Intel considers cutting ties with old tech support

Emre Çitak
May 26, 2023

Intel has put forth a proposal to streamline the x86 architecture by eliminating outdated features. The company's developer blog recently released a technical note that suggests the introduction of a new x86S architecture, aimed at simplifying the design of future processors and the booting process of personal computers.

The proposed change involves removing 16-bit and certain elements of 32-bit support from the hardware, resulting in a family of processors that would directly boot into x86-64 mode.

By bypassing the conventional series of transitions, which include shifting from 16-bit real mode to 32-bit protected mode and finally to 64-bit long mode, or even skipping the 16-bit mode altogether and directly entering the 64-bit mode, the envisioned processors would offer a more streamlined and efficient startup experience.

Intel has provided an accompanying 46-page technical white paper that delves into extensive details about the potential modifications. While some of these changes are quite drastic, their impact on the average computer user may go unnoticed – and that seems to be Intel's objective.

Intel old tech
Intel plans to remove old tech support from newer systems

Maximizing the modern software compatibility

The elimination of rings one and two, which are no longer utilized by modern software, aligns with historical usage patterns. In fact, the vast majority of PC operating systems rely solely on rings zero and three, with the exception of IBM's OS/2 and Novell Netware 4 and above.

Nonetheless, the absence of direct support for these operating systems on future hardware is unlikely to cause significant inconvenience, as they wouldn't even boot on contemporary UEFI machines.

While the removal of ring zero in 32-bit mode may prevent the execution of an x86-32 hypervisor, it is worth noting that the enhanced capabilities and increased memory provided by a 64-bit hypervisor make it a preferable choice. Although the x86-32 mode will not be completely eliminated, its functionality will be notably restricted.

However, it will still be possible to initiate an x86-32 operating system within a virtual machine, as these systems already rely on firmware emulation alongside emulated components such as graphics and network cards.

What does it all mean for newer systems?

Some people might be concerned about these changes, especially if they use older operating systems like DOS or Windows 9X. But the truth is, those systems wouldn't work on newer computers anyway. Intel is just making it official by removing the support for them.

Normally when you turn on your computer, it goes through different steps to get started. But with this new idea, the processors would skip some of those steps and go straight into the newest 64-bit mode. This could make things faster and more efficient.

So, what does all this mean for you? Well, if you have a newer computer, you won't even notice these changes. And if you use a 32-bit operating system, don't worry, you can still run it inside a special virtual environment. Intel just wants to make things simpler and faster for the future.


Tutorials & Tips

Previous Post: «
Next Post: «


  1. Jek they/them Porkins said on May 27, 2023 at 4:47 pm

    Intel did propose a new standard called IA-64. It also attempted eliminate “legacy” 16 and 32 bit code support. The beginning of the end for IA-64 was the cheaper and more compatible x86-64. IA-64 never caught on and was discontinued about 2 years ago.

    In my opinion this is just a sneaky anti free market attempt to regain their monopoly.

  2. Gerold Manders said on May 26, 2023 at 4:54 pm

    Is 64-bit preferable? Yes, it is. For big(ger) applications it is indeed preferable. For small and/or simple applications, 32-bit remains just fine. And as long as you do not envision your application to require more than 2 GByte of RAM at any given time, it is most of the time the faster version of the two.

    Do you compile software? If you do, there is quite some time difference compiling between 32-bit applications and 64-bit applications. 32-bit is always faster.

    A funny thing, Intel his the patentholder for the 32-bit part in the x86 architecture. AMD is the patentholder for the 64-bit part in the x86 architecture. As far as I understood, there is an agreement between these 2 companies, that Intel is not paying AMD for the 64-bit part and that AMD is not paying Intel for the 32-bit part.

    By scrapping (most of) the 32-bit section and older, would that mean that Intel has to start paying AMD for their 64-bit patents? Probably will be under the FRAND regulations. While these costs are (relatively) low, they are not zero.

    In any case, there is much legacy in the x86 architecture. And Intel is not wrong about simplifying it. Scrapping a lot of legacy, I envision that it would make x86 CPUs a lot more power-efficient. Probably much closer to ARM-based CPUs. Which would make a serious dent in sales of ARM-based laptops/desktops. Because the amount of software available for the x86 architecture is vast, even if all the 16-bit and most of the 32-bit software is scrapped. And there is much more expertise in writing software for x86 architecture. Don’t underestimate this either.

    ARM-based CPUs are efficient and at least Apple made their CPUs fast. But adoption of ARM is going very slow. If x86 becomes much more efficient without all the legacy, would that stop the adoption of ARM dead in its tracks? On desktops/laptops, I mean. ARM will remain popular in phones.

    Perhaps Intel should have proposed this 10 years ago, then x86 would have been more viable for phones or tablets too. Ah well, it is what it is.

Leave a Reply

Check the box to consent to your data being stored in line with the guidelines set out in our privacy policy

We love comments and welcome thoughtful and civilized discussion. Rudeness and personal attacks will not be tolerated. Please stay on-topic.
Please note that your comment may not appear immediately after you post it.