Do not fear, the multiplier is locked inside all coppermine CPU's (except engineering samples). So it disregards multipliers your motherboard supports or sets.
I think if your motherboard mentions coppermine CPU support, it supports them all.
Other then that I remember the 1133 was unstable, and the 1100 and 1000 will be much hotter then something like a 600MHz.
wrote:Deschutes are 2.0 Volt instead of 2.8. It would be nice If I could use a deschutes, they run cooler. At first I thought of getting a 350, but I think chances are better with a 333. Then pick the product numbers with the oldest revision (dA0 or dA1).
I just bought an SL2QF, so I can find out for myself when it arrives
It arrived, and I found out for myself. So it is a Pentium II 333MHz, 66MHz FSB, with MMX, Deschutes Core, 2.0 Volt, 512kB L2 cache at half speed. It is a retail processor in a solid plastic cartridge with a nice but worn Intel fan. I put it in a mainboard from Aopen, the AX6BC revision 1.4, it has a 440BX chipset of course.
Multiplier 5.0x, FSB 66MHz -> CPU runs at 333/66MHz, default setting.
Multiplier 1.5x, FSB 66MHz -> CPU protests* and runs at 133/66MHz instead.
Multiplier 2.0x, FSB 66MHz -> CPU runs at 133/66MHz. 155 pts**
Multiplier 5.5x, FSB 66MHz -> CPU protests* and runs at 133/66MHz instead.
Multiplier 4.0x, FSB 100MHz -> CPU runs at 400/100MHz, temp. is OK. 465 pts**
*CPU protests: warm rebooting does not work in this case.
** speedsys CPU benchmark points.
All other Pentium II/III/Celeron CPU's I have around here do only their default multiplier. So this P-II 333 is one of the most flexible types of Pentium CPU's. It accepts multipliers ranging 2x to 5x, the so called 'multiplier limited' Pentium. It has a speed range of 133 to at least 400 MHz.
At 133MHz and with the Throttle utility set to max slowdown I get 19.04 speedsys points (+-486DX50). But as I already said Throttle does not always work well with games and applications.
Edit: I read that due to L2 cache chip speed limitations it is not recommended to go over 400MHz with this CPU.
@gerwin
So i'm confused. According to the specs you posted, the PII 233 Klamath core should be multiplier unlocked, meaning you can run it at very slow speeds. Is this not correct? Also, what kind of performance does that PII @ 133Mhz give you? Is it near a Pentium 1 @ 133Mhz?
@retro games 100
You said you have a Klamath PII 233Mhz. Can you try and change the mulitplier? Will it actually go slower?
It seems to me that the PII 233Mhz would be the best thing to get, assuming it can, indeed, be unlocked.
Haven't done any real testing (bare board only, no disks etc), but tried to change multiplier just for the heck of it....
CPU: PII 400mhz (SL2S7)
http://processorfinder.intel.com/Details.aspx?sSpec=sl2s7
Mobo: Abit BH6
Tried with 2,0x66, and POST showed 133mhz... 😎
Mind you, I have no idea how fast it REALLY is working, it could very well(most likely) be running at 4x66mhz (locked multi?) for all I know.
But I'd then have to install some OS and Speedsys(?) for more accurate info... hmm, maybe later.
Oh well, at least a slower CPU is just a reboot away(no silly jumpers here, no sirr!), provided I have the appropriate CPU .
😁
I've only gotten my Klamath 233 down to 200 without a freeze. Anything below 200 or above 290 freezes me 🙁
@Farfolomew + leileilol
I think you can forget the pentium II Klamath core entirely. In my previous post I concluded: Best choice is an early Deschutes Core (example part numbers SL2KA or SL2QF)
wrote:So this P-II 333 is one of the most flexible types of Pentium CPU's. It accepts multipliers ranging 2x to 5x, the so called 'multiplier limited' Pentium. It has a speed range of 133 to at least 400 MHz.
The CPU's performance at 133MHz is depicted in the graph in my previous post: 155 pts, which is somewhat above a genuine Pentium 133 (This genuine P-133 is also in the graph). The L2 Cache was temporarely disabled. I played a few Dos games at that speed and experienced no problems.
leileilol, I am surprised to read the klamath freezes? are you changing the multiplier or the FSB? I read many Klamath Core CPU's should be multiplier unlocked, or at least be multiplier limited. However, I don't have any such CPU, and I don't need one. IMHO These things only use more power, and have a lower top speed.
Preferably you need a mainboard that allows you to set multiplier=2x in the BIOS or by using Jumpers. Like ASUS, ABIT and AOpen. Many other mainboards have a minimum of 3x. (Sometimes you can cheat by setting a higher multiplier like 6x which seems to revert back to 2x, not sure..)
@bestemor
Interesting, when I say early Deschutes core, I don't know exactly which ones are early enough. The SL2KA and SL2QF (stepping dA0) are early enough, maybe your SL2S7 is also an early release. I read it is the earliest in the 400 Speed range (stepping dA1). Please run speedsys to make sure the clock is actually slowed down. If so then this is another winner.
Look here for some additional info on this generation of CPU's:
Intel Slot 1 CPUs Uncovered
Ok then, surprise surprise !!
Now, I must admit I'm somewhat sceptical to that Speedsys program after seeing this, but...
Anyway, here are my stunning results:
Seems like a P2-400mhz actually is NOT locked, at least not for underclocking....?? very weird.... (!)
Haven't switched off any L1/L2 cache, so it may go even slower ?
(creation date for the report file looks a bit odd, 1999 etc, but I guess that's due to me not changing much in that bios)
PS: just tested L1/L2:
disable L2 = no effect
disable L1(L2 enabled) = veeery slow, benchmark score of a whopping.... 2.38 ! (386DX-40 country)
.
wrote:Seems like a P2-400mhz actually is NOT locked, at least not for underclocking....?? very weird.... (!) [...] PS: just tested L1 […]
Seems like a P2-400mhz actually is NOT locked, at least not for underclocking....?? very weird.... (!)
[...]
PS: just tested L1/L2:
disable L2 = no effect
That's no P2-400, but rather a Celeron-300.
That's no P2-400, but rather a Celeron-300.
What do you mean ??
Are you saying that the P2-400's really are Celerons ? 😕
Or is it the text output from Speedsys you are referring to ?
Well, either this must be a fake CPU - as outer markings printed on the plastic edge/cover say PY400512, and SL2S7.
(looks like it was made in week 17 1998)
Ooor.... Speedsys gets confused, due to the underclocking 😁
At least my CPUID is correct ('0651').
And... check the pic posted by 'gerwin' above, his P2-333 must also be a Celeron, obviously... heh....
Still, the lacking effect of disabling L2 in BIOS makes me wonder...
Maybe the test program has some unknown limitations ?
PS:
...just checked again, resetting to default speeds etc in BIOS.
POST now says 512kb cache, which is more like it.
And, Speedsys now correctly labels my chameleon CPU as a PII 400mhz Deschutes - thought test score is only 455.14, which seems to be slightly lower than an 'actual' P2-400(according to the Speedsys itself at least, blue line stops exactly before the 'P' in 'Processor benchmark')
Oh, and it also now shows both L1 and L2 sizes !
And disabling the L2 actually gives BETTER results (score=458.70) ??
And then... just changing the multi to 2x and FSB to 66mhz(with all cache still enabled in BIOS) results in POST showing cache = 'none' (!), maybe that's part of the problem ?
Well, are we confused yet ? 😵
.
wrote:Are you saying that the P2-400's really are Celerons ? 😕 Or is it the text output from Speedsys you are referring to ?
I was referring to the SpeedSys output. And the strange results with the L2 cache.
The SST CPU detection is rather good, at least with the CPUs I tested so far. Apparently it just reports the core as Celeron because lowering the multiplier somehow deactivates the caches... 😦
If it was possible to enable both caches AND run the CPU at 133 MHz, then the BX/Pentium2 platform would be a nice one for retro gaming after all!
Keep testing, guys, this is rather interesting... 😉
Seems like choosing a multiplier below 3x turns off the L2 cache for some strange reason....
Results from POST:
2x(or 2.5x) 100mhz = no L2
3x 66mhz = 512kb ....
Though Speedsys doesn't seem to care either way, giving out the same results(at same speed) no matter what L2 size.
BTW, lowest multi on my board is 2.0, no 1.5 sadly...
As for Speedsys reporting Celeron, tested with my 266mhz Klamath:
It STILL says it is a Celeron in the first line(if multi< 3x), but at least it now says 'Klamath' in the second line, probably due to the CPUID number/class.
But still no L2 cache at 2x....
@Gerwin
I found an SL2QF on ebay, but the FPO# is 98350465-0283. According to the guides i've found, that would place this CPU as being created in the 35th week of 1998, which is in September. I'm concerned that the multiplier will be locked. I'm looking at SL2KA processors, but none of them clearly show the FPO# in the pictures.
I suppose I could go with Bestemor's SL2S7 400Mhz processor too.
You keep mentioning stepping, but multiple SSPECs have the same stepping DA0 and DA1. Shouldn't I be looking for FPO#s less than August 1998, rather than stepping?
Thanks for the help!
Hmm, I've just had a thought that perhaps explains why Bestemor's proccesor has an unlocked muliplier.
Maybe it's not that ALL processors produced after August 1998 have locked multipliers, but rather all processors with NEW steppings than DA1 have locked multipliers. And that in August 1998 is when they went to the new stepping after DA1.
According to the Intel Proccessor finder, here are the DA0/DA1 stepping processors:
SL2SH 400 MHz N/A 1 100 MHz 0.25 micron DA1 512 KB S.E.C.C. N/A
SL2S7 400 MHz N/A 1 100 MHz 0.25 micron DA1 512 KB S.E.C.C. N/A
SL2S6 350 MHz N/A 1 100 MHz 0.25 micron DA1 512 KB S.E.C.C. N/A
SL2SF 350 MHz N/A 1 100 MHz 0.25 micron DA1 512 KB S.E.C.C. N/A
SL2S5 333 MHz N/A 1 66 MHz 0.25 micron DA1 512 KB S.E.C.C. N/A
SL2QH 333 MHz N/A 1 66 MHz 0.25 micron DA1 512 KB S.E.C.C. N/A
SL2KA 333 MHz N/A 1 66 MHz 0.25 micron DA0 512 KB S.E.C.C. N/A
SL2QF 333 MHz N/A 1 66 MHz 0.25 micron DA0 512 KB S.E.C.C. N/A
SL35V 300 MHz N/A 1 66 MHz 0.25 micron DA1 512 KB S.E.C.C. N/A
All .35 micron PIIs have a stepping of Cxx, whereas all .25 micron PIIs have a stepping of Dxx. Furthermore, most higher MHZ .25 micron PIIs have steppings of DBx.
So, starting with stepping of DBx (which may have started in August of 1998?) all processors would be multiplier locked. Does this make sense?
Sorry if this doesn't make sense; i'm just trying throwing out ideas hehe.
@bestemor
Can you try running the processor at 66 FSB with a 6.0 multiplier? I'm curious if your PII will allow you to adjust the mulitplier higher than the default 5.0. If so, then all I have to do to run at different speeds is to simply adjust the multiplier from 2.0 ~ 6.0, rather than having to change two settings (FSB and multiplier).
Thanks!
wrote:@retro games 100
You said you have a Klamath PII 233Mhz. Can you try and change the mulitplier? Will it actually go slower?
Um, sorry, I haven't got this thing immediately available to hand right now. 😦
@retro games
NP, thanks for checking back though 😀. I ordered a Klamath 233 for a few dollars on Ebay, so I should be getting that soon. However, I now wish I spent some more time investgating, as i'm about to buy a Deschutes 400Mhz (or less, still deciding).
@bestemor
Can you tell me what your FPO# is on your processor? It's the 2nd line of that long string of characters printed on the top of the slot cartridge. If yours was produced before Aug 1998, and so was Gerwin's, then that would mean the stepping has nothing to do with it, and it's just the Date of Manufacture. Thanks.
Can you tell me what your FPO# is on your processor?
Well, I DID say it was made in week 17 in my earlier post, didn't I ?.... 😎
So, 9817etc it is.
Can you try running the processor at 66 FSB with a 6.0 multiplier?
As the bios is quite old, there IS NO 6,0x(or higher) multiplier....
And trying to run 5,5x, weell...... it reverts to a unruly 2x mode, at least POST shows 5,5x like only 133mhz, and L2 cache is gone <sob>
So I have a feeling that your plan won't work quite like you hope, only changing the multi.
(3x-5x and between works fine though)
---
PS: tested with a second mobo, same model but newer bios:
Now I get 5,5x 66mhz to work, gives 366mhz AND 512kb.
6x 6,5x 7x etc all refuses to start (black screen), and after a hard reboot, it actually POSTs, but shows speed as 200mhz...
But when entering bios, my old setting is still there(6x66), weird.
Anyway, if you want the full 400mhz, seems like you'd have to go with 4x100.
And overclocking is not that great either, tried 112x4, but no-go.
.
Hmm, I missed that about Week 17. Sorry!
And I guess that helps explains things. It rather seems that going by the stepping alone (DA0 or DA1) is not enough to ensure an unlocked multiplier. You have to also make sure it was produced before August 1998. I did manage to find one 333 on ebay that was produced in Week 18, so I think i'll buy that. I'm also going to take a chance on a cheap $5 one, and report back my findings. I'm unsure of it's data of manufacture, but it's a 350Mhz with a DA1 stepping.
As for the 6.0x multiplier, I forgot how high my Abit BH6 board went. I thought it'd do 6.0...whoops! It sounds like the PII 400s won't do 450Mhz anyways, so if I need a faster chip than a PII 333Mhz, i'll throw in my good ol' Celeron 300a @450Mhz 😀
Well, I ended up purchasing a PII 350Mhz SL2S6 (DA1 stepping) that--I THINK--was produced in the last week of July '98 (the pic of the FPO# gets blurry right after the '3' hehe). According to a few websites, it looks like Intel implemented the actual multiplier lock on processors made AFTER August 19th, 1998, which is week 34. Any processor made in week 33 or earlier seems to be alright. So I'm hoping I have at least a ~20 day grace period 😜. I'll report back my (hopeful) successes!
On a side note, here's a guide to disabling the multiplier lock altogether:
http://www.dwpg.com/content.php?contid=2&arti … =89&orgartid=89
(I had no time to reply earlier)
Nice to read other peoples observations on this topic. Although it is rather disturbing that the L2 cache was not disabled by me at 133Mhz, but as an undesired side effect. I never really noticed this before, also because I was fairly new to speedsys. I fear there is not much one can do to fix this? Fortunately when slowing the system you might actually like the L2 cache being disabled. But still, it ruins the idea of having a Pentium-133 clone on your slot-1 board.
Actually my test setup is disassembled now, so I have to put it together again when I want to try things out.
Everyone please watch out with the term 'unlocked multiplier'. Most of the time it should say 'a limited multiplier'.
It is not suprising this subject was hardly documented in past, because in only a few cases a limited multiplier has use for overclocking. With the P-II 333/66 it was useful for overclocking because you could put the FSB to 100MHz and then lower the multiplier to stay at a reasonably CPU speed. With the P-II's that already have a 100MHz FSB (the P-II 350 and 400) the limited multiplier was useless.
wrote:So I have a feeling that your plan won't work quite like you hope, only changing the multi. (3x-5x and between works fine though […]
So I have a feeling that your plan won't work quite like you hope, only changing the multi.
(3x-5x and between works fine though)PS: tested with a second mobo, same model but newer bios:
Now I get 5,5x 66mhz to work, gives 366mhz AND 512kb.
This is not what I expected from a P-II 400MHz. I thought a P-II 333/66 can potentially be adjustable up to 5.0x whilst a P-II 400/100 does 4.0x max. But your feedback indicates a net-speed derived multiplier limitation, not just a fixed multiplier limitation.
I don't know if the stepping numbers are a foolproof guide to getting a multiplier limited CPU, or if the date-codes should be considered. My SL2QF says 98290382-0363. (We will know when we have knowledge of two identical part numbers, of which one is multiplier limited, and the other totally locked..)
Farfolomew, don't get your hopes up on that unlocking article. I know how it ends. 😜
I actually made a deal for an unlocked engineering sample coppermine CPU some days ago. I am still waiting for the other guys bank account nr. But he does no longer respond since 24-feb. 😠 Good chance the deal is off. He should hand it over quickly for the sake of scientific research. 😉
