been more fascinated by the Moon Landing seeing it live on the B/W TV and chosen a career in NASA as an astronaut than punch holes into cards just to run code on a mainframe in '69.

Intel is already planning to break XP in 2020 so you won’t have to wait until 2023. In any case, I doubt that Microsoft or Intel would even care enough about what you did to do anything about it. They never did anything about my enhancements. They will break things on their own. The intensity of your arguments for delaying any details is just making me more and more suspicious of your entire story. Even more now that you upped the ante to 70TB.

My primary business is selling Patches. Giving them away to sell Utilities is not a viable business model in the foreseeable future.
Some of my Patches are needed to make using modern hardware practical with 9x. I was the first to break the 137GB limit.
Want more than 1GB of RAM or to use a SATA Drive. Two more Patches. AHCI requires my Driver.
My main utilities are Partitioning and Formatting tools to support the enhanced storage capabilities made possible by my Patches.

I have seen various people in the MSFN forums spend days or even weeks trying to work around a problem that one of my Patches would have solved in an instant. An Uber looks a lot better in a blizzard.

FAT32 supports 16TiB Partitions in XP, 2000, and 9x (with my TBPLUS Patch) using 4K Sector mapped USB Drives with standard MBRs.
The 8TiB issue is a bug in Windows 9x that I discovered, and patched, many years ago that appears at 1TiB in standard 512 Byte Sector Drives.
DOS can support FAT32 Partitions as large as 128TiB. I have the snapshot. This allows access to 3.75PiB.

25 years ago I was selling Amiga software through a local Amiga store. I designed KickWork and a Networking Package for the Amiga. Before that, aerospace and military firmware.

I watched the Moon Landing on three TVs, one on each network, that I had found in the garbage and repaired previously.

@rloew



It looks like you had modified your message once and I didn’t see your original response so I’m not certain what you wrote prior to the final revision.

You’ll be amazed what Intel and MS have done. As for XP the easiest one is breaking off XP compatibility by stopping to create XP drivers even for their iGPU which I can’t see that being a huge toll on their company to create such drivers or official Intel USB 3.0 drivers for XP/2003 or even open sourcing it. So don’t doubt for a second they have an agenda and trying to off Windows 7 recently is just another indicator to push everyone to Windows 10. While in your case 98SE and DOS are too old to matter and not really optimal to run on multicore CPUs for them to need to do anything that advances in technology just make it less likely you’re going to use a multicore CPU with PCIe slots with them.

But I will say Intel does not hold a monopoly on the CPU market as you know and AMD could continue with producing BIOS support on their MBs. Second Intel would probably be smarter to focus on Server MBs to force the NO BIOS MBs unless they have a deathwish. A lot of users who still use DOS through Windows 7 will want their BIOS support intact for legacy software so removing BIOS support on Intel based CPU consumer MBs would backfire. And if MB manufacturers see a loss of $$$ they will know it’s due to users wanting a BIOS based MB alternative which I would pay a little more for if they created two versions. I’m really waiting for the Intel USB 4.0, PCIe 5.0, and DDR5 all to be finalized in the next chipset. With at least 256GB/512GB memory support on 4 or 8 memory slots with XP running on it would be the icing on the cake.

As for the 70TB XP limit I will share a little more now since I’ve heard you passed away recently and I don’t wish to speak ill of the dead and I’m also grasping with my relative fighting cancer so hopefully your death resulted from natural causes from old age and went peacefully and not from some sort of suffering as your own thoughts could have led me to believe some of the anger and trolling could have stemmed from some of this.

Hopefully in some digital heaven this could be received by you from the Earth Realm to satisfy some of your curiosity. But since you did not live till 2023 the 70TB limit for XP will come from the 4 NTFS Partitions of 17.59TB that MBR can create. I documented with photos I took when I performed the test which would be getting close to 2 years ago. Times files. I might locate them at some point and upload them here for some future post. But again that would only be experimental proof of obtaining 70.36TB on XP and seeing it. But realistically only 35TB might be the true max that is usable without patching the OS which is why I would say 64TB is when I’d start using GPT for compatibility reasons with Server 2003.

I know your business is about selling software which is why I said we would be incompatible. Strangely enough if your death was approaching that you did not consider releasing whatever your research/source code/software written and contributed for the public freely upon or prior to your death. I think most would have appreciated that as a last gesture and maybe continued 9X support by the public would have continued. Perhaps one of your children or relatives you may have left a will to do this.

Sadly enough your websites are down so I have no idea whether any of your software was useful or if you had a 3rd party seller to continue releasing your software post death. I might have contributed some of my minimal funds to purchase something but alas my own loans and other financial priorities prevented it.

Amiga software I have an abundance now but time is always a factor to explore it all. I did read about your Amiga contribution but never tried any of that so who knows if somehow some of that software will find its way onto archive.org where it will be preserved and I can write a review on your work should I make use of it?

I do find it hilarious you watched the Moon Landing on three TV stations simultaneously on NBC, CBS, and ABC must have been quite the experience seeing how each TV station at the time reported it. My follow up question had you been alive would have been did you record all of it but most likely the answer you would have responded with would be no since VCR / Betamax was too new and it’s doubtful you had any way to store all that video footage unless you were working at a TV station.

I’m certain a large part of your knowledge will be lost.

Godspeed… in the Binary Realm Mr. Rudolph Loew.


XPLives

Regards,

anyone knows if replacing disk.sys and partmgr.sys from server 2003 makes data corruption after writing > 1 TB ?

i saw here : GPT disks on windows XP SP3 (partially solved) - Page 2 \ VOGONS
an user report that writing > 1 TB of data disk becomes unstable or corruption occur

anyone has info about this?

My Best Regards