According to this thread it should be safe to run an MMX chip at 3.3v.

Using a pentium mmx on a 3.3/3.3 motherboard.

I haven't tried it though.

"Safe" is stretch. That topic has n=1 example of where it didn't end in flames within a few months.

The P55C Pentium MMX was massively over-engineered and capable of much higher speeds than actually sold at; Intel limited it so as not to compete with the Pentium 2. As such it can survice abuse like overvolting it by 0.5V. That will definitely reduce its lifespan, but if that was going to be centuries the CPU might still outlive you. That said, using a decent VRM rather than using the CPU to burn off the excess power is a lot more elegant.

In addition, the CPUs may be bulletproof, but you'll be drawing more current at 3.3V than the board is designed for and on an AT system, it's the motherboard VRM that makes 3.3V from 5V. If this is an AT board, check that as its VRM is almost certainly less overengineered than the P55C. That was also mentioned in the topic...

Note that this board is an Intel OEM design and Intel BIOS is notorious for refusing to boot with unknown CPUs. So even if it might technically work, BIOS could ruin the party. That said, there were MR BIOS available for most Intel OEM 430FX boards. MR BIOS will not have this limitation.

Which board does your system have, exactly?

dionb wrote on 2025-05-29, 10:04:

"Safe" is stretch. That topic has n=1 example of where it didn't end in flames within a few months.

True. Possibly not worth the risk, long term, if you can't afford to lose the hardware. But maybe could try it for a short period to check for compatibility before considering VRM mods?

dionb wrote on 2025-05-29, 10:04:

Which board does your system have, exactly?

Thanks for the reply! It's an Intel 652605-501/Intel Advanced ATX/Thor. Currently have the 1.00.06.CN0 BIOS.

RetroPCCupboard wrote on 2025-05-29, 11:43:

[...]
True. Possibly not worth the risk, long term, if you can't afford to lose the hardware. But maybe could try it for a short period to check for compatibility before considering VRM mods?

The big risk in short term is 100% on the motherboard's VRM, the CPU will 100% survive a test.

Tbh the only dead P55C I ever say was one that came from a failed server. It failed after over a decade of 24/7 use. Turns out during assembly the heatsink fell of the P166MMX. The CPU had been running without a heatsink for years. The OPGA housing had been baked brown and was cracked. And even then it took years for it to fail. Don't try that with an Athlon Thunderbird, for example 😉

But regarding the motherboard VRM - what board is it exactly?

dionb wrote on 2025-05-29, 12:41:

But regarding the motherboard VRM - what board is it exactly?

Seems to be effectively this board but without the J7B1 recovery boot jumper. It is identical to the board shown here.

dionb! Thank you for mentioning MR BIOS. I knew nothing about it until now (still getting my bearings). I was able to successfully flash my board with it. Looks like it includes a BIOS option to disable CPU cache. I bought a 166mhz MMX and will see if it all works out!

UPDATE: MR BIOS working well. My HDD no longer boots to Win 95 though. BIOS sees the drive just fine, and I set it to boot from C:, but essentially nothing happens

Possible change to/from LBA and ECHS?

wierd_w wrote on 2025-05-29, 18:53:

Possible change to/from LBA and ECHS?

Good question, I don't see any options in bios to change that though.

UPDATE: Claude recommended that I create a win95 boot disk, and run
fdisk /mbr
then
sys C:

Worked like a charm!

aderius wrote on 2025-05-29, 17:27:

[...]

Seems to be effectively this board but without the J7B1 recovery boot jumper. It is identical to the board shown here.

Thor? The first ATX board.

Well, no worries about 3.3V current then as ATX gets that from the PSU.