VOGONS


Reply 40 of 505, by debs3759

User metadata
Rank Oldbie
Rank
Oldbie

Well - I'm not delidding my K6-2+/570 or K6-3+/550 😄

See my graphics card database at www.gpuzoo.com
Constantly being worked on. Feel free to message me with any corrections or details of cards you would like me to research and add.

Reply 41 of 505, by rasz_pl

User metadata
Rank Oldbie
Rank
Oldbie
Sphere478 wrote on 2022-01-29, 08:29:

For those just tuning in we are still waiting on someone to show up who has a k6-3+ or a k6-2+ and has de lidded it or is willing to. We need high res pics and resistance measurements.

those are 99.999% decoupling capacitors, not configuration straps

Reply 43 of 505, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
BitWrangler wrote on 2022-01-28, 16:30:

As far as I understood it, K6 2/3 plus were fabbed at the dresden plant as soon as it was upgraded to 180nm and it was the Duron that pushed them out of production there. Though they might have been making both concurrently for a bit but duron demand made them go 100% duron.

Ah, so I can still be right. The III+/2+ cores could've been pipe-cleaners but in that specific fab. It would be a bit weird if AMD had a working shrink of K7, they could've just used that, but perhaps it was a combination of several factors like the contracts I've mentioned, or design rules for K6 being more easily adaptable to new process, etc.

However I'd like to warn people NOT TO TRY any resistance measurement on the parts around the die, which all look like capacitors anyway. Not unless you know very well what you are doing. Most people would just use auto-ranging DMM and these will just charge these caps to their full open-circuit voltage, or rather to the point someting in the CPU starts conducting, and you might easily permanently damage the core this way. So unless you know how to setup and use a low-voltage (on these CPUs up to 0.5V or so) testing rig, keep your grubby fingers well away from these CPUs. And that's not even touching the subject of proper ESD protection.

Reply 44 of 505, by TrashPanda

User metadata
Rank Oldbie
Rank
Oldbie
NostalgicAslinger wrote on 2022-01-29, 11:28:

HERE are very high quality Die Shots of an AMD K6-III+ 400 ATZ and K6-2+ 500 ACR!

They are for the most part the exact same die, like the 2+ simply has fused off cache.

Oh noes, the cap let the shmooo out 😁

Reply 45 of 505, by cyclone3d

User metadata
Rank l33t++
Rank
l33t++

All of those pics that show the capacitors are labeled k6-2+.

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 46 of 505, by Sphere478

User metadata
Rank l33t
Rank
l33t
rasz_pl wrote on 2022-01-29, 10:19:
Sphere478 wrote on 2022-01-29, 08:29:

For those just tuning in we are still waiting on someone to show up who has a k6-3+ or a k6-2+ and has de lidded it or is willing to. We need high res pics and resistance measurements.

those are 99.999% decoupling capacitors, not configuration straps

If you look closer there are very tiny pads around them

NostalgicAslinger wrote on 2022-01-29, 11:28:

HERE are very high quality Die Shots of an AMD K6-III+ 400 ATZ and K6-2+ 500 ACR!

Ooo sweet! Lemme look!

cyclone3d wrote on 2022-01-29, 15:30:

All of those pics that show the capacitors are labeled k6-2+.

Yeah, the pics while cool don’t seem to give us a good view of a 3+ yet.

Deunan wrote on 2022-01-29, 11:40:
BitWrangler wrote on 2022-01-28, 16:30:

As far as I understood it, K6 2/3 plus were fabbed at the dresden plant as soon as it was upgraded to 180nm and it was the Duron that pushed them out of production there. Though they might have been making both concurrently for a bit but duron demand made them go 100% duron.

Ah, so I can still be right. The III+/2+ cores could've been pipe-cleaners but in that specific fab. It would be a bit weird if AMD had a working shrink of K7, they could've just used that, but perhaps it was a combination of several factors like the contracts I've mentioned, or design rules for K6 being more easily adaptable to new process, etc.

However I'd like to warn people NOT TO TRY any resistance measurement on the parts around the die, which all look like capacitors anyway. Not unless you know very well what you are doing. Most people would just use auto-ranging DMM and these will just charge these caps to their full open-circuit voltage, or rather to the point someting in the CPU starts conducting, and you might easily permanently damage the core this way. So unless you know how to setup and use a low-voltage (on these CPUs up to 0.5V or so) testing rig, keep your grubby fingers well away from these CPUs. And that's not even touching the subject of proper ESD protection.

Yeah, I’m kinda hoping someone comes along with some ones with busted pins or something that we can in good conscious start probing. Yeah, a dvm may actually cause harm to these chips. So at your own risk. But if we manage to start unlocking 570s woohoo!!

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 47 of 505, by Deunan

User metadata
Rank Oldbie
Rank
Oldbie
Sphere478 wrote on 2022-01-29, 19:36:

Yeah, I’m kinda hoping someone comes along with some ones with busted pins or something that we can in good conscious start probing. Yeah, a dvm may actually cause harm to these chips. So at your own risk. But if we manage to start unlocking 570s woohoo!!

A CPU with busted pins would be a good candidate for tests, yup. However the tiny golden dots you can see around the caps, and the core itself, are most likely not any test points or configuration jumpers but simply positioning marks for assembly purposes.

Reply 48 of 505, by Sphere478

User metadata
Rank l33t
Rank
l33t
Deunan wrote on 2022-01-29, 20:18:
Sphere478 wrote on 2022-01-29, 19:36:

Yeah, I’m kinda hoping someone comes along with some ones with busted pins or something that we can in good conscious start probing. Yeah, a dvm may actually cause harm to these chips. So at your own risk. But if we manage to start unlocking 570s woohoo!!

A CPU with busted pins would be a good candidate for tests, yup. However the tiny golden dots you can see around the caps, and the core itself, are most likely not any test points or configuration jumpers but simply positioning marks for assembly purposes.

🤔 let’s hope they are enabling pads for 1mb cache!!!🤪

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 49 of 505, by BitWrangler

User metadata
Rank l33t
Rank
l33t

Better start eyeballing those scrap listings real close... beware though some of the bulk bulk lots "kilogram of CPU" etc only have sample pics, not what you're actually getting.

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 50 of 505, by Sphere478

User metadata
Rank l33t
Rank
l33t
BitWrangler wrote on 2022-01-29, 20:23:

Better start eyeballing those scrap listings real close... beware though some of the bulk bulk lots "kilogram of CPU" etc only have sample pics, not what you're actually getting.

indeed, I wish those guys with piles of cpus on ebay would respond to requests for selective bulk buys.

Still trying to find people to help with high res pics, and voltage measurements of a delidded k6-2+ and k6-3+

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 51 of 505, by Sphere478

User metadata
Rank l33t
Rank
l33t

here are the measurements I need from a k6-3+ and a K6-2+

Attachments

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 52 of 505, by Tetrium

User metadata
Rank l33t++
Rank
l33t++

Regarding the K6/K7 thingy, I don't really understand why AMD even made the K6-2+ in the first place, unless AMD could disable part of the cache to sell off partially defective K6-3+ chips as K6-2+ chips. If this ended up actually happening is the question though as yields seem to have been really good.
K6+ was in my eyes clearly made for the laptop market. K7 was a power hog compared to K6+ and in laptops battery duration is all important to drive sales.
Perhaps K7 could actually be clocked lower, but it's conceivable that to reach K6-3+ 400 ATZ 1.6v power dissipation, Athlon would need to be clocked to such a low amount of MHz that this would hurt sales (because who back then would want to buy a (just pulling this out of my magic hat for a bit for explanatory reasons) 200MHz Athlon if a 400MHz K6-3 sounds faster (or Athlon would need to be clocked higher with the battery lasting less long as well?) and it's probably also cheaper on the chipset side. K6 was also old tech at the time, perhaps being more reliable, cheaper and easier to implement in laptops.
It was all about the megaherz back then, IPC didn't really exist as a concept at the time.

Deunan wrote on 2022-01-28, 11:29:
AFAIR the III+ chip dies were considerably bigger due to having extra cache on them, that would typically result in more load on […]
Show full quote
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?

AFAIR the III+ chip dies were considerably bigger due to having extra cache on them, that would typically result in more load on the internal signal paths so extra latency cycles were added to help with that. Depending on how it was done, and manufacturing process limitations, you could either get extra latency but higher clocks, rarely same latency with higher clocks, often same latency/clock or even regression but made up for by the cache being larger now.

Point is, there is no reason AMD (or any other chip manufacturer) would cripple their best dies, expensive to make on the best process available, to sell them as parts with lower margins. Not even Intel was that rich and stupid, except when the dies were so easy to make it wasn't much of a loss to rebrand them. AMD at that time had it factories running 24/7 to meet demand and they wanted to sell the latest and greatest Athlons for big bucks. I just don't see someone making an insane decission to forget that, make more III+ cores instead of Athlons and then rebrand them as 2+ to sell at lower price.

As I see it, the only III+ cores that would be rebranded to 2+ (if any) were either defective cache dies or didn't meet other specs (clocks or power draw at rated speed). Now, it might be possible to find a part that is fully functional, and can somehow be unlocked to full cache size, but will require more juice to run and will generate more heat. But that would be like winning a lottery ticket I think. But, good luck, I just hope you people won't kill a ton of 2+ chips trying to delid them to find that one III+ core.

One reason a CPU company would want to do this, is to flood the lower segments with cheaper CPUs to harm CPU sales of your main competitors. Intel kinda did actually do this and other companies probably did something like this as well.
Well it was not literally their very best dies, but they certainly did sell better parts underclocked for a lower price than it was arguably worth at the time, just to be as much of a hindrance to their competitors as they could.
Because of their competitors made less money, they would be less able to invest in faster CPUs and be less of a competition, indirectly increasing revenue.

Whats missing in your collections?
My retro rigs (old topic)
Interesting Vogons threads (links to Vogonswiki)
Report spammers here!

Reply 53 of 505, by Tetrium

User metadata
Rank l33t++
Rank
l33t++
NostalgicAslinger wrote on 2022-01-29, 11:28:

HERE are very high quality Die Shots of an AMD K6-III+ 400 ATZ and K6-2+ 500 ACR!

These photos look absolutely stunning!

But which ones are from the K6-3+ and which ones are from the K6-2+?

Whats missing in your collections?
My retro rigs (old topic)
Interesting Vogons threads (links to Vogonswiki)
Report spammers here!

Reply 54 of 505, by Sphere478

User metadata
Rank l33t
Rank
l33t
Tetrium wrote on 2022-02-18, 10:03:
NostalgicAslinger wrote on 2022-01-29, 11:28:

HERE are very high quality Die Shots of an AMD K6-III+ 400 ATZ and K6-2+ 500 ACR!

These photos look absolutely stunning!

But which ones are from the K6-3+ and which ones are from the K6-2+?

according to file names the de lidded ones are from a 2+

Sphere's PCB projects.
-
Sphere’s socket 5/7 cpu collection.
-
SUCCESSFUL K6-2+ to K6-3+ Full Cache Enable Mod
-
Tyan S1564S to S1564D single to dual processor conversion (also s1563 and s1562)

Reply 55 of 505, by BitWrangler

User metadata
Rank l33t
Rank
l33t
Tetrium wrote on 2022-02-18, 10:00:
Regarding the K6/K7 thingy, I don't really understand why AMD even made the K6-2+ in the first place, unless AMD could disable p […]
Show full quote

Regarding the K6/K7 thingy, I don't really understand why AMD even made the K6-2+ in the first place, unless AMD could disable part of the cache to sell off partially defective K6-3+ chips as K6-2+ chips. If this ended up actually happening is the question though as yields seem to have been really good.
K6+ was in my eyes clearly made for the laptop market. K7 was a power hog compared to K6+ and in laptops battery duration is all important to drive sales.
Perhaps K7 could actually be clocked lower, but it's conceivable that to reach K6-3+ 400 ATZ 1.6v power dissipation, Athlon would need to be clocked to such a low amount of MHz that this would hurt sales (because who back then would want to buy a (just pulling this out of my magic hat for a bit for explanatory reasons) 200MHz Athlon if a 400MHz K6-3 sounds faster (or Athlon would need to be clocked higher with the battery lasting less long as well?) and it's probably also cheaper on the chipset side. K6 was also old tech at the time, perhaps being more reliable, cheaper and easier to implement in laptops.
It was all about the megaherz back then, IPC didn't really exist as a concept at the time.

Deunan wrote on 2022-01-28, 11:29:
AFAIR the III+ chip dies were considerably bigger due to having extra cache on them, that would typically result in more load on […]
Show full quote
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?

AFAIR the III+ chip dies were considerably bigger due to having extra cache on them, that would typically result in more load on the internal signal paths so extra latency cycles were added to help with that. Depending on how it was done, and manufacturing process limitations, you could either get extra latency but higher clocks, rarely same latency with higher clocks, often same latency/clock or even regression but made up for by the cache being larger now.

Point is, there is no reason AMD (or any other chip manufacturer) would cripple their best dies, expensive to make on the best process available, to sell them as parts with lower margins. Not even Intel was that rich and stupid, except when the dies were so easy to make it wasn't much of a loss to rebrand them. AMD at that time had it factories running 24/7 to meet demand and they wanted to sell the latest and greatest Athlons for big bucks. I just don't see someone making an insane decission to forget that, make more III+ cores instead of Athlons and then rebrand them as 2+ to sell at lower price.

As I see it, the only III+ cores that would be rebranded to 2+ (if any) were either defective cache dies or didn't meet other specs (clocks or power draw at rated speed). Now, it might be possible to find a part that is fully functional, and can somehow be unlocked to full cache size, but will require more juice to run and will generate more heat. But that would be like winning a lottery ticket I think. But, good luck, I just hope you people won't kill a ton of 2+ chips trying to delid them to find that one III+ core.

One reason a CPU company would want to do this, is to flood the lower segments with cheaper CPUs to harm CPU sales of your main competitors. Intel kinda did actually do this and other companies probably did something like this as well.
Well it was not literally their very best dies, but they certainly did sell better parts underclocked for a lower price than it was arguably worth at the time, just to be as much of a hindrance to their competitors as they could.
Because of their competitors made less money, they would be less able to invest in faster CPUs and be less of a competition, indirectly increasing revenue.

Yes, they were certainly intended to keep a foot jammed in the door of the mobile market while a mobile Athlon wasn't ready. The argon, pluto and orion cores were unsuitable for power reduction, and at the 500-600 Mhz level mostly performed on a par with the K6-3+ due to having baked in pipelining and timing designed for higher speeds, kinda like the 286-16 386-16 thing. So there was no point clocking them further down. Thunderbird came along, and while it was good enough to push the clock speed on desktop, it wasn't quite good enough to get acceptable mobile power figures. The next stopgap as it were was the cache trimmed version, the Duron Spitfire core, which could with the hot running cache reduced, be made into a mobile model to patch the gap from 550 to 1Ghz. It wasn't until the Palomino core that a "proper" Athlon mobile existed, and maybe you can see how much they felt this mobile deficit, by the fact that palomino wasn't released on desktop, it was only available as mobile CPU for a number of months.

Where are they all? You might ask of the K6-2+/3+ laptops, or even the early Duron ones. Well I think the cap plague applied much slaughterage to the population in the wild, and also what I call the "everybody forgot how to make hinges" problem. Which is really a story of how you can't make mass market consumer priced laptops with a full magnesium alloy frame that you had in the 90s $3000 business models. So cheaper materials and lower part count in the consumer laptops had screens snapping off left and right, until they figured out how to do cheap and reasonably durable, or found the floor of the absolute minimum you could get away with. Which was actually the ceiling above what they were trying to do in 99-01. Another factor to that is that screens had got bigger and heavier, a little 9" didn't cut it any more, ppl wanted full 14", 15" panels which were quite heavy up to around late noughties when they began to get slimmer and lose mass.

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 56 of 505, by Fritzchens Fritz

User metadata
Rank Newbie
Rank
Newbie

Hi there!

Interesting topic, but I don't think it is possible to reactivate the cache. The cache will most likely have been fused off or disconnected via lasercut on the die itself. But the chance that it depends on the package configuration, can also be true. So, I have picked out my old chips and take a package photo of the K6-2+ and finally the III+. Just the capacitor on the bottom right corner is in a different position. Resoldering the capacitor on the K6-2+ would be the only thing I could try. I still have a working K6-2+.

https://flic.kr/p/2n4J2wx

Attachments

  • DSC05049##.jpg
    Filename
    DSC05049##.jpg
    File size
    132.94 KiB
    Views
    1185 views
    File license
    Public domain

Reply 57 of 505, by BitWrangler

User metadata
Rank l33t
Rank
l33t

Vielen danke Fritz. Many thanks... that's what we've been waiting for.

I may try it when I can order my 6, at least 2 of them are going to be in un-super 7 boards at 4-500, so won't matter if I end up with a K6-3 I have to clock slower to get cache 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 58 of 505, by The Serpent Rider

User metadata
Rank l33t
Rank
l33t

If we consider how "sloppy" AMD were with Slot A being fully unlocked via small attachment with switches and Thunderbird pencil mod, it won't be a big surprise, if replacing just one cap can change cache config. And these CPUs were not meant to be used in regular PCs anyway, plus you need to delid it.

Get up, come on get down with the sickness
Open up your hate, and let it flow into me

Reply 59 of 505, by Fritzchens Fritz

User metadata
Rank Newbie
Rank
Newbie

Looks like, replacing the capacitor is the way to active the full amount of 256KB L2 Cache. CPUz recognize my K6-2+ now as a III+ with 256KB cache. 3DMark2000 score is also higher. Will now place the capacitor back to the original position, just to be sure...