Something similar was done years ago using Socket 3 and a POD on an ASUS board.

Some So5 boards had *very* low FSB options (as they were basically using 486 PLLs). One I had myself: Spring Circle P5C01. It had options down to 16MHz. That might not sound slower than the other options but...

This is a board with SiS 5511 chipset and a 6205 VGA chip using UMA. So the memory bandwidth with 16MHz FSB is shared between CPU and VGA. It was *SLOOOOOW*, running Windows XP not quite as slowly as on my Pentium Overdrive I was running at 16MHz (16MHz 32b bus x 1 by removing fan), but being in same ballpark.

Ugh, sounds like torture to me. I recently used my TXP4 to test a K6 300 I have to make sure it works. And it does, but the difference between 6x83 with 256K L2 cache and 4.5x66 with only the motherboard's 512K L2 was night and day.

PLL modding when? 😜

Shuttle HOT-541 (430 FX) and HOT-553 (430 HX) have very flexible PLL. You may go down to 16MHz FSB.
Below is table (frequencies are measured with osciloscope). At 16MHz FSB motherboard was still able to boot-up with PCI Trio64V+
This is my HOT-553 with Pentium 120 (CPU multiplier set up as 1.5), I also have Cyrix 6x86 so I may try with 1x multiplier as well (in the future).

I just had a thought... while Cyrix 6x86 and MII class might have a functional 1x multiplier (Implemented multis can vary with time, fab, and OEM version, apparently identical ST/IBM/Cyrix versions of same model have shown differences.) ... it might not be slower than a P54 or P55 at 1.5x ... because that core is particularly fast on 16 bit code... anything you need about "486" fast is probably mostly 32bit by then... anything you want TurboXT/286 fast is probably 16 bit code. It might get brought more into line with L1/L2 kneecapping, as then the 50% loss of memory speed is gonna show.

Still waiting for someone to hit 7mhz core clock on socket 5/7

Some socket 4 boards had very low undocumented FSB settings, when I’ve tried very low FSBs on either socket 4 or 5 the board would usually freeze during post
.

Dumber idea but several socket 5/7 boards would allow a voltage greater than 3.5 volts.

My m571 could overclock a k5-or133 to pr200 using 4.1 volts (undocumented)

Such voltages would enable a socket 4->7 interposer and a 1x multiplier using a p60/66

rmay635703 wrote on 2024-12-06, 06:34:

Some socket 4 boards had very low undocumented FSB settings, when I’ve tried very low FSBs on either socket 4 or 5 the board would usually freeze during post

Yes the same clockgen chips that gave some 486 boards "high" 60 and 66 settings were in use on some socket 4, so you can get the "486" speeds like 40, 33, 25, and maybe 20, 16.

edit: Oh by the way, I noticed in some clockchip data sheet, that when the lowest frequencies were used, the DMA/floppy and USB if used reference frequencies went to something different also. It seems possible therefore, that some low speeds may be floppy and soundblaster incompatible, and USB may need to be disabled. Ideally you can look up the datasheet and read what your clockchip does, but many refs only have basic settings or you need to infer things from "close family" chips because that specific one is missing. Anyway, get a hang on boot, try PIO HDD only, no floppy or USB and see if it goes further.

Sphere478 wrote on 2024-12-06, 04:02:

Still waiting for someone to hit 7mhz core clock on socket 5/7

I tested this configuration a long time ago, even slower than 7MHz using Socket 7 board with working Turbo header on a K6-III+, slower than an IBM XT, measured using various benchmarks:

WIP 2: The 6-in-1 Turbo-switched Socket 7 - from XT to 500MHz; dual Tseng powered...

Speedsys 0.45

NSI 0.9

Topbench 3

Landmark 2MHz

A 4.77MHz IBM XT should be around 0.5 Speedsys, NSI around 1.0 and Topbench around 4; so these benchmarks give an idea of how slow it is. Board boots fine and managed to test with a speed sensitive game as well. Currently all in storage but will return someday and finish.

EDIT:

Dhrystone 268 (IBM XT around 300)

Wow! Very nice!

j^aws wrote on 2024-12-06, 18:45:

Speedsys

Please may you post the cache chart of SpeedSys?
(You can skip extended memory test of course)

Disruptor wrote on 2024-12-09, 06:50:

j^aws wrote on 2024-12-06, 18:45:

Speedsys

Please may you post the cache chart of SpeedSys?
(You can skip extended memory test of course)

Sure, luckily managed to grab these at the time:

K6-III+ at 100MHz with Turbo enabled:

K6-III+ at 100MHz with De-Turbo enabled:

Sphere478 wrote on 2024-12-09, 06:32:

Wow! Very nice!

Thanks!

Impressive chart.

j^aws wrote on 2024-12-09, 18:58:

Thanks!

Yes but SpeedSys shows now 9.26 and not the 0.45 from above. Has this low value not been reproduceable?

double post

double post

Disruptor wrote on 2024-12-10, 14:37:

Impressive chart.

Yes but SpeedSys shows now 9.26 and not the 0.45 from above. Has this low value not been reproduceable?

Yes, it's reproducible - the test just takes a very long time to complete. The 0.45 Speedsys is with all caches disabled (L1, L2 and L3), so there isn't a cache graph from Speedsys. Were you expecting L1 still to be enabled for 0.45? You can see in the table below the various baseline configurations for caches and Speedsys scores:

Wonder if this could be used to reach the "sweet spot" for old games, requiring 286'es and such (reason for the turbo button on 3/486 cases)

Ahrle wrote on 2024-12-10, 19:02:

Wondering if this could be done to reach the "sweet spot" for old games requiring 286'es and such (reason for the turbo button)

Yes, indeed this does occur with the range of speeds available. Something to also consider is the VGA speed where animations can be too fast still, even with the correct CPU speed; so slowing down this aspect also helps.