pcwizard7 wrote:

I am wanting to know the last pentium 2 or 3 that can be downclock by bios?

Thanks

all of them can be if the BIOS allows one to change the FSB. OR maybe via Setfsb
As for multi's, one would need a older Pii or Piii ES cpu.

You need a Pentium II produced before week 34 1998. Retail Pentium II (and III) processors produced after that date are multiplier locked.
Good chances with 233/266/300/333 MHz Models, especially Klamath core. Only a minority of the 350/400 MHz Deschutes core processors are not locked.

gerwin wrote:

You need a Pentium II produced before week 34 1998. Retail Pentium II (and III) processors produced after that date are multiplier locked.
Good chances with 233/266/300/333 MHz Models, especially Klamath core. Only a minority of the 350/400 MHz Deschutes core processors are not locked.

There were actually 350MHz Deschutes which were not fully locked? Any idea which ones? Or is it just a date code thingy? 😜

Tetrium wrote:

There were actually 350MHz Deschutes which were not fully locked? Any idea which ones? Or is it just a date code thingy? 😜

When searching these topics on keyword '350', I get no reports of an unlocked one. But no reports of pre week 34 locked one either.
- Slowing a 440BX based PC
- Which Pentium IIs can be underclocked?
- Quick question about underclocking a Pentium 2

There is insufficient data. I just throw them in the general week 34 rule, as a brother of the 400MHz one, until proven otherwise.
Please don't expect me to buy a a handful of 350MHz ones just to test, I already did that with the Pentium MMX 😉

Edit: just checked my notes, nothing much on 350Mhz, but there is one report of a 450MHz unlocked one. Anyways I will just insert all of it here:

Klamath core: SL2HD 233 - Gerwin has '98-08 Malay - unlocked SL2HE 266 - AdamP has an unlocked one Deschutes core: SL2KA 333 (dA […]
Show full quote

Klamath core:
SL2HD 233 - Gerwin has '98-08 Malay - unlocked
SL2HE 266 - AdamP has an unlocked one
Deschutes core:
SL2KA 333 (dA0) AdamP has an unlocked one
SL2QF 333 (dA0)
Gerwin has '98-29 Malay - unlocked ( has cache rated for 200MHz )
Tetrium has '98-19 Malay - unlocked
Tetrium has '98-29 Malay - unlocked
ux-3 has an unlocked one
SL2VY 300 (dA1) Tetrium has a locked one, Philippines
SL2S5 333 (dA1) AdamP has a locked one
SL2ZP 333 (dA1) Gerwin has '98-39 Costa rica - locked
SL2SH 400 (dA1) Bestemor has an unlocked one
SL2S7 400 (dA1)
Tetrium has '98-30 Malay - unlocked
Malik has '98-25 Philippines - unlocked
Skyscraper has '98-31 Malay - unlocked
Gerwin has '98-30 Malay - unlocked
Gerwin has '98-33 Philippines - locked
SL2U3 350 (dB0) Tetrium has a locked one, Ireland
SL2U6 400 (dB0) AdamP has a locked one
SL2U7 450 (dB0) ux-3 has two locked ones, but also '98-30 Malay - unlocked

Note: unlocked is not entirely the right term, it is more like 'multiplier is adjustable mostly downwards'.

I actually have a 400Mhz one from week 33 which is locked!? Week 33 is not safe either 😵
Tetrium, what are the date codes on your SL2VY 300 and SL2U3 350?

gerwin wrote:

When searching these topics on keyword '350', I get no reports of an unlocked one. But no reports of pre week 34 locked one eith […]
Show full quote

Tetrium wrote:

There were actually 350MHz Deschutes which were not fully locked? Any idea which ones? Or is it just a date code thingy? 😜

When searching these topics on keyword '350', I get no reports of an unlocked one. But no reports of pre week 34 locked one either.
- Slowing a 440BX based PC
- Which Pentium IIs can be underclocked?
- Quick question about underclocking a Pentium 2

There is insufficient data. I just throw them in the general week 34 rule, as a brother of the 400MHz one, until proven otherwise.
Please don't expect me to buy a a handful of 350MHz ones just to test, I already did that with the Pentium MMX 😉

Edit: just checked my notes, nothing much on 350Mhz, but there is one report of a 450MHz unlocked one. Anyways I will just insert all of it here:

Klamath core: SL2HD 233 - Gerwin has '98-08 Malay - unlocked SL2HE 266 - AdamP has an unlocked one Deschutes core: SL2KA 333 (dA […]
Show full quote

Klamath core:
SL2HD 233 - Gerwin has '98-08 Malay - unlocked
SL2HE 266 - AdamP has an unlocked one
Deschutes core:
SL2KA 333 (dA0) AdamP has an unlocked one
SL2QF 333 (dA0)
Gerwin has '98-29 Malay - unlocked ( has cache rated for 200MHz )
Tetrium has '98-19 Malay - unlocked
Tetrium has '98-29 Malay - unlocked
ux-3 has an unlocked one
SL2VY 300 (dA1) Tetrium has a locked one, Philippines
SL2S5 333 (dA1) AdamP has a locked one
SL2ZP 333 (dA1) Gerwin has '98-39 Costa rica - locked
SL2SH 400 (dA1) Bestemor has an unlocked one
SL2S7 400 (dA1)
Tetrium has '98-30 Malay - unlocked
Malik has '98-25 Philippines - unlocked
Skyscraper has '98-31 Malay - unlocked
Gerwin has '98-30 Malay - unlocked
Gerwin has '98-33 Philippines - locked
SL2U3 350 (dB0) Tetrium has a locked one, Ireland
SL2U6 400 (dB0) AdamP has a locked one
SL2U7 450 (dB0) ux-3 has two locked ones, but also '98-30 Malay - unlocked

Note: unlocked is not entirely the right term, it is more like 'multiplier is adjustable mostly downwards'.

I actually have a 400Mhz one from week 33 which is locked!? Week 33 is not safe either 😵
Tetrium, what are the date codes on your SL2VY 300 and SL2U3 350?

Thanks for your extensive re-research 😜

I'd have to go check, but I basically tossed them back into the box after testing so I can't affirm for sure which date codes these exact chips were unless I find some old pics of said CPUs while undergoing testing.

And I never even knew there were 'unlocked' 450MHz Deschutes! I'd definitely have to take a closer look to those when I get to it (which alas may be for a long while).

All of the Klamath chips I tried for my recent reviews were unlocked. Will try some Deschutes soon, but not holding my breath 😀

I'm especially wondering which CPUs have which multipliers enabled and if there is a difference in what board one uses. There seems to be a difference even amongst CPUs of the same stepping and speed rating. Not sure about the Slot 1 board that is used though. Details are sketchy atm.

I heared that the abit BH6 was capable to unlock some locked pentium 2s by making them "think" they were in 66MHz FSB mode instead of 100MHz FSB. More or less the reverse thing to the tape mod on celerons, but still pushing the cpu to it's original FSB.

I'm going to have one soon with a 350MHz pentium 2. I already have one 350 and a 400MHz model, so this makes 3 potential unlocked/unlock-able CPUs to test

PhilsComputerLab wrote:

Will try some Deschutes soon, but not holding my breath 😀

Regardless of wheter the are locked or not, I would be glad to add them to the list above with SSPEC and datecode.

Deksor wrote:

I heared that the abit BH6 was capable to unlock some locked pentium 2s by making them "think" they were in 66MHz FSB mode instead of 100MHz FSB. More or less the reverse thing to the tape mod on celerons, but still pushing the cpu to it's original FSB.

I'm going to have one soon with a 350MHz pentium 2. I already have one 350 and a 400MHz model, so this makes 3 potential unlocked/unlock-able CPUs to test

Again, please share your findings.
About that Abit board 'trick'. It is the second time it is mentioned, but it is all rather vague.

Well I don't have that board at the moment so I can't test that. I'm just quoting what I've read on old reviews for that board. It might be just some bs, but I remember hearing about unlocking a pentium 2 somewhere else, so I don't know ... I might get it next week, but then I don't know if I will find the right pentium 2

I did find this link http://forum.oc-forums.com/vb/showthread.php? … ight=unlock+PII from this old thread http://www.overclockers.com/forums/showthread … ntium-II!!!-hlp but unfortunately I think that it's gone forever as the wayback machine doesn't have this page 🙁

Oh sorry I was editing my post and you replied already. You are right these old posts about Abit are there, I found them.
One thing is this though - locked is locked, nothing a motherboard can do about it. But for the 'not locked Pentium IIs', their maximum multiplier is hard to predict. And also it is not fully understood what affects it.
But I figure it is also of limited benefit, at least to me. Like I briefly used a Pentium II 333Mhz with 200MHz cache modules. So I can set it at 2x50=100MHz when I want it to be slow, and reboot at 4x100=400MHz when I want it to be fast. I don't really care if it cannot do 5.5x66=366MHz. And faster then 400MHz is impossible with that cache anyways.

Edit: Oh, you added the links, I will read it. thanks.

Here is where I found the "abit trick" : http://www.hardware.fr/articles/75-1/abit-bh6.html

This is in french, so I will translate the most interesting part :

"Tout comme pour les fréquences de bus, les coefficients multiplicateurs on aussi droit à leur petite innovation. Ainsi, l´option ´SEL100/66# Signal" vous permettra tout simplement de débloquer le coefficient des Pentium II 350 et 400. En théorie, il vous suffit donc de donner la valeur "Low" pour cette option et vous pourrez utilisé tout les coefficients ce situant entre x3.5 et x5, pour par exemple faire de l´overclocking à 450 MHz @ 4.5 x 100. Dans la pratique, cela marche sur tout les Pentium II 350 et 400 sauf les derniers qui disposent d´un blocage du coefficient différent, et qui ne peut donc pas être débloquer par la BH6."

Means :
"Just like the bus frequencies, the multiplier are getting improvementd too. Thus, the option "SEL100/66# Signal" will simply allow you to unlock the multiplier from the Pentium 2 350 and 400. In theory, you just need to set it's value to "LOW" and then you will be allowed to use any multiplier between 3.5x and 5x in order to, for example, overclock to 450MHz @ 4.5*100. In the reality, this works for all (or on all, I dont know what the author meant) the Pentium 2 350 and 400 except for the latest ones which are locked differently and so can't be unlocked by the BH6"

So I'm just guessing, but maybe you can fully unlock the pentium 2s that were partially locked ? Looks like when you put one of those on a motherboard that can do only 66mhz fsb, the cpu would unlock it's highest multipliers in order to make you able to set it to a more accurate speed. This would require some testing

This have been discussed by people here http://forum.hardware.fr/hfr/OverclockingCool … jet_19143_1.htm

I also found another article where they said again that the abit BH 6 could unlock pentium 2s, but I can't find the link 🙁

Inside the forum thread you linked to in the previous post, this in-depth article was mentioned:
Clocking and locking the Pentium II - by Jon "Hannibal" Stokes
Saved for offline reading:

On page four is this bit, where the Abit boards are mentioned:

The B21 trick I imagine that, by now, many of you are wondering how Tom’s B21 trick plays into all of this. The answer is, it re […]
Show full quote

The B21 trick
I imagine that, by now, many of you are wondering how Tom’s B21 trick plays into all of this. The answer is, it really doesn’t. The SEL 100/66# pin is a separate issue, and it has nothing to do with multiplier locking. Let me give a quick an dirty explanation (for the real details, head over to Tom’s) and you’ll see what I’m talking about. The idea behind Intel’s introduction of the SEL 100/66# pin (or pin B21) is that the chipset and CPU would use it to cooperate with each other on setting the bus speed. When you put a CPU into the slot, the chipset is supposed to check SEL 100/66# to see if it’s low or high. If it’s low, it knows to set the bus speed to 66MHz. If it’s high, it knows to set the bus speed to 100MHz.
.....
Actually, if you’ve got a motherboard that doesn’t play by the rules (like the Abit BH6), then you don’t have to worry about this signal. The BH6 motherboard itself just ignores the SEL 100/66# signal (and ignores Intel’s BX chipset specifications too) and sets the bus speed to whatever you want.

This doesn't sound any different to how my Gigabyte GA-6BXC motherboard works, where I can set things any way I want to (including AGP divider). True though, a lot of i440BX motherboards try to be 'helpful' and use this FSB autodetection to make life harder. Like this one did before I put an end to it 😉

But as to the main point, what can I say? The sensible thing to say is that the CPU cannot know the actual FSB speed (as is also stated in that article), and therefore the FSB speed cannot influence the multipliers that are possible (besides reaching the overclocking limit). Though I remember in older topics here, that the upper multi was temperamental in weird way. Even at the end of that arstechnica article they admit that maybe intel added something nifty....

Hm in the arstechnica article you found, they are talking about german people that modded their CPUs, and there are some links to two pictures that apparently someday was hosted on "sandpile.com". That is really infortunate ...

Deksor wrote:

Hm in the arstechnica article you found, they are talking about german people that modded their CPUs, and there are some links to two pictures that apparently someday was hosted on "sandpile.com". That is really infortunate ...

Yes i noticed that too. You could ask the sandpile.com webmaster? But in the followup article they observed modications around the BSEL (pin B21) trace, which would indicate that they only changed a 66FSB CPU to a 100FSB one.

I did two thing yesterday:
First, I tried my locked and not-locked Pentium IIs with the Gigabyte GA-6BXC motherboard again. Nothing new: Locked remains locked, regardless of taping pin B21. Not-locked Pentium II 400 does not do a multi higher then 4.0x (Regardless of FSB speed. Regardless of AGP Divider, EDIT: Regardless taping pin B21), whilst the Pentium II 333 does up to 5.0x.

Second, I read this again: Which Pentium IIs can be underclocked?. And if you look at ux-3's posts, het also talks about how his not-locked Pentium II 400 refuses to do a multiplier above 4.0x, except in some rare occasions. There is also the confusion about how two people with Abit boards manage to run identical CPUs at 5.5x.

So the Abit mystery is mainly about setting multiplier 4.5x to 5.5x on a not-locked Pentium II 400. A mystery insofar I don't know how that works...

gerwin wrote:

Yes i noticed that too. You could ask the sandpile.com webmaster? But in the followup article they observed modications around t […]
Show full quote

Deksor wrote:

Hm in the arstechnica article you found, they are talking about german people that modded their CPUs, and there are some links to two pictures that apparently someday was hosted on "sandpile.com". That is really infortunate ...

Yes i noticed that too. You could ask the sandpile.com webmaster? But in the followup article they observed modications around the BSEL (pin B21) trace, which would indicate that they only changed a 66FSB CPU to a 100FSB one.

I did two thing yesterday:
First, I tried my locked and not-locked Pentium IIs with the Gigabyte GA-6BXC motherboard again. Nothing new: Locked remains locked, regardless of taping pin B21. Not-locked Pentium II 400 does not do a multi higher then 4.0x (Regardless of FSB speed. Regardless of AGP Divider), whilst the Pentium II 333 does up to 5.0x.

Second, I read this again: Which Pentium IIs can be underclocked?. And if you look at ux-3's posts, het also talks about how his not-locked Pentium II 400 refuses to do a multiplier above 4.0x, except in some rare occasions. There is also the confusion about how two people with Abit boards manage to run identical CPUs at 5.5x.

So the Abit mystery is mainly about setting multiplier 4.5x to 5.5x on a not-locked Pentium II 400. A mystery insofar I don't know how that works...

My Deschutes 400 ran with 5.5x multi on a Chaintech motherboard, jfyi 😀

I found some old pics and the date code of my unlocked Deschutes 400 SL2S7 was indeed 9830 and the other unlocked ones also have the correct datecodes. But alas I never made pics of the locked ones so I can't tell without retesting all these chips.

Tetrium wrote:

My Deschutes 400 ran with 5.5x multi on a Chaintech motherboard, jfyi 😀

Ah, so the mystery is not just an Abit thing.

Tetrium wrote:

I found some old pics and the date code of my unlocked Deschutes 400 SL2S7 was indeed 9830 and the other unlocked ones also have the correct datecodes. But alas I never made pics of the locked ones so I can't tell without retesting all these chips.

Thanks for checking. In case you have only one SL2VY and one SL2U3; retest would not be necessary.

gerwin wrote:

Ah, so the mystery is not just an Abit thing. […]
Show full quote

Tetrium wrote:

My Deschutes 400 ran with 5.5x multi on a Chaintech motherboard, jfyi 😀

Ah, so the mystery is not just an Abit thing.

Tetrium wrote:

I found some old pics and the date code of my unlocked Deschutes 400 SL2S7 was indeed 9830 and the other unlocked ones also have the correct datecodes. But alas I never made pics of the locked ones so I can't tell without retesting all these chips.

Thanks for checking. In case you have only one SL2VY and one SL2U3; retest would not be necessary.

Chances are good I have several of each stepping. Would make for an interesting chain of testing 🤣

But if it comes to that, I'll surely let you guys know 😀

True experts can buy FAKED Pentium II processors and downclock to the nominal frequency 🤣

https://books.google.cz/books?id=2AEAAAAAMBAJ … um%20II&f=false