News Tiny Intel 486 Runs Windows 95 and MS-DOS with Ease at 100 MHz

[Admin]

Eric Mackrodt from The Eric Experiment has created a super tiny Intel 486 desktop from scratch using an Intel 486DX4 CPU and housing the unit in a custom 3D-printed shell.

Tiny Intel 486 Runs Windows 95 and MS-DOS with Ease at 100 MHz : Read more

 

[Dr3ams]

Nice! I like it. In the 90s I had a 386, 486 DX and a Pentium 133 MHz.

Also...because there are some games I still want to play, I have Windows 98SE and XP installed in virtual machines.

Some of the games I still play that won't run on Windows 10/11:
- Act of War (XP)
- Red Alert 2 (XP)
- Frogger (Windows 98SE)
- Marble Drop (Windows 98SE)
- Microsoft Arcade (Windows 98SE)
- Think-X (Windows 98SE)
- Yahtzee from Hasbro (Windows 98SE)

 

[Eximo]

The headline is weird. Why wouldn't a 100Mhz 486 be able to run Win95 and MS-DOS? Right in the correct era...

Has he not heard of PC104 form factor? Could do it with off the shelf parts if you wanted to.

 

[bit_user]

It would've been really nice to include a couple pictures of the guts. I'm not going to watch the Youtube video.

the whole PC is based around a small SBC with an Intel Intel 486DX4 CPU with about 100 MHz of power.

Huh. I had an AMD 486DX4-100. I didn't think Intel made a DX4, though. That was well into the Pentium era, for Intel, which makes me further doubt it.

For graphics, he’s using a Tseng ET4000 video card.

I had one of these. It was an ISA card, though. In the DX4 era, you'd normally have had a VESA Local Bus card or even PCI. IIRC, VLB was like 4x as fast as ISA, but still like half or a quarter as fast as PCI. My DX4 motherboard had PCI, but the weird thing is it was tied to the CPU's base clock. So, that meant it ran at 25 MHz instead of the typical 33 MHz. I think my dad had a Pentium 75, which did the same thing.

 

[Amdlova]

My first computer comes from garbage pentium 100 with some edo Ram. The motherboard have jumps to set clock on cpu and pci first interaction with math. Got some overclock from cpu. After I got a p2b with a pentium 2 350 and a voodoo 2 card gold old days

 

[Eximo]

I inherited a Pentium 83Mhz overdrive chip in a 486 Zeos desktop. Paradise Pipeline 64 and a Voodoo 2 8MB, at the end I managed to get 4 4mb sticks into that thing.

We had a similar system with the 100Mhz DX in it I think. Bumped it up from a 33 if I recall. I don't remember what the original processor in my 486 was, might have been a 66mhz. I know we had more 486 processors than computers. Probably sitting in a drawer somewhere, just know that my brother bought himself a PII and Matrox card and gave me the Pentium. Not that I kept it too long. AMD K6 sometime after that. Man the jumps in performance were staggering back then.

 

[JamesJones44]

I didn't think Intel made a DX4, though

They did manufacture a DX4 100 and 75. They didn't sell all that well in OEM form because the Pentium 60 and 66 were quicker in most benchmarks for about the same price. It also didn't help that the Intel DX4 came out after the Pentium 60 and 65 were released.

 

[Alvar "Miles" Udell]

My first Windows computer had a 75mhz AMD K5 and 8MB RAM, ESS audio, S3 Virge DX GPU, 600mb HDD, ran Windows 95, and was able to play Doom and navigate the internet with dial up, just like thousands of others did in the mid 90s.

It's a nice project though.

 

[blppt]

The headline is weird. Why wouldn't a 100Mhz 486 be able to run Win95 and MS-DOS? Right in the correct era...

Has he not heard of PC104 form factor? Could do it with off the shelf parts if you wanted to.

Exactly. I ran Windows 95 on a 486SX-25 back in the day. It didn't run GREAT, but it got the job done.

DX4/100 was a dream for me.

 

[blppt]

Huh. I had an AMD 486DX4-100. I didn't think Intel made a DX4, though. That was well into the Pentium era, for Intel, which makes me further doubt it.

Intel actually released their DX4/100 first---I think you are remembering the AMD DX4/120, which Intel never made a 486 chip that fast.

 

[blppt]

I had one of these. It was an ISA card, though. In the DX4 era, you'd normally have had a VESA Local Bus card or even PCI. IIRC, VLB was like 4x as fast as ISA, but still like half or a quarter as fast as PCI.

IIRC, VLB's problem wasn't that it was significantly slower than PCI, it was that its clock speed was locked to the bus speed, meaning that for certain CPUs like the DX50 (not DX2), it would be problematic running a VLB card at that bus speed (obviously the clock multiplier DX2 50-66-75-100 would operate at 25 or 33mhz bus, and not suffer that problem).

PCI, on the other hand, had a locked clock speed of 33mhz. Also, the connectors for VLB were downright massive in comparison.

 

[bit_user]

PCI, on the other hand, had a locked clock speed of 33mhz. Also, the connectors for VLB were downright massive in comparison.

No, the first version of PCI might have had a max speed of 33 MHz, but I assure you that there were boards like my dad's Pentium 75 which ran it at 25 MHz and that definitely bugged me. I think my 486DX4-100 did, too.

 

[Endymio]

The headline is weird. Why wouldn't a 100Mhz 486 be able to run Win95 and MS-DOS? Right in the correct era...

A side effect of the death of Moore's Law. There's not enough new news; sites must recycle the old.

 

[blppt]

No, the first version of PCI might have had a max speed of 33 MHz, but I assure you that there were boards like my dad's Pentium 75 which ran it at 25 MHz and that definitely bugged me. I think my 486DX4-100 did, too.

IIRC the DX4/100 was 3x clock multiplier, not 4X. The DX4/75 was based on the 25mhz bus. I guess there could have been a DX4/100 based on the 25mhz bus and clock quadrupled, but the most popular one was the 3x33mhz model.

If your PCI bus was running at 25mhz, it was running out of 'official' spec, which was set at 33mhz. But generally, slower clock speeds are more tolerable than faster ones, which is what you were going to start getting as the base system clock crept up to 100, making VLB untenable, even if it wasn't a highly inconvenient slot type that was designed with video in mind, which is why PCI quickly replaced it. Kinda like how the specialized AGP didn't really stick around that long once PCI-E was developed.

 

[bit_user]

IIRC the DX4/100 was 3x clock multiplier, not 4X. The DX4/75 was based on the 25mhz bus. I guess there could have been a DX4/100 based on the 25mhz bus and clock quadrupled, but the most popular one was the 3x33mhz model.

Okay, you win. My aging brain remembered there was something about PCI on my board that was sketchy, and you've jogged my memory. What I think I had was a AM486DX4-120, with PCI running at 40 MHz.

https://en.wikipedia.org/wiki/Am486#Am486_models
​

However, this confirms what I recall about the Pentium 75 running its PCI bus at 25 MHz:

https://www.vogons.org/viewtopic.php?t=71797
​

So, I'll take that win, thank you!

If your PCI bus was running at 25mhz, it was running out of 'official' spec, which was set at 33mhz.

Heh, my next PC was a Pentium Pro 180, which ran PCI at 30 MHz! That's not even the worst part. It had the write-combining bug, which hampered performance by like a factor of 2. At least I got it cheap, at some hardware liquidator.

Again, this backs me up on its 30 MHz PCI speed:

http://philipstorr.id.au/pcbook/book1/speed.htm