The DX4 Overdrive 100MHz should be a drop-in replacement for your Intel 486DX2-66, assuming all jumpers were set correctly for the DX2-66.
The Overdrive version is supposed to handle the 5V to 3.3V regulation and multiplier setting on its own.
It is no surprise that it detects as 80 MHz. The 100 MHz didn't exist yet and the test just benchmarks the CPU into one of a set of known "buckets." If it thinks it is a DX2-80, then it will expect the bus to be 40 MHz and will apply more conservative settings if it is in autodetect mode. That would not be causing your issue.

The manual (https://theretroweb.com/motherboard/manual/4m … 20353321113.pdf) has some strange jumpers. For example, J5 is 1-2 for Intel CPU, or 2-3 for Cyrix M6/M7 CPU, *but* if your hard drive is a WD WDAC2340 then use 2-3... TheRetroWeb also doesn't yet have a picture of this board, if you want to take and send one. I have the EFA 3486UC-L4 which is similar, but ISA-only, and is a combo 386/486 board.

If you're not trying to shatter benchmarks, you might keep the 486DX2-66 anyway, as this is a very early VLB board and the DX2-66 is quite fitting.

FlynnTinkers wrote on 2026-01-08, 03:17:

labelled a Pentium Overdrive 100 MHz (full part label: Intel Overdrive DX40DP100 C6210632 SZ956 V1.1).

AFAIK SZ956 is a 75 MHz rather than a 100 MHz part, see https://www.cpu-world.com/sspec/SZ/SZ956.html

There is no 100 MHz Pentium overdrive.
Pentium overdrives with intact cooler are 2,5x clocked and exist as 63 (62,5) FSB 25 and 83 (83,3) FSB 33 MHz.
But there are 486 overdrives too. All DX2 overdrives are 2x clocked and go up to 66 MHz. The DX4 overdrives are 3x cloked and exist in a 75 (FSB 25) and 100 (FSB 33) MHz flavour.

Perhaps you try it with FSB 25.
According to the manual you have to flip both switches 2 and 3 on your "switch 1".
https://theretroweb.com/motherboards/s/efa-4m50hl3

Please report.

Next thing i can consider is that your BIOS misconfigures your board when facing such a fast processor.
You should use avoid auto configuration then.
Remember that it gets interesting with 486 then. 3,3 Volt, Write-Back-Cache, Cache and DRAM-Timing and so on...
If you don't have that much experience it is better to remain with your DX2 66. But remember that it needs a heatsink and a fan to stay cool.

Thanks all. Submitted the attached to Retroweb.

Switched to 25 MHz settings (flipped 2/3) and it's stable; still self-reports 80 MHz. The Cirrus Logic video card and Winbond I/O card both seemed stable as well, though I'm not sure how I'd specifically try to test VLB instability. I have them in slots 2 and 3, due to the mini-tower's drive carriage barely occluding both cards' length in slot 1.

It seems your BIOS tries to configure that Overdrive as DX2 with 80 MHz, that means timings for 40 MHz FSB.

It will use DX2-80 timing for 25x3 MHz because this table entry is the closest one at this frequency.
It will use DX2-80 timing for 33x3 MHz because there is no faster table entry than this one.
Just as jakethompson1 said.

I'd go with the DX2-66 on this semi-early 486 motherboard.

Perhaps you also can post picture(s) of your overdrive processor.

If you really want to use the DX4-75, look into AMISETUP, which will let you disable Autoconfig of the chipset and set parameters manually. This avoids the 80 MHz detection issue.
Because you have 20 ns cache and tag, it's possible you will achieve the fastest timings at 25 MHz but not 33 MHz.

Yeah they are just graded at 20ns maximum, so they are think of a number between 15.1 and 19.9 ns, if they were setting up intending to make 15ns you'll have a lot of "just missed" and they'll do as good as offbrand 15s maybe, or if they were shooting only for 25 and cherry picking the 20s to sell at a premium they might have only just made 20 so not go past 25mhz without wait states ... silicon lottery again.

Though weirdly there seemed to be more faster cache around earlier in the 90s. it might have been a demand thing though PC adoption got exponential from there.

Haven't played with settings yet but here is the CPU in the meantime.

just to be clear this is NOT a pentium overdrive, it is a 486 DX4 overdrive

Part number should be read like DX4 ODP 100, "DX4 486" "Overdrive processor" "running at 100" not "random intel nonsense" "OD = overdrive" "P100 = Pentium 100"

Since the trailing P stands for processor, this is why you'll see actual pentium overdrives called PODP with a leading P for Pentium.