VOGONS


SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Topic actions

First post, by Sphere478

User metadata
Rank l33t
Rank
l33t

EDIT: this thread has come a long way since my first post We are now seeing that it’s probably a bad idea to try this on chips that weren’t binned as high speed chips. The 570 2+ chips seem to have very good luck while the chips even 70 mhz slower seem to be having trouble. All in all it seems to be luck of the draw.

On a personal note, I’d like to emplore you guys to not try this mod on your rare chips. At the time of writing this there are thousands of 570 2+ chips on ebay right now that are not only not rare, but are also having very good success Rates with this mod. If you do do this mod please read from the link below all the way to the end of the thread before doing so. And please do reinstall the IHS so that the core is not damaged by your heatsink. They are very delicate. I think no matter how good of a method we refine for this mod the major challenge will be the actual soldering. The processor sucks the heat out of your iron when you try to do this mod, moving that resistor is not a easy thing to do. Please get help from a professional if you need it.

If you try this mod, we need your help. Please report the following information so we can gather data to help future people who want to try this.
Please include (if you can):

-cpu-z before and after screenshot
-max stable clock
-full processor model
-minimum stable voltage at max over clock,
-wether or not stock clock and voltage was stable
-any other notes of interest

Here is a collection of results for specific chips:
Admins/mods you may add new entries to this list as time goes on if I have become inactive, or have failed to add one.

AMD-K6-2E+/350AUZ
Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

AMD-K6-2E+/350IUZ

AMD-K6-2E+/400ACR

AMD-K6-2E+/400ATZ

AMD-K6-2E+/400ICR

AMD-K6-2E+/400ITZ

AMD-K6-2E+/450ACR

AMD-K6-2E+/450APZ

AMD-K6-2E+/450ICR

AMD-K6-2E+/500ACR

AMD-K6-2+/450ACZ
Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

AMD-K6-2+/450ACZM

AMD-K6-2+/450ADZM

AMD-K6-2+/475ACZM

AMD-K6-2+/475ADZM

AMD-K6-2+/500ACZ

AMD-K6-2+/500ACZM
Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

AMD-K6-2+/533ACZ
Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

AMD-K6-2+/533ACZM

AMD-K6-2+/550ACZ

AMD-K6-2+/570ACZ
Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

-

-

-

-
—————————————————————————————
Some links for those who want to avoid reading the whole thread:

Action starts here: Re: (K6-2+/3+ OWNERS SEE THIS) possible Cheap k6 3+ 570 chips.

Fast forward to Mod guide here:
Re: SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod

—————————————————————————————
<End of update>
-

-

-

-

-

-

-

OP:
Have an idea,

Need help confirming,

I’m wondering if it may be possible to enable the 256k cache on k6 2+ 570 cpus by de-lidding them and making resistor/pad mods

It’s not so crazy considering the socket athalon/durnon cpus from amd had mods like this right after the k6 line

I need help tracking down pics of a 2+ and 3+ without a lid on them. So we can visually compare.

Last edited by Sphere478 on 2022-04-17, 18:43. Edited 55 times in total.

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 1 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t

I think this is a pic of a non plus sharptooth.

🤔

Need pic pf 2+ and 3+ though.

Attachments

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 2 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t

Found these images, no indication of what the chips are but it’s clear that heat is needed to remove the chip without destroying it

Attachments

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 3 of 491, by Repo Man11

User metadata
Rank Oldbie
Rank
Oldbie

I can provide a photo of a delidded K6-2+, but not a K6-3+.

I've taken the IHS off of several K6-2+ CPUs over the years, but the actual increase in clock speed I achieved by doing so has been pretty minimal. Perhaps I've been lucky, but by using a single edged razor on all four corners the shield has come right off with no damage to the core.

"Everyone is ignorant, only on different subjects." - Will Rogers

Reply 4 of 491, by BitWrangler

User metadata
Rank l33t
Rank
l33t

When they were less than 5 years old, the non-plus K6-2 had a "wet" thermal paste on the die, the corners were more like hard grey silicone or slightly rubbery epoxy or something, that you could rock a razorblade around to eventually slice through. However, the dies were still on the crunchy side, and I've seen pics of durons, thunderbirds, coppermines have their dies get ripped up from thermal paste stiction and rough HSF removal, so not sure what to make of the pics here. Don't know whether they are worst case hamfisted spreader removals (not a fan of the clamp it in a vice method) or extremely careful ones by ppl who knew what they were doing that failed anyway.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 5 of 491, by BitWrangler

User metadata
Rank l33t
Rank
l33t
Repo Man11 wrote on 2021-12-09, 13:56:

I've taken the IHS off of several K6-2+ CPUs over the years, but the actual increase in clock speed I achieved by doing so has been pretty minimal.

Yah that's what I've found in general on prior K6 class, doesn't get you higher POST speeds, may stabilise marginal overclocks. Probably worth about "half" a speed grade, ~25Mhz. So if highest stable is say 475 and you can POST 500 but it's crashy, then maybe spreader delete and decent sink give you 500 stable.

Unicorn herding operations are proceeding, but all the totes of hens teeth and barrels of rocking horse poop give them plenty of hiding spots.

Reply 6 of 491, by Repo Man11

User metadata
Rank Oldbie
Rank
Oldbie

This is a decapped K6-2E+ 570.

Attachments

"Everyone is ignorant, only on different subjects." - Will Rogers

Reply 7 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t
Repo Man11 wrote on 2021-12-10, 03:29:

This is a decapped K6-2E+ 570.

Sweet! Can we get a high res of the middle all cleaned up showing all the tiny gold pads?

Eventually hopefully someone shows up with a 2+ and a 3+ and can do some continuity measurements I’m wondering if amd didn’t pop a fuse in the package in order to kill the 128k of cache. Maybe it can be re bridged

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 8 of 491, by cyclone3d

User metadata
Rank l33t++
Rank
l33t++

I have both K6-2+ and K6-III+ CPUs but not sure I want to decap any of them.

Do we really think that the K6-2+ and K6-III+ were the same die?

Maybe I will as they are really easy to decap unless they are wildly different than decapping regular K6-2 CPUs.

Yamaha YMF modified setupds and drivers
Yamaha XG resource repository - updated November 27, 2018
Yamaha YMF7x4 Guide
AW744L II - YMF744 - AOpen Cobra Sound Card - Install SB-Link Header

Reply 9 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t
cyclone3d wrote on 2021-12-10, 05:35:

I have both K6-2+ and K6-III+ CPUs but not sure I want to decap any of them.

Do we really think that the K6-2+ and K6-III+ were the same die?

Maybe I will as they are really easy to decap unless they are wildly different than decapping regular K6-2 CPUs.

We got a pic of the 2+ so let’s see a 3+ and go from there.

If you have a caliper and can measure the die that would help

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 12 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t
kixs wrote on 2021-12-10, 09:36:

I might help. I think I have one dead K6-3+ 400ATZ. Will check tonight...

what did you find?

Repo Man11 wrote on 2021-12-10, 03:29:

This is a decapped K6-2E+ 570.

do you see any places on that that may make sense to get measurements from?

like places they may have used current to blow links to sever the cache?

I'm hoping that when in the factory before the heat spreader that they all basically were the same until amd did something

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 14 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t
kixs wrote on 2022-01-20, 07:16:

As it happens I didn't get to it yet. But I'm also not sure if you still need it?

yes we haven't found anyone with a de lidded 3+ yet

may call on you for some measurements also

need one or two people with a 2+ and a 3+ and ability to upload high res pic and take measurments

we are hoping to figure out how the cache was disabled and turn 2+ into 3+

I would help but I don't want to de lid any of mine (yet)

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 15 of 491, by majestyk

User metadata
Rank Member
Rank
Member

Intel was selling Pentium CPUs that had defect(s) in their on-chip L2 caches as Celerons by deactivating the half with a defect.
What if AMD did the same with their III+ / 2+ CPUs? "Reactivating" the missing half would result in a CPU with unusable L2 cache.

Reply 16 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t
majestyk wrote on 2022-01-20, 07:40:

Intel was selling Pentium CPUs that had defect(s) in their on-chip L2 caches as Celerons by deactivating the half with a defect.
What if AMD did the same with their III+ / 2+ CPUs? "Reactivating" the missing half would result in a CPU with unusable L2 cache.

Enable it, find out, disable if no go. 😀

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 17 of 491, by majestyk

User metadata
Rank Member
Rank
Member

Chances are that AMD followed a price discrimination policy and sold faultless "castrated" III+ chips as 2+ to meet the demand for cheaper mobile CPUs.

I´m not sure if I remember this correctly but didn´t a larger cache (in the same CPU) mean a lower maximum CPU speed?

Reply 18 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t
majestyk wrote on 2022-01-20, 07:57:

Chances are that AMD followed a price discrimination policy and sold faultless "castrated" III+ chips as 2+ to meet the demand for cheaper mobile CPUs.

I´m not sure if I remember this correctly but didn´t a larger cache (in the same CPU) mean a lower maximum CPU speed?

Well, since 2+ chips have the mhz records on socket 7 yeah, it seems less cache more clock.

I wonder if these overclockers sometimes disabled cache entirely?

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of

Reply 19 of 491, by Sphere478

User metadata
Rank l33t
Rank
l33t

Still haven’t been able to find anyone who owns a k6-3+ that has been de lidded or anyone willing to do it. Anyone? I just can’t bring myself to do it to my 550 🤣

SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)
-
Socket 5/7/SS7 (Motherboard) Tweaker
-
The coolest socket 7 motherboard that you’ve never heard of