Something I came across which made me cease interest in 266 MHz. http://www.cpu-world.com/forum/viewtopic.php?t=23838

On a side note, a long time ago i made the mistake of trying to overclock a 1M cache PPro on an Intel VS440FX motherboard.. It ran fine, until i ran Prime95 on it, then the VRM MOSFET exploded into shrapnel.. Hence why i'm using a 256K cache chip to do this little experiment. I might try a 512K L2 PPro, but a 1M cache chip is probably asking too much.

Might need to add a fan to the VRM heatsink with 1M at 233 MHz. Another comment that was interesting,

I would suspect that the 512k cache versions would be better since the sram was mfg at .35u rather than the .5u of the 256k cache.

Some statistics from another user,

Once upon a time, I gave all my 200MHz PPro's a run @ 266MHz. I had a dozen of 256K, 4 512K and 6 1M. I kept the voltage at the default 3.3V for a reason. One 256K could pass Memtest86+ and another 256K hanged up in the process. The other 10 were unable even to start it. Two 512K hanged up in the process, the other two didn't even make to it. None of the 1M could even POST.

And this poster had to add the missing capacitors to the top of the PPRO-256 to get stable operation at 266 MHz and 3.4 V. http://www.cpu-world.com/forum/viewtopic.php?p=263604#263604 He used 1uF caps from a dead s370 copermine.

My conclusion is that getting 266 MHz to run stable to the point of satisfaction (memtest, prime, etc) requires too much effort, especially for a VS440FX board (need to hack the multiplier jumpers; need to hack the VRM to get greater than 3.3 V; need to solder on 10 ceramic capacitors; need to get lucky).

Even with liquid nitrogen you can achieve only around 300mhz.

Well ,i run my ppros at 233 stable all day on my micronics w6-li mobo.I have 2 sa1 cpus and one sb1 .I managed to run these on my s1668 tyan mobo but i sold the board back in the days in favor for the micronics which has edo dimm ram sockets.

I could only run 233 MHz reliably on 512K or lower CPUs. Were you able to get 233 MHz out of a 1024K chip without increasing the voltage?

Without voltage, I never did.

The Serpent Rider wrote on 2019-10-21, 08:37:

Even with liquid nitrogen you can achieve only around 300mhz.

I vaguely recall a few PPRO boards with 70mhz, 75mhz,80 or even 83mhz FSBs (sometimes undocumented)

In theory you could then test some inbetween speeds like 240mhz , 245 or 250mhz assuming the FSB and cards are stable.

feipoa wrote on 2024-04-18, 18:42:

I could only run 233 MHz reliably on 512K or lower CPUs. Were you able to get 233 MHz out of a 1024K chip without increasing the voltage?

I have one such setup. Blacktop stable at 233MHz in a VS440FX, stock voltage. The regulator heatsink on the motherboard gets blazing hot but I installed a bevy of case fans to blow on it. Will run prime95 for 24+ hrs.

feipoa wrote on 2019-10-20, 09:43:

Thanks guys. I'll see what PLL is on the board to see if there is a hidden 75 MHz option. EDIT: Its an ICS9169M.. datecode with a -01 suffix. https://www.idt.com/document/dst/9169-01-datasheet - 50/60/66 MHz options + TestMode freq, whatever freq. that is. Datasheet does mention "Output frequency ranges to 100 MHz (depending on option)".

Since this subject is revived anyways and I know you're still an active user, I guess, I'd think that the 100Mhz mentioned in the datasheet might be if using another crystal for input?

ok everyone relax, i'm still here...

To wrap it all up in a bow - The main limit to the "overclocking" of the ppro is the L2 cache. The core seems to run just fine at any FSB (within reason, obviously) but the main holding factor is the L2 cache. This is because the L2 cache runs at the same as the core speed - So even with a 100mhz fsb - you're likely limited to a 2.5 multi at best - and few of the 256k chips will run at that speed. Anecdotal evidence being what is is, 233/240mhz on the 256k chip is easily obtainable - 266 drops off QUICK and 300mhz is <10% (probably closer to 1%). Speeds over 233 likely requiring additional voltage for stability. Going to the 512 and 1mb cache chips drops this down by 50% each increase of cache.

According to the Pope of the Church of the Pentium Pro; increasing voltage on a retro processor for an overclock is a SIN.

maxtherabbit wrote on 2025-06-07, 22:55:

feipoa wrote on 2024-04-18, 18:42:

I could only run 233 MHz reliably on 512K or lower CPUs. Were you able to get 233 MHz out of a 1024K chip without increasing the voltage?

I have one such setup. Blacktop stable at 233MHz in a VS440FX, stock voltage. The regulator heatsink on the motherboard gets blazing hot but I installed a bevy of case fans to blow on it. Will run prime95 for 24+ hrs.

This is certainly interesting. Did you run any other software to stress the CPU with random, yet still heavy, tasks? How many 1M chips did you test before finding one that ran at 233 MHz? Could cooling the VRM be the cure?

its about 1:3 for me - if thats helpful

Mine PPro 1M ran stable at 233 MHz, stock voltage. The board was Intel RC440FX. I'm not sure if it was perfectly stable - but could run my whole test suite just fine. https://hw-museum.cz/article/5/cpu-history-to … -1995---1999-/1

And regarding my other PPros - 180/256, 200/256, 200/512 - all of them ran fine at 233... and none of them at 266. Perhaps some posted but couldn't boot OS.

luckybob wrote on 2025-06-10, 04:03:

its about 1:3 for me - if thats helpful

Thanks. I suppose my sample size of two CPUs was too small.

feipoa wrote on 2025-06-10, 08:12:

I suppose my sample size of two CPUs was too small.

Yes, I have 4 x SL25A and all 4 happily run @ 233 MHz at least long enough to do a SIV [Save Local], I just wish I had a Quad Socket system for them.

Below you can see this took about 3 minutes and for much of this time the CPU was 100% loaded. I don't know about 266 as x3.5 is the maximum I can set.

I have a dual setup with PPro 200MHz 1MB chips running at 233MHz. I need better heatsinks as standard ones gets really hot, too hot for my taste.

I wonder do SL259 and SL25A overclock equally well.

Below is the one I have and it was not too bad heat wise, I could run two in my Gigabyte GA-686DX, but I only have the one good cooler, at the moment it's got 2 x SL2KE @ 333 MHz