[airspirit]

I always find myself needing/wanting more power. Whether it is long compilating sessions, running games faster with more eye candy, or just for general responsiveness of the system, overclocking is the shee-it. It also has a way of saving you money long term if you're like me and upgrade quite a bit, since you can go between larger jumps but keep the prices lower. For example, I started with a $100 1700+ that I raised from 1466 to 1733. That was replaced by a 1900+ ($100) when it burnt due to personal stupidity that I pushed over 1900 Mhz. After that I went to a 1700+ TBredB that clocked me as high as 2550 ... for $70. After that came the god chip for $50 (1700+) that got me to 2850, while my other chip went into a different box. I sold that for a small fortune and picked up a Barton 2500+ that runs at 2550 for $75, and that's where I'm at today. Mind you, if I was to have bought just one chip at retail prices with today's speed cap, I would have spent more on that one chip than all the chips that I've had for the last few years, mkay? (Mkay). The benefit to me is that I was constantly running at top-of-the-chart speeds during that time period (and in many cases well above the top speed range) but still kept my total hardware costs under the cost of one retail chip.

With this, however, comes other forms of insanity such as watercooling. Again, I see this as an investment. If you dump an added $200 onto the cost of the chips above (typical watercooling system from scratch) you still aren't much above the cost of that one retail chip (especially on the Intel side of the house) but you can swap it from system to system effortlessly as you upgrade.

Overclocking gets you more bang for your buck, and regardless of what the naysayers may hold as true, some of us do need the extra speed and reach for it even when we can't necessarily afford it.

As far as the risk of premature failure, you can look at it this way: if buying a package that can be replaced (such as a socket A system for the past three years), you are better off picking up and burning eight 1700+ chips running at godawful speeds over the course of a five year period than buying one 3200+ and running it stock both cost-wise and speed-wise. Of course, you won't go through that many (I've only burnt one chip and the rest were sold or bartered), but it is a good way of looking at it, isn't it? The only risk factor is your board ... though even if you fry a board (not common unless you're insane like me or just plain stupid) you still come out ahead cost-wise.

Are you starting to see the benefits yet? For those that clock a measly 5-15% speed over stock there won't be much benefit, but for those of us that cherry-pick chips and bottom-feed on the cost leaders (in an educated manner), there are only benefits to doing this (especially when you get into the 50-100% OC range like that 1700+ god chip I described above).

If you have doubts, then just don't do it. We won't laugh at you. Better to stay with the safe and normal than do something from an uneducated standpoint. If you know what you're doing, however, it is almost foolish not to push the envelope.