Surely the 486's that are still being made are for industrial control machines and such?

ElectroSoldier wrote on 2025-05-03, 23:44:

Surely the 486's that are still being made are for industrial control machines and such?

But are any still being made by anyone?

I'd like to see an example of a 486 of some form from the last decade

myne wrote on 2025-05-04, 03:38:

I'd like to see an example of a 486 of some form from the last decade

I believe 2014 date code ones exist but were a little poopy quality wise. (Lots of duds)

darry wrote on 2025-05-04, 03:03:

ElectroSoldier wrote on 2025-05-03, 23:44:

Surely the 486's that are still being made are for industrial control machines and such?

But are any still being made by anyone?

I seriously doubt it. Even for basic control systems technology has moved on considerably and given that its not cheap to make even a 486 anymore.

ElectroSoldier wrote on 2025-05-04, 14:24:

darry wrote on 2025-05-04, 03:03:

ElectroSoldier wrote on 2025-05-03, 23:44:

Surely the 486's that are still being made are for industrial control machines and such?

But are any still being made by anyone?

I seriously doubt it. Even for basic control systems technology has moved on considerably and given that its not cheap to make even a 486 anymore.

It will probably eventually be economically feasible to just emulate one on an SBC and find a way to interface it with socket 3, Pistorm style.

ElectroSoldier wrote on 2025-05-04, 14:24:

darry wrote on 2025-05-04, 03:03:

ElectroSoldier wrote on 2025-05-03, 23:44:

Surely the 486's that are still being made are for industrial control machines and such?

But are any still being made by anyone?

I seriously doubt it. Even for basic control systems technology has moved on considerably and given that its not cheap to make even a 486 anymore.

When did they quit making vortex CPUs? (Those are just a 486)

rmay635703 wrote on 2025-05-05, 02:05:

When did they quit making vortex CPUs? (Those are just a 486)

They didn't. But those are not "just a 486". They need to be clocked much higher to attain the speed of a 486.

rmay635703 wrote on 2025-05-05, 02:05:

When did they quit making vortex CPUs? (Those are just a 486)

The Vortex86DX and later models support the CMPXCHG8B instruction, which is basically what makes a CPU "Pentium class".

Quoting the Phoronix article:

The kernel patches would remove support for CPUs lacking TSC and CX8/CMPXCHG8B capabilities.

It's for this reason that the Vortex86SX (one model lower/earlier than the DX) cannot run Windows XP.

So basically, as of 2025, the Linux kernel is setting the same support baseline that Microsoft did with Windows XP all the way back in 2001.

It's fun to note that someone managed to hack Windows XP to not require CMPXCHG8B. But the practical use of that novelty hack is limited, since application and driver developers that targeted Windows XP and later would likely compile their code to make use of the CMPXCHG8B instruction even if the underlying OS wouldn't, so those would still fail to run on such older processors.

I also found this recent message in a Linux kernel discussion thread, which also mentions how Vortex86SX support will be dropped, but the Vortex86DX and later SoCs in the Vortex family should continue to be supported.

It's more annoying that Shift+PgUp and Shift+PgDown were dropped on text mode virtual consoles in 5.x

kagura1050 wrote on 2025-04-29, 09:54:

(I don't know much about i486 Linux other than opencocon, so if anyone knows of any other projects that run i486 with a new kernel, please let me know!)

AOSC Retro is one (releases are here), has been discussed a few times on this forum, and ArchLinux32 is another, although that has mostly non-graphics packages for its i486 build and the latest i486 ISO is from December 2022.

jakethompson1 wrote on 2025-05-06, 13:45:

It's more annoying that Shift+PgUp and Shift+PgDown were dropped on text mode virtual consoles in 5.x

Yes. Absolutely.

About the 486 support - yes it is symbolically sad, but modern linux cannot work on a 486 anyway, without some serious manual labour and trimmings, and without any graphics. Nothing "modern" (i.e. web browser based) will work properly if at all.

Kernel <= 2.4 with Xfree86 and accelerated drivers is the way to go.

Edit: Just started reading about AOSC retro in the link above - so far I like what I see there 😀

digger wrote on 2025-05-06, 10:46:

It's fun to note that someone managed to hack Windows XP to not require CMPXCHG8B. But the practical use of that novelty hack is […]
Show full quote

It's fun to note that someone managed to hack Windows XP to not require CMPXCHG8B.
But the practical use of that novelty hack is limited, since application and driver developers that targeted Windows XP
and later would likely compile their code to make use of the CMPXCHG8B instruction even if the underlying OS wouldn't,
so those would still fail to run on such older processors.

Hi, Windows 2000 still ran on i486 - maybe Windows 2000 drivers/applications didn't require CMPXCHG8B yet?
Just came to mind, because Windows XP drivers were actually "Win2k/XP" drivers.
The INF folders, for example, were called "Win2K_XP" and similar.
So they may or may not have required CMPXCHG8B?
Or maybe Windows 2000 had featured CMPXCHG8B emulation?

Edit:

Kernel <= 2.4 with Xfree86 and accelerated drivers is the way to go.

Maybe. But it doesn't solve the (my) "problem".
Namely, that current Linux was useful to keep old hardware running.
On the internet, not just isolated on an old PC in the corner in the cellar.

If Linux becomes an OS that targets modern hardware all the time, then what's the purpose of it? 😟
Then it's nolonger useful to the hardware hacker who wants to get a 35 year old webcam going.
Or likes to use Linux for testing purposes, for all sorts of hardware, no matter the age.

Because, if you have to assemble an old PC to run old Linux 2.4 to test old hardware, then it's silly.
Owners of a 1995 webcam or flatbedscanner etc. want to use the vintage hardware in daily life, purposefully, with modern software.
That's what Linux used to be good for, that's were it did stand out.

The closest to this in Windows world was Windows 98, maybe.
Windows 98SE, for example, was very popular among sellers of second-hand hardware.
Since it had excellent legacy support, eBay sellers had used Windows 98SE to test hardware.
Windows 98SE had shipped drivers (both VXDs and WDMs) for about anything at the time.

Ironically, Windows 11's higher hardware requirements (TPM2) actually have a purpose, but half humanity cries "it's so unfair!".
Here, it's apparently sort of a problem that old hardware nolonger is being supported (officially, there are workarounds).

Typo fixed.

Win2k service packs say no - at least for the memory requirement alone. I could only get the RTM going on my AM5x86 at a slug's pace a decade+ ago.

but this is probably a big "see? win3x-is-best. whats point of linux" nothing thread of no resolve anyway

Jo22 wrote on 2025-05-08, 20:40:

Hi, Windows 2000 still ran on i486 - maybe Windows 2000 drivers/applications didn't require CMPXCHG8B yet? Just came to mind, be […]
Show full quote

digger wrote on 2025-05-06, 10:46:

It's fun to note that someone managed to hack Windows XP to not require CMPXCHG8B. But the practical use of that novelty hack is […]
Show full quote

It's fun to note that someone managed to hack Windows XP to not require CMPXCHG8B.
But the practical use of that novelty hack is limited, since application and driver developers that targeted Windows XP
and later would likely compile their code to make use of the CMPXCHG8B instruction even if the underlying OS wouldn't,
so those would still fail to run on such older processors.

Hi, Windows 2000 still ran on i486 - maybe Windows 2000 drivers/applications didn't require CMPXCHG8B yet?
Just came to mind, because Windows XP drivers were actually "Win2k/XP" drivers.
The INF folders, for example, were called "Win2K_XP" and similar.
So they may or may not have required CMPXCHG8B?
Or maybe Windows 2000 had featured CMPXCHG8B emulation?

Edit:

Kernel <= 2.4 with Xfree86 and accelerated drivers is the way to go.

Maybe. But it doesn't solve the (my) "problem".
Namely, that current Linux was useful to keep old hardware running.
On the internet, not just isolated on an old PC in the corner in the cellar.

If Linux becomes an OS that targets modern hardware all the time, then what's the purpose of it? 😟
Then it's nolonger useful to the hardware hacker who wants to get a 35 year old webcam going.
Or likes to use Linux for testing purposes, for all sorts of hardware, no matter the age.

Because, if you have to assemble an old PC to run old Linux 2.4 to test old hardware, then it's silly.
Owners of a 1995 webcam or flatbedscanner etc. want to use the vintage hardware in daily life, purposefully, with modern software.
That's what Linux used to be good for, that's were it did stand out.

The closest to this in Windows world was Windows 98, maybe.
Windows 98SE, for example, was very popular among sellers of second-hand hardware.
Since it had excellent legacy support, eBay sellers had used Windows 98SE to test hardware.
Windows 98SE had shipped drivers (both VXDs and WDMs) for about anything at the time.

Ironically, Windows 11's higher hardware requirements (TPM2) actually have a purpose, but half humanity cries "it's so unfair!".
Here, it's apparently sort of a problem that old hardware nolonger is being supported (officially, there are workarounds).

Typo fixed.

I like running Linux on older hardware too, but I consider that being able to go back about 30 years on the hardware front while using mainstream current distros is reasonable enough. I think it's OK to draw the line somewhere. While we may disagree on exactly where that line should be drawn, we can't reasonably expect support to extend forever.

I recently tried a KL5KUSB101 based USB Ethernet adapter from 1998ish on a modern x86-64 Linux setup and it just worked, at least enough to ping. For almost everything else PCI(E) or USB we can have VMs with passthrough (banked VGA modes seemed to have performance issues for me when using a PCI VGA card).

For ISA passthrough, if someone could manage to interface/bridge something similar to a dISAppointment LPC to ISA bridge to a VM, we might have more ISA options too, even on modern hardware.

Anyway, I'm veering off topic here.

darry wrote on 2025-05-08, 21:52:

I like running Linux on older hardware too, but I consider that being able to go back about 30 years on the hardware front while […]
Show full quote

Jo22 wrote on 2025-05-08, 20:40:

Hi, Windows 2000 still ran on i486 - maybe Windows 2000 drivers/applications didn't require CMPXCHG8B yet? Just came to mind, be […]
Show full quote

digger wrote on 2025-05-06, 10:46:

It's fun to note that someone managed to hack Windows XP to not require CMPXCHG8B. But the practical use of that novelty hack is […]
Show full quote

It's fun to note that someone managed to hack Windows XP to not require CMPXCHG8B.
But the practical use of that novelty hack is limited, since application and driver developers that targeted Windows XP
and later would likely compile their code to make use of the CMPXCHG8B instruction even if the underlying OS wouldn't,
so those would still fail to run on such older processors.

Hi, Windows 2000 still ran on i486 - maybe Windows 2000 drivers/applications didn't require CMPXCHG8B yet?
Just came to mind, because Windows XP drivers were actually "Win2k/XP" drivers.
The INF folders, for example, were called "Win2K_XP" and similar.
So they may or may not have required CMPXCHG8B?
Or maybe Windows 2000 had featured CMPXCHG8B emulation?

Edit:

Kernel <= 2.4 with Xfree86 and accelerated drivers is the way to go.

Maybe. But it doesn't solve the (my) "problem".
Namely, that current Linux was useful to keep old hardware running.
On the internet, not just isolated on an old PC in the corner in the cellar.

If Linux becomes an OS that targets modern hardware all the time, then what's the purpose of it? 😟
Then it's nolonger useful to the hardware hacker who wants to get a 35 year old webcam going.
Or likes to use Linux for testing purposes, for all sorts of hardware, no matter the age.

Because, if you have to assemble an old PC to run old Linux 2.4 to test old hardware, then it's silly.
Owners of a 1995 webcam or flatbedscanner etc. want to use the vintage hardware in daily life, purposefully, with modern software.
That's what Linux used to be good for, that's were it did stand out.

The closest to this in Windows world was Windows 98, maybe.
Windows 98SE, for example, was very popular among sellers of second-hand hardware.
Since it had excellent legacy support, eBay sellers had used Windows 98SE to test hardware.
Windows 98SE had shipped drivers (both VXDs and WDMs) for about anything at the time.

Ironically, Windows 11's higher hardware requirements (TPM2) actually have a purpose, but half humanity cries "it's so unfair!".
Here, it's apparently sort of a problem that old hardware nolonger is being supported (officially, there are workarounds).

Typo fixed.

I like running Linux on older hardware too, but I consider that being able to go back about 30 years on the hardware front while using mainstream current distros is reasonable enough. I think it's OK to draw the line somewhere. While we may disagree on exactly where that line should be drawn, we can't reasonably expect support to extend forever.

I recently tried a KL5KUSB101 based USB Ethernet adapter from 1998ish on a modern x86-64 Linux setup and it just worked, at least enough to ping. For almost everything else PCI(E) or USB we can have VMs with passthrough (banked VGA modes seemed to have performance issues for me when using a PCI VGA card).

For ISA passthrough, if someone could manage to interface/bridge something similar to a dISAppointment LPC to ISA bridge to a VM, we might have more ISA options too, even on modern hardware.

Anyway, I'm veering off topic here.

When the GNU project and later Linux got started, they had to draw the line at what architectures to support. They required that machines be at least 32 bits. As the GNU Coding Standards (https://www.gnu.org/prep/standards/html_node/ … CPU-Portability) say, "However, don’t make any effort to cater to the possibility that an int will be less than 32 bits. We don’t support 16-bit machines in GNU."

286 machines were far more practical in 1991 than a 486 is today, yet they still avoided support--because it cuts down on the #ifdefs, needing to code everything to operate in blocks smaller than 64K (huge pointers), etc. drastically.

Supporting 486 doesn't butcher the code with anywhere near as many #ifdefs as compared to something like supporting 16-bit, but the kernel needs a critical mass of testers and maintainers for that part of the code and if they had that, they wouldn't be dropping it. In particular, floppy controller support was on the chopping block a while back, but someone stepped up to maintain it, so now it isn't.

My personal opinion is that the DJGPP project, FreeDOS, etc., are far more relevant to running an open source OS on your 486 machine, need help, and are glad to accept it. They won't drop support for your machine anytime soon. If anything, with MS-DOS 4.x being open source now, it's arguably an exciting time for open source DOS development and efforts are better spent there...

Jo22 wrote on 2025-05-08, 20:40:

If Linux becomes an OS that targets modern hardware all the time, then what's the purpose of it? 😟
Then it's nolonger useful to the hardware hacker who wants to get a 35 year old webcam going.
Or likes to use Linux for testing purposes, for all sorts of hardware, no matter the age.

@Jo22 There's always NetBSD. NetBSD prides itself on portability across processor architecures. The 32-bit x86 port of NetBSD still supports 486 processors. Interestingly enough, it's still called NetBSD/i386, even though it requires at least a 486.

Anyway, the NetBSD developers maintain support for a high number of CPU architectures. Moreso than Linux does. Plus NetBSD has a Linux emulation layer that allows it to run Linux software as well.

And then there is that 3rd party project that I mentioned earlier, which adds FPU emulation, which allows NetBSD to to run even on 486SX systems.

So perhaps that would be a good alternative for you?

Jo22 wrote on 2025-05-08, 20:40:

If Linux becomes an OS that targets modern hardware all the time, then what's the purpose of it? 😟
Then it's nolonger useful to the hardware hacker who wants to get a 35 year old webcam going.

It's a good question, i guess answered by numbers to some extent. My guess is that the number of people with a computer having a later than 486 processor (or even later than 32 bit for that matter) far outweighs those without.

Within that the number that have interests similar to "get a 35 year old webcam going" is going to be very low compared to those who just want to go online and those who are hackers, but with interests in 'modern' devices

I'd even guess a the ratio of "486 and /or old stuff hacking" to "later machines and/or new stuff hacking" is somewhere on the order of 1:10000 or so, maybe even larger

That leaves linux with plenty of purpose, but sadly cuts a few out who then will consider alternatives to achieve the same goals, like playing with older tech but using a slightly later machine

Jo22 wrote on 2025-05-08, 20:40:

If Linux becomes an OS that targets modern hardware all the time, then what's the purpose of it? 😟

I'm trying not to start sounding like that "Um Ahktually" linux nerd, but this question tells me you don't really 'get' linux..

Excluding Nvidia proprietary shenanigans, Linux is mostly targeted for reasonably old hardware, i.e. not 15-20+ years old crust.
Outside of that, there are some specific distros targeted for old stuff like Puppy Linux. And well, I suppose you can clump into that pile Debian and some forks like LDME6, because Debian is notorious to be slow with software updates on stable release. Debian 12 is still rocking X11 and all that jazz, still kinda compatible with Kepler GPUs.