Know your CPUs and what they do

BY DAVE THOMPSON
Last updated 09:03 03/08/2010

Relevant offers

Gadgets

Review: Acer Aspire S3 Review: Sony Ericsson Xperia arc S 3D printing: saviour or piracy tool? Apple in talks on iTV Get ready for the Apple ... Treadmill? What's USB 3.0? Review: Dell XPS 14z notebook Robots bring beer but Rosie still takes cake Duck for cover, here comes the laser-guided bullet Angelic 'Jobs' loves Android in Taiwan ad

Today, I'm writing about CPUs. Wait, before you self-harm, the world of the CPU is not as boring as you might think. In fact, it's worse.

But if you are buying a new computer and are completely confused by the salesperson's slick sales pitch and the befuddling array of colourful stickers adorning machines, sporting terms like "Core Duo" and "AMD Athlon II", then pay attention.

To gain even a basic understanding of any of this, we have to go back to the dark ages of computing, the 1970s. At a Silicon Valley company called Intel, technicians squashed as many transistors as they could on to a wafer of silicon, creating the first microprocessor, which led shortly thereafter to the first central processing unit (CPU). The age of the microcomputer had begun.

To understand yet more, you have to know how the computer controls the CPU. Uber- simplistically (note to anoraks considering emailing me about this: please don't), the CPU is managed by the motherboard, which runs it at a certain clock frequency, measured in gigahertz.

Generally, the faster the clock frequency, the more instructions per second the CPU can do - up to a point. Speed also very much depends on the architecture of the processor, with technologies like hyper and multi-threading and pipelining making a vast difference to overall number crunching.

As technology advances, the size of transistors, interconnecting wiring and other atomic-level architectures gets smaller, meaning faster and more efficient components.

Moore's Law states something about the number of transistors on a given-sized blob of silicon doubling every six months, but since there is a limit to how small these things can be made, instead of cramming in more transistors, boffins theorised that two CPUs are better than one.

By linking two together, they created the dual-core processor. However, this doesn't necessarily mean double the calculations; it's a bit like using two hands to do a one-handed job. It is just more efficient.

There are also issues introduced with a dual-core setup that affect overall speed, though it stands to reason a dual- core CPU will outperform a single-core unit, even if the latter is running at higher clock frequencies. This is why we don't simply buy the chip with the fastest claimed clock speed.

These days we also have quad- core processors as well as laptop CPUs designed for low power use and heat output, which creates a bewildering array of options, well beyond the scope of this article.

Ad Feedback

Throw in the fact that AMD and Intel use different measurement conventions to rate their products and it is no wonder punters are stumped.

What CPU you choose has a lot to do with what you want it for; 3D animators and high-end gamers have vastly different CPU requirements than someone just wanting to surf the web, answer emails and write documents. Fortunately, most modern machines are relatively fast, once the crapware is removed.

My advice is to research very carefully to ensure you get the right tool for the job, or do what most of us guys do - buy the biggest, baddest and most expensive CPU you can find.

* Dave Thompson runs a computer-services company in Christchurch. Contact: dave@computerkungfu.com.

- © Fairfax NZ News

Special offers
Opinion poll

At what age is it OK for children to have a smartphone?

Four

Seven

10

13

16

Vote Result

Related story: (See story)

Featured Promotions

Sponsored Content