How it went...
Now for some action. To get the chips to 450MHz, I noticed AMD had to turn the core voltage up to 2.4v. Because these chips were specked for anywhere from 2.0v to 2.5v, this seemed perfectly OK. I followed suit with 110 x 4 = 440MHz. This however, seemed just as fast (or maybe just as slow) as the 120x3.5 I had earlier! It looked like more drastic measures where needed. I upped the Vcore to 2.6v and managed to crank the chip up to 473MHz (105x4.5).I had now noticed why there were so few K6-2 clocking articles. Its not an easy chip to clock! On a celeron, 25% isn't a great deal, however here, it had taken a great deal of effort. The flaming L2 cache was still only at 105MHz as well, holding things back. I was desperate for more and more speed, not realizing the damage I could cause.
I set my sights on around 500MHz. You know how you do, you pick a target and you do anything you can to get there. If you don't do it, you feel you have failed (whether or not the target was reachable never really matters). I *was* going to get this chip to 495Mhz (110x4.5). So, up went the Vcore again, to 2.8v. I was now 0.3v over AMD's maximum limit and 0.6v over the chips stated Vcore. The chip refused to post at 495. Up it went again (where had I put my brains at this point?) to 2.9v, and yes! It posted at 495! Windows even managed to boot and hold down a Quake II demo, before it died.
It still wasn't the magic 5-ton tho' was it? Inspired by this new-found speed, I set the Vcore to 3.0v and tried 115x4.5 (518). No luck. I set the Vcore to 3.1v (0.9v over the chips stated limit) and it all went very very wrong.
The machine did post at this insane speed, and I believe I had the fastest K6-2 400 for a few seconds. The screen then went all fuzzy, and some funny smells emitted from the PC, before a total lockout. I rebooted, and nothing happened. Hmmm, I turned the Vcore down a touch to my previous level and tried 495MHz. Nothing. I continued turning the clock and the Vcore down, until I was at 400Mhz, 2.2v. Nothing. The chip refused to even post, at *any* voltage, unless it was at 350MHz.
So there we have it. The poor K6-2 cooked itself and is now a 350Mhz part. All that effort, for all the extra speed, and now it's slower than when I started. :( Oh well, on with the Celeron :)
