Chapter 5. Maximizing Performance

Although your computer spends 99.9% of the time waiting for you to do something, what concerns us is that other 0.1% of the time when eight seconds can seem like an eternity.

A common misconception is that a computer with a fast processor, say 1.6 GHz, will automatically be faster than an 800-MHz system. While the increased processor speed is an obvious benefit in some specific circumstances, such as when performing intensive statistical calculations, using 3D modeling software, or playing some high-end games,^[1] the real-world performance of a computer is really measured differently.

^[1] See Section 5.1.5 later in this chapter.

In most cases, our qualitative assessment of a computer's speed is based on its ability to respond immediately to mouse clicks and keystrokes, start applications quickly, open menus and dialog boxes without a delay, start up and shut down Windows quickly, and display graphics and animation smoothly. For the most part, all of these things depend more upon correctly optimized software, the amount of installed memory, and the amount of free disk space than on raw processor power.

Because financial limitations prevent most of us from simply buying new hardware every three months, most of this chapter is devoted to solutions that will help improve the performance of your existing system without requiring any additional monetary investment. For example, the way Windows uses the swap file (virtual memory) can be inefficient, and dealing with this bottleneck can result in performance increases all across the system. In your approach to these problems, consider that your computer has a given theoretical top speed, so all you need to do is fix whatever is slowing it down so that you can approach that speed.

Upgrading, discussed at the end of this chapter, is a somewhat different story there, your aim is to raise the theoretical top speed. Start by asking yourself where your money is best spent, which isn't always obvious. For example, there are always faster processors available, but often something as simple as adding more memory can have a much bigger impact on performance for much less money.

Now, I'm the last one to condone throwing money at a problem. Even if money were no object and we could simply buy a new computer or component whenever the proverbial ashtray gets full, we'd still have to take the time to install and troubleshoot the new hardware and to reconfigure all the software. So, upgrading is not necessarily the best choice, either to resolve a problem or to improve performance. Spending a little time fine-tuning your hardware and software and perhaps spending a little money replacing certain components can make a difference.

Naturally, there is a certain point past which your computer is going to turn into a money and time pit. The older your system is, the less vigorously you should try to keep it alive. It's easy to calculate the point of diminishing returns: just compare the estimated cost of an upgrade (both the monetary cost and the amount of time you'll have to commit) with the cost of a new system (minus what you might get for selling or donating your old system). I stress this point a great deal, because I've seen it happen time and time again: people end up spending too much and getting too little in return. A simple hardware upgrade ends up taking days of troubleshooting and configuring, only to result in the discovery that yet something else needs to be replaced. Taking into account that whatever you end up with will still eventually need to be further upgraded to remain current, it is often more cost effective to replace the entire system and either sell or donate the old parts.