Overclock.net › Member Blogs › More Thoughts On Quot Stability Quot

More Thoughts On Quot Stability Quot

The question of stability in computers is a bit of a contentious point, I'd say. The best way I can express it is that it is an objective condition defined subjectively. In other words, there is such a thing as "true" or "absolute" stability, but because it seems to be a bit of an idealized condition, the best overclockers can shoot for is the closest possible approximation of this ideal.

In truth, I don't think absolute stability in computing is actually possible. There are simply too many variables, too many factors, which can compromise operational stability. There are a million billion individual pieces (i.e., transistors) trying to work in concert, and if just a few (or even one, if it's the critical one) craps out and causes a calculation mistake, the whole notion of stability falls apart.

However, even knowing this we as computer builders (I personally prefer the term "system integrator") still must strive towards creating as stable a machine as possible. Whether it's for our personal machine or for a client's, we must exercise the proper discipline and work hard towards assuring operational stability as much as possible.

This is especially true for overclocked machines. By definition, our overclocked computers have parts that are running at speeds beyond spec. While they may still be operating within the limits of safety, the faster we force these machines to go, the closer they get to these limits, and the smaller that margin of safety becomes. You get too close to the edge of stability, you definitely increase the risk of falling over to oblivion.

It's personally disturbing to see people with such flippant attitudes towards overclocking and achieving some modicum of stability. Why should I care so much about why they talk about not doing enough (if anything at all) to make their machines stable? Well, it's one thing to be a benchmark junkie; I can understand the idea that they simply want ultimate performance. They don't actually use their computers for anything remotely important. For most other people, though, even for people who just use their computers for gaming, stability is important. A crash is a crash, and system corruption is a headache for anyone who relies on his/her computer in any way.

The danger is that there are far too many people who don't care enough about overclocked stability, and that these people often are louder and more brash than more sensible folk. Let's just face it: It's hard WORK to make a machine stable. It's also a blow to one's ego (if you're inclined to think in these terms) to willingly slow your machine down just so it can be more stable than its actual top speed limit. Testing and tweaking is likely boring to most people. I can understand all these "negatives" to the imperative of achieving operational stability while overclocked.

Nevertheless, I personally cannot live with an unreliable machine. A reliable machine is one over which I have complete mastery; it will do what I want it to do. In contrast, an unreliable machine makes you do what you need to do. It's a role reversal I personally detest. I will never be any machine's slave. I will rely on the machine, and not the other way around, if I can possibly help it.

Thank you so much for reading. I appreciate your time and attention. As always, I invite comments and discussion.

Comments

There are no comments yet
Overclock.net › Member Blogs › More Thoughts On Quot Stability Quot