AdamP wrote:

http://i825.photobucket.com/albums/zz180/adamplenty/P1010007.jpg […]
Show full quote

Tetrium wrote:

AdamP wrote:

Well, I can't try the SL2KA because I think the heatsink's on the wrong way around so I can't plug the CPU in and I can't remove it..

Sounds strange...
Could you take a pic please?

Definitely looks like it's on the wrong way around to me.
[/img]

Maybe it's meant to be snapped in two like Kit-Kat?

😜 jk -- you must detach it then -- is there any screwhole or some kind of a clip in the back?

if not i'm afraid you have to snap it open... maybe wrap-in-a-plastic-bag-and-put-it-in-the-freezer trick?

Sure you can remove it. It is attached by 4 Screws in the side of the cooler. These screws may look like small bolts, but they have a telling hexagonal core, which fits an inbus key. I was fooled first and broke open a locked CPUs case till I found out.

If you are lucky, you can even turn it upside down.

I've already removed the screws. But I just can't get the heatsink to budge. All I've managed to do so far is open one side of the case.

Not sure if this'll help but this is what it looks like:

My guess is someone simply mounted the cooler on backwards.

Could you post a pic of what the front of the CPU now looks like?
I've removed coolers from these processors before but often those oem coolers are a pain to remove without breaking anything.

No, that is the way it was to be. Give us a picture from the cooler side.

Are these what you want?:

Do different PIIs have different heatsink locking mechanisms? I removed the heatsink on my SL2U6 without much difficulty, but it wouldn't fit properly on my SL2HE so I had to put it back.

AdamP wrote:

Do different PIIs have different heatsink locking mechanisms?

Yup, theres the earlier fully enclosed P2's and the 'half open' model that was also used on the slot 1 P3's.

Have you succeeded in removing that heatsink now btw?

Sorry to jump in, just a quick question...

What CPU equivalent can emulate with a Pentium 2? E.g. enabling caches...

Anyone disabled the caches on a PII? Will it be even slower than P IIIs or are they quite different...

Take a screwdriver and pop the cooler off, it is probably glued.

The P2 without cache can run as slow as an XT

ux-3 wrote:

Take a screwdriver and pop the cooler off

That worked! 😀 Thanks

But I couldn't fit it on the right way around. I also tried the heatsink from my SL2U6, which also didn't fit properly. I ended up using the heatsink from my SL2S5. My guess is that the heatsink was meant for another PII.

The CPU isn't locked 😀 although I had to up it to 200mhz to keep the L2 cache. Why doesn't the L2 cache work at 166mhz when it did for the SL2HE? (My wild guess is that the multiplier has to be higher than half the default)

Is what this says true? If so, it seems that it was produced week 23 1998 (and my SL2S5 was produced week 42 1998 and my SL2HE was produced week 24 1998).

Thanks for your help

Quick question, I would love to know what applications do you under clock these cpus for?

I got the impression that they are to fast to "turn into" a 386 and when disabling L1 cache you end up with a XT machine?

Mau1wurf1977 wrote:

Quick question, I would love to know what applications do you under clock these cpus for?

I got the impression that they are to fast to "turn into" a 386 and when disabling L1 cache you end up with a XT machine?

As I told you before, a number of Windows games are speed sensitive:
Magic the Gathering's Shandalar comes to mind immediately.

As I have also answered you before, the 486 region is a blindspot you won't be able to hit with a Slot1 CPU in hardware. That is why I have adopted a real 486.

There is however a utility that will help you. It is called "Throttle". I have only used it shortly, until I discovered that joysticks don't work well with it. If your game needs no joystick, this is an option.

If you only disable L1 cache but keep the clock up at about 400 MHz, you are getting a 286 for many purposes.

No worries thanks for the info! I'm in "386 mode" so I might have missed your other comments regarding pentium and 486...

Yes I saw your post about throttle and joustick drift. Definitly an issue for me as love Wing Commander!

Cheers for the input!

Ok, I've got today SL2SP Deschutes 333 (stepping dA1), I though it is one of the earliest steppings, so it will be unlocked. No luck. 🙁 It is locked! So only dA0 is safe.

What matters is production week, not stepping. Only CPUs before August 1998 had unlocked multiplier.

W.x. wrote on 2021-03-04, 11:58:

Ok, I've got today SL2SP Deschutes 333 (stepping dA1), I though it is one of the earliest steppings, so it will be unlocked. No luck. 🙁 It is locked! So only dA0 is safe.

You also have the Date Code?

gerwin wrote on 2021-03-04, 13:47:

You also have the Date Code?

Yes, mine is 0832. Which is pointing just at august. Damn, I probably missed it by week probably.

https://www.cpu-world.com/forum/viewtopic.php … ocked+deschutes

Seems to be date rather than stepping, yes.

[Edit]: Or maybe the span in production dates is due to different phasing in of new stepping in different production plants ..

You probably may have luck within 1-2 weeks depending on CPU factory. But I have no idea which factory implemented multiplier lock first.

Just thought I'd share, I was playing around today with a motherboard from an old Compaq Deskpro along with a P2 Klamath 300/66 processor.

The board has a 6-switch dipswitch for setting the multiplier as well as a few other things. Switches 2 through 4 control the multiplier.

As documented in the manual for this computer, the clock settings are as follows:

1100   	3.0x	200
2110	3.5x	233
3001	4.0x	266
4011	4.5x	300
5101	5.0x	333

Note that there are 8 possible positions and thus 3 that are not documented as follows:
000
010
111

So I tested the first 2 and found that:

000 results in a 2.0x multiplier and a 133 MHZ speed (also disables the L2 cache)
010 results in a 2.5x multiplier and a 166 MHZ speed (L2 cache remains enabled)

...at least as reported by the BIOS. Now I haven't installed an OS yet for testing stability at these lower settings, but it does at least appear that this can be clocked all the way down to 133 MHZ just via mobo switch settings.

I'm not brave enough to test 111 - I suspect it could possibly be a 5.5x multiplier and I don't want to overheat my CPU.