I hope you have a very good cooling system, and I'd not run it for extended periods of time regardless, but congratz!
Btw isn't the 2x multiplier remapped to 6x on those chips? You could get it to 400MHz that way.

Believe it or not, with stock Intel's 370 cooler it gets barely warm. Same behavior than p55c 200 mhz. I ran prime95 fft for about an hour and it's super stable!
I'll try 2x setting. If it works that will be too damn impressive! It's the first time I see something like this.

I can't believe it... It's running stable at 400 mhz!!!!! 😵 😵 😵

So, you are basically running a 2.2V CPU at 3.3V?

That does take some guts. And here I was scared to run my K6-2 @2.8V (my board doesn't have any voltage options between 2.2V and 2.8V). 😉

It's not totally surprising that it runs stable at 400MHz. More voltage makes the chips more stable, not less. That is, until they get too hot and burn. Since the K6 has no overheating protection like modern CPUs, I'd say you better monitor the temperatures closely, and not overdo this, because at the end of the road you have a toasted chip. 😜

Perhaps he could test for us to see if the theory of electromigration is true or not? 😁

Some Mainboards autodetect Voltage,
what mainboard do you use ??

The original 0.35u K6 233Mhz at 3.3v still probably wins in the cooling requirement. 😁

386SX wrote:

The original 0.35u K6 233Mhz at 3.3v still probably wins in the cooling requirement. 😁

Probably, but the original K6 should also have a larger die size (since K6-2 is basically the same thing but with a die shrink), which means it should be able to dissipate more heat before it fails.

matze79 wrote:

Some Mainboards autodetect Voltage,
what mainboard do you use ??

Probably this: FIC PA-2001

Tetrium wrote:

Probably, but the original K6 should also have a larger die size (since K6-2 is basically the same thing but with a die shrink), […]
Show full quote

386SX wrote:

The original 0.35u K6 233Mhz at 3.3v still probably wins in the cooling requirement. 😁

Probably, but the original K6 should also have a larger die size (since K6-2 is basically the same thing but with a die shrink), which means it should be able to dissipate more heat before it fails.

matze79 wrote:

Some Mainboards autodetect Voltage,
what mainboard do you use ??

Probably this: FIC PA-2001

Interesting point, have anyone seen both core without the metal cover?

By the way from my tests the K6-233 really need eveything you'd use to cool a socket 7 system. Ok 30W of TDP is not a "Barton" but definetely the highest point they probably could reach at that time.

386SX wrote:

Tetrium wrote:

Probably, but the original K6 should also have a larger die size (since K6-2 is basically the same thing but with a die shrink), […]
Show full quote

386SX wrote:

The original 0.35u K6 233Mhz at 3.3v still probably wins in the cooling requirement. 😁

Probably, but the original K6 should also have a larger die size (since K6-2 is basically the same thing but with a die shrink), which means it should be able to dissipate more heat before it fails.

matze79 wrote:

Some Mainboards autodetect Voltage,
what mainboard do you use ??

Probably this: FIC PA-2001

Interesting point, have anyone seen both core without the metal cover?

I did delid a K6-2 once (I was curious and it was defective anyway) and without it's metal heatspreader it very much reminds me of a sA Thunderbird, but with a smaller die. Never delidded a 0.35µ K6 though.

dr_st wrote:

So, you are basically running a 2.2V CPU at 3.3V?

That does take some guts. And here I was scared to run my K6-2 @2.8V (my board doesn't have any voltage options between 2.2V and 2.8V). 😉

It's not totally surprising that it runs stable at 400MHz. More voltage makes the chips more stable, not less. That is, until they get too hot and burn. Since the K6 has no overheating protection like modern CPUs, I'd say you better monitor the temperatures closely, and not overdo this, because at the end of the road you have a toasted chip. 😜

A K6-2 500 is common enough and cheap enough to experiment with.

Just don't try it with a K6-2+

I have a FIC PA-2001. Is the same rig I'm restoring. I found these K6-2 yesterday and I got curious jajaja. I'm sure my motherboard doesn't have autodetect feature.
The only issue I saw is, again, the voltage regulator. At 400mhz is burning hot. At 233 mhz is ok (same as P55C 200) but with the benefit that a k6-2 233 is closer to a Pentium II 233 than a P55C 233.

I will return to the P55C 200 (and maybe next week I will buy a P55C 233). But for experimental purposes this was so awesome! I never saw something like this before. Maybe it serves to encourage some other people to do some experiments.

Tetrium wrote:

Perhaps he could test for us to see if the theory of electromigration is true or not? 😁

I agree, the K6-2 wouldn't survive that voltage for long. The absolute maximum voltage any K6-2 can be run without destroying it (Core Voltage Upper Limit) is 2.6V.