Windows 7 32-bit with full 4 GB or 8 GB RAM support


Recommended Posts

Are you still talking about PAE? Reading this thread it seems to me that nobody denied the existence of PAE or its ability to extend memory usage in a 32-bit OS under the right conditions, but rather that it isn't practical for consumers using standard consumer hardware because of incompatibilities.

Actually, ShamRocker1 has repeatedly stated that it isn't possible.

Because I can. LoL

My point was what were you trying to say when you did that? I see that you highlighted the word patch, and this was after you asked me what patch I was talking about. Have you actually READ the first post of the thread?

I do recognize the PAE aspect. There is nobody who has given proof of this working. So, I say BS. Yes the concept is there but no actual proof. Good enough for you?

That's better. At least you aren't now repeatedly saying "it won't work" without saying why you think that. Most of your comments in this thread have reminded me of those people who will stick their fingers in their ears and say "la la la la la" when someone tries to tell them something they don't want to hear.

I'd be happy to test it, but my only 32-bit system running 7 only has 1GB in it, and to try to put more in would mean shutting down my main system and raiding the RAM from it.

I do recognize the PAE aspect. There is nobody who has given proof of this working. So, I say BS. Yes the concept is there but no actual proof. Good enough for you?

Right, so that since PAE does work, and since the server editions of Windows (which are virtually identical to the client editions) have fully supported it since 2000, how can you completely discard the possibility that it can be enabled in the client edition as well? Do you have some insight into how memory is initialized in the two editions to make this claim?

Right, so that since PAE does work, and since the server editions of Windows (which are virtually identical to the client editions) have fully supported it since 2000, how can you completely discard the possibility that it can be enabled in the client edition as well? Do you have some insight into how memory is initialized in the two editions to make this claim?

Why do you keep avoiding my question? SHOW ME PROOF! Otherwise go away. TY

Why do you keep avoiding my question? SHOW ME PROOF! Otherwise go away. TY

Well, someone wrote the patch, so presumably they have run it, as have various people who are posting about it, as well as other sites (such as the one linked earlier, dating back years) that talk about how the restriction works from a technical aspect.

You, on the other hand, simply proclaim that it's "BS!" You need to get off your high horse, and recognize that it is very possible that the patch can work, especially when you consider the technical side. It's not some magic feat we're talking about here. Since you have provided exactly zero evidence or explanation for why you believe it simply can't work (a claim you have continually made), while others have theorized that it can and given some technical background as to why, I hardly think you are in any position to tell people to "go away." The truth is that you have no idea at all.

Well, someone wrote the patch, so presumably they have run it, as have various people who are posting about it, as well as other sites (such as the one linked earlier, dating back years) that talk about how the restriction works from a technical aspect.

You, on the other hand, simply proclaim that it's "BS!" You need to get off your high horse, and recognize that it is very possible that the patch can work, especially when you consider the technical side. It's not some magic feat we're talking about here. Since you have provided exactly zero evidence or explanation for why you believe it simply can't work (a claim you have continually made), while others have theorized that it can and given some technical background as to why, I hardly think you are in any position to tell people to "go away." The truth is that you have no idea at all.

Show me PROOF. All you have done is spread a bunch of crap and have nothing to back it up as well as rework the dialog to suit your needs or agenda.

Most probably would LIKE to go to 64 bit, BUT 32 bit ONLY CPU's still do exist. And obviously 64 bit wont install if the CPU is only 32 bit. And the only way they'll be able to get out of 32 bit, is if they update everything. So, it'll support it

That is true but by the time windows9 comes out, i dont think there would be 32bit only CPUs about... @ the moment i cant see this happen cause and with windows8 ment to be 2012 i would see 99% of the market using a 64bit CPU and with 50% of those people using a Quad Core or higher..

Show me PROOF. All you have done is spread a bunch of crap and have nothing to back it up as well as rework the dialog to suit your needs or agenda.

Of what? I haven't claimed that the patch works, only that it easily could. I have no interest in running it. There's nothing for me to "prove."

You on the other hand keep insisting that it simply can't, with no explanation why. At one point you even claimed PAE doesn't work. You're just being silly.

As for me "spreading a bunch of crap," you'll have to back that up with quotes. Actually, don't, as I no longer have any interest in talking to someone that willfully ignorant.

Of what? I haven't claimed that the patch works, only that it easily could. I have no interest in running it. There's nothing for me to "prove."

You on the other hand keep insisting that it simply can't, with no explanation why. At one point you even claimed PAE doesn't work. You're just being silly.

As for me "spreading a bunch of crap," you'll have to back that up with quotes. Actually, don't, as I no longer have any interest in talking to someone that willfully ignorant.

Go back through the posts.

I have a 32 Bit PC ? 3,4 GHz, single core. And I think it's running fine and fast enough for me. I can't install Windows 7 64 Bit.

I have installed 4 GB of RAM (4x1 GB) ? but with Windows 7 32 Bit I only can use 3.25 GB of RAM. What for I have paid for 4 GB RAM - when I only can use 3.25 GB ?

If I had paid for a 1 terabyte hard disk and could only use 750 GB I also would be angry.

I borrowed from a friend 4x2 GB RAM modules ? and the test shows that I can use all of this 8 GB RAM.

OK ? you are right: With the patch I still have a 32 Bit operating system and any single application can only use 2 GB of RAM. But, hey, Windows 7 is a multitasking system. So I can run several applications to use all of my 4 GB RAM.

And it is even possible that specific applications can use more then 2 GB of RAM. This feature is named "AWE" ("address windowing extension"- see wikipedia http://en.wikipedia.org/wiki/Address_Windowing_Extensions). But this is a Windows API ? so the software must support it.

With the borrowed 8 GB RAM I do this test:

I installed "VMware Player 3" and created a virtual machine with 5 GB of RAM. And this works.

And about driver stability: All drivers that are included in Windows 7 are WHQL certificated. So all pointers within these drivers must be 64 bit able. That's because DEP (data execution protection) and the NX bit. So this driver works fine and with no problems ? even in a 32 bit enviroment. If you use "not WHQL" certificated drivers it's your own risk. Windows 7 warns you to use them. And in Windows 7 64 bit it is even not possible to use "not WHQL" certificated drivers.

I have a 32 Bit PC ? 3,4 GHz, single core. And I think it's running fine and fast enough for me. I can't install Windows 7 64 Bit.

I have installed 4 GB of RAM (4x1 GB) ? but with Windows 7 32 Bit I only can use 3.25 GB of RAM. What for I have paid for 4 GB RAM - when I only can use 3.25 GB ?

If I had paid for a 1 terabyte hard disk and could only use 750 GB I also would be angry.

If you brought a pre-made system with that much memory and 32-bit Windows preinstalled, then you would have a point. That would be a ripoff. If you go and buy Windows yourself though, then it is your own responsibility to ensure that it is compatible with your hardware. The restrictions in Windows are widely published.

And about driver stability: All drivers that are included in Windows 7 are WHQL certificated. So all pointers within these drivers must be 64 bit able. That's because DEP (data execution protection) and the NX bit. So this driver works fine and with no problems ? even in a 32 bit enviroment. If you use "not WHQL" certificated drivers it's your own risk. Windows 7 warns you to use them. And in Windows 7 64 bit it is even not possible to use "not WHQL" certificated drivers.

Well, if you go back in time, Windows has actually shipped with several drivers that are not PAE-aware and would crash. In Windows 7, who knows, but it's not impossible since the scenario in question would never actually occur unless someone hacked the OS.

The real issue here though, is that most people end up also running drivers that aren't bundled with Windows. This is where the problem could theoretically occur. Windows runs on so many systems that they have to make sure it's as stable as possible for absolutely everyone. This happens at the expense of people like you, but that's the price.

And in Windows 7 64 bit it is even not possible to use "not WHQL" certificated drivers.

They do not have to be certified, they just have to be signed. Anyone can go buy a certificate and sign their driver with it. There are no requirements regarding code quality. The signing requirements exists only to make drivers more identifiable.

A 32bit driver uses 32bit pointers, a 64bit driver uses 64bit pointers, a driver that runs under an OS with PAE uses 36bit pointers, the WHQL "sticker" doesn't mean it'll use all 3 dynamically.

There are no 36-bit pointers. It uses 64-bit addressing. All drivers that are correctly designed have to support this, which doesn't really mean much since there are thousands of devices out there of extremely poor quality.

The main thing to bear in mind with running a true x64 system is that the processes/applications themselves have a wider memory range. 32bit system with PAE will still restrict each process to 2GB. I grant you that many applications are still 32bit however with the push to 64bit we will start to see more and more apps take advantage of this.

I don't understand why someone would still push for a 32bit system, as 64bit systems are pretty new with relevant driver support.

...

There are no 36-bit pointers. It uses 64-bit addressing. All drivers that are correctly designed have to support this, which doesn't really mean much since there are thousands of devices out there of extremely poor quality.

Doh, of course, there's no 36bit data type. You'd just use a 64bit pointer and ignore 28bits of it.

And yeah, I've got a webcam that causes crashes whenever apps try to access it, and that's a using the driver that Windows automatically installs for it.

And in most CPU's made in the last decade, they aren't 32-bit either, they are 36-bit.

There are no CPUs with 36-bit registers. PAE doesn't change register sizes. It just adds another level of hierarchy to the page table structures.

A 32bit driver uses 32bit pointers, a 64bit driver uses 64bit pointers, a driver that runs under an OS with PAE uses 36bit pointers, the WHQL "sticker" doesn't mean it'll use all 3 dynamically.

There is no such thing as a 36-bit pointer.

And about driver stability: All drivers that are included in Windows 7 are WHQL certificated. So all pointers within these drivers must be 64 bit able. That's because DEP (data execution protection) and the NX bit. So this driver works fine and with no problems ? even in a 32 bit enviroment.

32-bit client versions of Windows always run with PAE enabled in the CPU. This is what allows hardware DEP / NX to work. This does NOT result in 64-bit pointers magically appearing, and has no impact at all on driver compatibility.

Allowing support for greater than 4GB of memory is what causes compatibility problems, not PAE or DEP.

If you use "not WHQL" certificated drivers it's your own risk. Windows 7 warns you to use them. And in Windows 7 64 bit it is even not possible to use "not WHQL" certificated drivers.

This is utterly false. 64-bit Windows does not prevent you from using non-WHQL drivers. It prevents you from using unsigned drivers. You can sign a driver without going through WHQL. How do you think 64-bit users install non-WHQL display drivers all the time?

Driver signing is done by the manufacturer and has nothing to do with driver quality.

For some background on the complications which resulted in >4GB of addressing using PAE being disabled for client editions of Windows:

http://www.microsoft.com/whdc/system/platf...PAE/PAEdrv.mspx

For some background on the complications which resulted in >4GB of addressing using PAE being disabled for client editions of Windows:

http://www.microsoft.com/whdc/system/platf...PAE/PAEdrv.mspx

Looks like it isn't limited in Windows 2000. Someone should give it a go.

This topic is now closed to further replies.
  • Recently Browsing   0 members

    • No registered users viewing this page.
  • Posts

    • KillerPDF 1.4.2 by Razvan Serea KillerPDF is a lightweight, portable PDF editor for Windows built for users who want full control without subscriptions, installers, or telemetry. It runs as a single executable, making it ideal for USB use and field work. You can view PDFs with smooth PDFium rendering, navigate quickly with thumbnails, zoom, and shortcuts, and reorganize pages using drag-and-drop. It supports merging multiple PDFs, splitting documents, and extracting selected pages. KillerPDF also allows inline text editing with font matching to preserve the original layout, plus annotations like text boxes, freehand drawing, highlights, and reusable signatures. You can search full text, copy content easily, and print documents with flattened annotations. Designed as a free and open alternative to bloated PDF tools, it works fully offline on Windows 10/11 x64. No runtimes install. Everything needed is inside the EXE (targets .NET Framework 4.8, which ships with every supported Windows release). KillerPDF key features: High-quality PDF rendering via PDFium Edit PDF text inline (double-click to modify text) Page thumbnails and fast navigation with zoom and shortcuts Merge multiple PDFs into one Split PDFs and extract selected pages Drag-and-drop page reordering Font matching to preserve original document appearance Text boxes for notes Freehand drawing tools Highlight overlays with adjustable color, size, opacity Undo actions and clear per-page annotations Create, draw, and save reusable signatures Click-to-place signatures anywhere Full-text search with highlighted results Drag-select or Ctrl+A to copy text Print with annotations flattened Portable single-file app (~10 MB) No installer, no admin rights required No account, no telemetry KillerPDF 1.4.2 changelog: What's new PDF form filling. Interactive PDF forms now render their fields (text inputs, checkboxes, radio buttons) as live controls. Fill them in directly and save — field values are written back into the PDF. PDF outline (bookmark) navigation. A new OUTLINES tab in the sidebar displays the document's bookmark tree. Click any entry to jump to that page. The sidebar auto-fits its width to the longest entry on open and can be dragged wider; switching back to PAGES snaps to the pages-mode width. Fixed Page rotation no longer reverts after saving. Rotations applied via the sidebar context menu now persist correctly through the save pipeline. Copied text words were out of order on PDFs where glyphs are stored in non-reading order (Issue #66). Text extraction now sorts words by position and uses a dynamic line-grouping threshold so both drag-select and Select All produce correctly ordered output. PDFs with malformed or non-standard XRef tables now open in read-only mode instead of showing "Invalid entry in XRef table" and failing entirely. Download: KillerPDF 1.4.2 | 6.1 MB (Open Source) Link: KillerPDF Home Page | Github | Screenshot Get alerted to all of our Software updates on Twitter at @NeowinSoftware
    • "...a low price of just $340..." I don't think it means what you think it means.
    • This Corsair Vengeance DDR5-6000 32GB RAM with RGB is a great deal for limited time by Sayan Sen Memory prices have been through the roof for a while, though it seems like things might finally be getting better. If you are in the market for one, then grab this Corsair Vengeance DDR5 32GB (2x16GB) DDR5 6000 CL36 kit with RGB for a low price of just $340 (purchase link under the specs table down below). The kit is compatible with both AMD and Intel systems as it supports both EXPO and XMP overclocking profiles, respectively. 6000 MT/s is often the sweet spot for many systems as it provides ample data transfer speed while still being on Gear 1 mode. This Vengeance variant has RGB so if you love bright setups with such lighting, this is a win-win for you. The technical specifications of the Corsair Vengeance memory kit are given in the table below: Specification Value Memory Type DDR5 Memory Size (Total) 32GB Kit Configuration 2 × 16GB Form Factor UDIMM (Desktop) Pin Count 288-pin Speed (Data Rate) 6000 MT/s Speed Rating PC5-48000 Tested CAS Latency 38-44-44-96 Voltage (Tested) 1.35V Performance Profile AMD EXPO & Intel XMP Heat Spreader Aluminum heatspreader Cooling Type Passive (Heatsink) Lighting Ten Zone RGB Software Support Corsair iCUE Get it at the link below: CORSAIR Vengeance RGB DDR5 32GB (2 x 16GB) 6000 CL38 – Gray (CMH32GX5M1E6000Z38): $339.99 (Sold and Shipped by Woot US, Fulfilled by Amazon US) This Woot deal is US-specific and not available in other regions unless specified. This is a first-party seller link (at the time of article publishing); ensure that you also purchase from a first-party seller link only. If you don't like it or want to look at more options, check out the previous deals that we have covered, OR you can also visit Amazon US deals page. Get Prime (SNAP), Prime Video, Audible Plus or Kindle / Music Unlimited. Free for 30 days. As an Amazon Associate, we earn from qualifying purchases.
    • The very fact that a TPM (v2.0 specifically which is part of the issue I suspect) is now a baseline for any supported Windows installation will naturally mean other vendors will start to leverage it as they know it'll be there. It's called progress, and it's always been the way. A TPM isn't a windows thing, it's just a module designed to securely store keys. Secure boot isn't a Windows thing (although MS are the TCA as I recall hence the upheaval this year as the 2011 certs expire), it's just a way to verify a bootloader is signed. Windows simply leverages them.
    • It's a local account with the ability to reset a password at a very base level. I really don't get the issue that gets whipped up around it But you do you
  • Recent Achievements

    • Proficient
      Eric Biran went up a rank
      Proficient
    • Dedicated
      Conjor earned a badge
      Dedicated
    • Week One Done
      Windows Guy earned a badge
      Week One Done
    • Dedicated
      Mark Spruce earned a badge
      Dedicated
    • Collaborator
      conkir earned a badge
      Collaborator
  • Popular Contributors

    1. 1
      +primortal
      479
    2. 2
      PsYcHoKiLLa
      244
    3. 3
      Steven P.
      72
    4. 4
      +Edouard
      66
    5. 5
      Skyfrog
      65
  • Tell a friend

    Love Neowin? Tell a friend!