Not all 200MHz MMX CPUs have the 3.5x multiplier avaliable even though their core could probably handle 233MHz just fine.

melbar wrote on 2020-08-26, 03:18:

Does every P200 MMX run at 233 MHz?

If it's the case, this would mean also that every P200 MMX has a unlocked multi...?

Locked multipliers were not a thing at the times of the Pentium MMX. The introduction of locked multipliers happened in response to the rise of counterfeit Pentium 2, Pentium III and early Athlon Slot A CPUs.

I am quite confident that every Pentium MMX supports four different multipliers. All Pentium MMX CPUs support x2, x2.5 and x3 (133, 166, 200 at FSB66). Dornkaat might very well be right that there are early P55C dies that Interpret the fourth setting as x1,5 (100) just as the P54C does, not as x3,5 (233).

I have two Pentium MMX'es, and one refuses to take higher multiplier than 3, while other happily goes to 3.5. They look absolutely identical.

So in practice, locked multiplier PMMX's do exist.

GigAHerZ wrote on 2020-08-26, 06:18:

I have two Pentium MMX'es, and one refuses to take higher multiplier than 3, while other happily goes to 3.5. They look absolutely identical.

So in practice, locked multiplier PMMX's do exist.

Maybe they're not locked, but produced before the 1,5/3,5 remapping was introduced?

H3nrik V! wrote on 2020-08-26, 06:37:

GigAHerZ wrote on 2020-08-26, 06:18:

I have two Pentium MMX'es, and one refuses to take higher multiplier than 3, while other happily goes to 3.5. They look absolutely identical.

So in practice, locked multiplier PMMX's do exist.

Maybe they're not locked, but produced before the 1,5/3,5 remapping was introduced?

Isn't that just an implementation detail of upside locked multiplier? 😜

GigAHerZ wrote on 2020-08-26, 06:43:

H3nrik V! wrote on 2020-08-26, 06:37:

GigAHerZ wrote on 2020-08-26, 06:18:

I have two Pentium MMX'es, and one refuses to take higher multiplier than 3, while other happily goes to 3.5. They look absolutely identical.

So in practice, locked multiplier PMMX's do exist.

Maybe they're not locked, but produced before the 1,5/3,5 remapping was introduced?

Isn't that just an implementation detail of upside locked multiplier? 😜

I think that's a question of definition 😉

IIRC P54 (and P55) only had 2 bits for multiplier selection, which led to 1.5, 2, 2.5 and 3. In order to get higher clock speeds, Intel remapped the 1.5 multiplier to 3.5 in the later P55's

Maybe the first P54's actually only had one bit, selecting 1.5 or 2?

Maybe they're not locked

They are locked toward higher multiplier, just as early Pentium II. Early manufactured CPUs didn't had that lock, probably because Intel weren't sure yet about PMMX 233 production yields, so they just kept it undocumented. Not like it was their first time too. Pentium 120/133 also had semi-locked multiplier.

I had a p200mmx that only was stable to 225, 233 it wasn’t stable in Windows YMMV

My "unlocked" P200MMX runs at 3.5x75 = 260MHz with no problems. If anything, then some PCI cards are a bit picky because of the higher PCI bus clock. 😜

GigAHerZ wrote on 2020-08-26, 06:18:

I have two Pentium MMX'es, and one refuses to take higher multiplier than 3, while other happily goes to 3.5. They look absolutely identical.

So how do you know that this is a multiplier problem, and it's not just the core crapping out at 233MHz? Did you try to run the processor at lower FSB speed, like 3.5 * 60 = 210MHz?

I think the greater number of them supports the 3,5 multiplier. I have four 200mmx chips and all like the 3,5 very much. Two of them also like 2,5x100 (250MHz) without overvolting on super 7 boards. One even got to 3x95 (285Mhz) at 3,0v. This could be an escape if your MMX dislikes the 3,5 multiplier.

I know for sure, stepdad's computer was 200MMX and is locked. Had to overclock the fsb to 75 to get 225MHz. That was on SP97-V.

Cheers,

mkarcher wrote on 2020-08-26, 20:59:

GigAHerZ wrote on 2020-08-26, 06:18:

I have two Pentium MMX'es, and one refuses to take higher multiplier than 3, while other happily goes to 3.5. They look absolutely identical.

So how do you know that this is a multiplier problem, and it's not just the core crapping out at 233MHz? Did you try to run the processor at lower FSB speed, like 3.5 * 60 = 210MHz?

The PC boots, but still doesn't have 3.5x multiplier. I don't remember, did it still run at 3.0x or 1.5x multiplier. It just didn't eat that configuration.