The 2008 Guide to Gaming PC SpecsLooking AheadPredicting the future is always difficult, but looking at past trends it's possible to make good, educated guesses. We are currently in the middle of a console cycle - all manufacturers have an up-to-date product out on the market, we have games for them that take advantage of most of the resources available, and there is no new console hardware on the horizon for 2008. Realistically it will be late 2009 or early 2010 before you can spend money on the next WiiPSXBox720 or whatever they decide to call their next piece of kit, and that's so far into the future that anything past that point is a wild guess - yours is as good as mine. The Xbox 360 is no PC, but its capabilities dictate what gets done these days as the development budgets keep growing. Consoles are obviously not PCs, but they directly affect PC game development. More and more titles are developed on multiple platforms, and the capabilities of Xbox 360 and PS3 dictates what gets done. While there are some PC-only developers who couldn't care less about consoles, the development budgets these days are so large that it just makes sense to do the game for all possible platforms simultaneously, especially as the engines and middleware used often allow you to do it by simply recompiling the existing console title to the PC (or the other way around). This can lead to ugly PC ports of console games, frequently with no effort being made to "localize" the game for the PC player. During 2007 some of the publishers did put more effort towards offering proper PC versions, and the "crappy console port" became the unfortunate exception instead of the norm. So, based on this, the effective minimum gaming system for the foreseeable future (next two years) should be a system that can run these multiplatform games. We might see some very forward-looking PC-only games that push beyond what's possible on the current consoles, but they will be rare exceptions. DX10 is the (far) futureDX10 adds its own twist to the equation - my personal educated guess is that we are still a year (or more) from the first DX10-only game, and at least two years away from a situation where DX10 hardware is commonly required. Add to the mix the fact that none of the current cards have managed to demonstrate good performance while running DX10 material on Windows Vista, and I'm ready to say that while it's smart to pick DX10-level hardware today, that is true only because it offers the best DX9 performance available. The DX10 capabilities are just a nice bonus. It is still a new technology, and just as we saw with the first DX9 cards years ago, the first generation hardware for a whole new architecture tends to underperform with the new code, and is aimed mostly at the developers working on the new stuff. NVIDIA and ATI would want you to think otherwise, but that's a given - they want to sell you new video cards. DX10 performance just isn't a factor when picking a system today - and by the time it is, you can probably obtain a faster replacement video card at a fraction of the price asked for today's top card (NVIDIA's GeForce 8800 Ultra). I'd love to see the PC market advance faster, but my pessimistic view is that the consoles will hold things back until we start hearing rumbles about the next console generation - probably in 2009. Current "reasonably priced high end cards" - such as ATI's Radeon HD 3870 and NVIDIA's GeForce 8800 GT - will very likely be valid gaming cards two years from now, even as DX10-exclusive titles start to crop up. As a comparison, I built my current home system back in the spring of 2006 with a brand new Radeon X1900 XT inside. Two years later, and it's still good enough for almost any game out there, and will continue to be good enough well into 2008 - sure, it doesn't run Crysis maxed, but come on, it's two years old. The fact that the GeForce 8800 GTX, released over 12 months ago, is still almost the king of the hill on the market reinforces this argument. The video card development cycles have slowed down; sure, the vendors still put out something "new" every six months - they have to, to stay in business - but the improvements are evolutionary rather than revolutionary, bringing down the price, the manufacturing costs and the cooling requirements. Part of the reason for this can be traced to the competitive situation - ATI has kept stumbling in the past year, and there has been no need for NVIDIA to aggressively push for a new high end generation until ATI has something competitive out there. In fact, the next step seems to be to glue together two cards, and call them "X2" - ATI have their Radeon HD 3870X2 and NVIDIA, the GeForce 9800 GX2. Out of these two, ATI's card is labelled honestly - it is effectively two HD 3870s planted onto a single circuit board, while NVIDIA tries to market two 8800GT boards glued together as The Next Generation with the 9800 label. It remains to be seen how long we have to wait for true second generation high end DX10 hardware instead of the current crop of die shrinks. NVIDIA might do an announcement before the summer, but I wouldn't be surprised if it took another six to nine months to see a true successor to GeForce 8800 and Radeon HD 2900/HD 3870 that you can actually buy from a store.
Related StuffTags |
|||||||||||||||||||||||||||||
![]() |