What’s the point?

Today’s ultra high-end components command outlandish price tags and, in some cases, offer basic results. So, is there a point to their existence? Our WinLabs Editor certainly thinks so…

  • E-Mail
By  Jason Saundalkar Published  April 1, 2008

Most markets are comprised of three different segments: low-, medium- and high-end. The IT industry however is a little more diverse - in that the market doesn’t top out at high-end but rather at the ‘ultra high-end’ as we’ve come to call it. (Many companies refer to this as the enthusiast performance segment.)

A number of products and technologies fall into this category, including Intel’s Extreme Edition quad-core processors and Skulltrail platform, nVidia’s quad-SLI and AMD’s quad-CrossFire X technology, Phase-Change cooling solutions and  many more. Each of these is aimed at the the ultra high-end category and demands a steep asking price.

Take multi-GPU technology as an example. To do this dance, you need a multi-GPU compatible motherboard, and however many graphics cards you want to combine. Let’s say you’re combining four high-end cards on a compatible motherboard, you’re looking at an investment of well over US $1500 for just these components. Plus, you’ll still need to factor in the cost of a processor, memory and more; quite an expensive proposition then. But here’s the real kicker, regardless of whether you’re running a CrossFire X or SLI powered machine, your game’s framerate will never go up by a factor of four when you go from one card to four. I’ve observed a 60% improvement over using a single card in most cases but this again depends on the resolution the software is being run at as well as the rest of your components and how optimized (or not) the software itself is. I’m not pointing the finger at multi-GPU technology by itself either; almost everything in this pricey segment will rarely produce results that are worth shouting about.

Besides the code and the actual kit, you could point the guilty finger at the operating system, drivers and almost everything else inside the rig. The fact is that it’s very hard to get ultra high-end kit to perform and return linear results because of the various factors mentioned above. With that said, why should you and I care?

In the long run, having kit of this class on the market ultimately helps us - the consumers. You see, whenever something ultra high-end is released, it automatically pushes down the pricing of low-, medium- and high-end kit. Going further still, when a brand new quad-core ultra high-end CPU is unleashed for example, the previous super chip may become a high-end part with a drastically reduced price. This could also knock the previous high-end chip into the mid-range category and so on.

Another benefit is that ultra high-end hardware pushes software developers to better optimise their code. I for one am sure that AMD, Intel and nVidia for example, are far from happy that their respective platforms and technologies are being restricted in the performance department because of un-optimised coding. The good news here is that while those running this ultra high-end hardware will eventually see bigger performance boosts as developers beef up their code, those running kit from any of the other three segments can also look forward to better performance.

Super-fast kit also spurs competition and innovation from vendors and this is always a good thing. When nVidia launched its multi-GPU SLI technology for instance, AMD (ATI at the time), was also prompted to release its own multi-GPU technology. If it didn’t, the firm would have lost a lot of business to nVidia. Now that both AMD and nVidia have two-, three- and four-way multi-GPU technology, users have a choice of which technology and company to go with, and choice is critical when building your own rig.

In the end, although ultra high-end kit is very expensive at launch and provides basic benefits, at least at present, we ultimately benefit in the long run. We just need to play a bit of a waiting game, and that’s just fine with me.

4009 days ago
Tom L

"Another benefit is that ultra high-end hardware pushes software developers to better optimise their code." I have to disagree with you on this - I believe the dramatic increases in the performance of hardware components has made it very easy for software developers to be lazy. If you are a developer faced with either going through and optimising your code - a very tedious and time-consuming process that could potentially delay release - or rely on the fact that ever-increasing hardware specs will make your sloppy programming less relevant, which are you going to choose? Hardware development is great, but I believe that if the pace of new releases of GPUs, for example, was to slow, this might force developers to make better use of what they already have.

Add a Comment

Your display name This field is mandatory

Your e-mail address This field is mandatory (Your e-mail address won't be published)

Security code