Overclocking
| Related Externals |
| nVidia |
| PowerStrip |
| GeForce Roundup |
| Leinen Software |
On Processors
This story is pretty short, at least the "successes"-part of it. There have actually been a few experiments now that I think of it, it's just that none of them have had any positive results. Mind you, they haven't had any damaging results either, which can of course be put on the positive side ... The only manouvers I've attempted are the classig jumper setting, which naturally doesn't produce very much on recent Intel procesors ...
The P-133 have seen the most results of any attempts. In fact, it's the only computer that's seen any results. Seems it was old enough to allow changing processor speed using the jumpers, but new enough to at least restrict change upwards. Clocking down worked without a hitch, but that's not really all I had hoped for :-) ...
On Videocards
This is the bit of the page with anything even close to results. This little story begins with the Voodoo2 and the program I used to tweak it's settings, Voodoo2 Tweaker (that's a creative name!). Raising the clock speed quickly gave insight in the backsides of overclocking and the signs of said backsides. The side being heat and the sign being graphical errors. Turned out my dear Voodoo2 was quite adapt at producing heat and not all that good at getting rid of it. In short, it wouldn't overclock much at all. A few MHz generated tell-tale errors after a while, a few more produced a black screen and the need for a reboot ...
Quick jump to present day. 3dFX is gone and nVidia is on the performance throne and the GeForce is inside my computer. Somehow I suddenly found the interest for attempting overclocking again, and soon ventured to the right settings tab to try my luck. Seemed the thoughtful driver wizards had decided to protect the overzealous user from itself, in order to use a new setting you had to test it first. Fine, I thought, raised the speeds a little, and hit test.
Garbled screen, locked system. Reboot.
Right, I thought, appearantly that was too high. The only problem was that one single MHz of change was too high too. My suspicions about a little bug in the driver were easily confirmed, as an attempt to underclock the card produced the exact same crash ...
Change of tool followed. And yes, victory at last! The card happily accepted its new, minimally higher, speed. Of course, this created the a slight curiousity about just how much more I could push out of it. A few sites were consulted, and a few numbers were collected, to be sure I didn't go completely into the dark with my only graphics board. Extra cooling was achieved by opening the case and aiming my trusty table fan directly at the card. Cool, eh? The clock settings went up, and it didn't crash at all! Didn't even garble a single little pixel. In a rare more of caution, I considered the experiment a success and went back to more conservative settings after just a few timedemos. Not all the way back to zero of course, just so everything runs stable with the case properly shut ...
