Pushing the boundaries in graphics hardware
01 December 2007
The folks at RARE, one of the iconic pioneers in computer gaming, like to joke that the quiet surrounds of their headquarters, deep in the UK countryside in Warwickshire, is the reason for their company's consistently strong record in innovation. “From the early days it was intentional to be outside the city,” says Nick Burton, a Senior Software Engineer at RARE. “Being somewhere quiet lets you concentrate on making games, and not on where to go at lunchtime!”
RARE's track record is also built on teamwork and putting the gaming experience first. “We don't do technology for technology's sake,” Nick says. “The technology has to facilitate what the concept artists and designers want to do. The best term is ‘cyclical': the designers ask for something, the programmer usually sucks his teeth and goes away to think about it, and then delivers something close to what the designers want.” Since 2002, RARE has been part of Microsoft Games Studios where they are the biggest studio by headcount. “We've always had close links with UK universities and now we have access to Microsoft Research and its network as well,” Nick points out.
Working on core technology as well as broader demographic games such as Viva Piñata means that RARE is constantly pushing the boundaries. “We're entering a very exciting phase of computer programming right now,” says Mike Boulton, also a Senior Software Engineer at RARE. “In graphics we're seeing how a pragmatic approach to fidelity improvement is actually driving us to generate new algorithms and new ways of doing things. So you don't feel limited by the hardware as much as by your ability to come up with novel approaches. It's an open problem but we're chipping away at it.”
“The focus is shifting to clever means of extracting greater fidelity without resorting to the brute force method of manually increasing asset complexity,” Mike continues. “Exploiting various types of coherency will also be important. When you look into the future, you are going to see machines with a bit more RAM, but a lot more processing power. There might also be a generalisation between the CPU and the graphics processing unit – that is going to be a very interesting time. With the Xbox360, we are just at this threshold with a number of threads on the CPU and a powerful GPU.”
What this means is that as users walk through a game, they will find the fidelity of the light and shadow around them changing dynamically, or that something really does jump up at them. At the moment it is impossible to calculate these for every frame as it would just take too long. “We're moving from the situation of calculating something every frame to a continual ‘moving point average' calculation where assets, such as a background, will be stored in a more raw and verbose format. The machine will be constantly deciding and self-analysing what to build and re-build,” Mike explains. “It is all about focusing the power of the system on what you see and what you appreciate. The surface exposed to the player needs to have as high a fidelity as possible and there will be a much more sophisticated algorithmic approach to this challenge.”
Nick Burton (left) with Mike Boulton, both Senior Software Engineers at RARE.
Nick Burton says the other big development is that computers will decide where to spend their power – which will prove liberating for all developers. “It feels like we entered a renaissance with the Xbox360, allowing us to do what we wanted - so as the hardware challenges evaporate even further we can increasingly focus on creating the best assets and the best games.”
Another challenge that RARE has embraced is broadening the appeal of games to new audiences, particularly those who are intimidated by status quo gamers. “We want to push the boundaries technically but also demographically,” Nick says. “Viva Piñata appeals very broadly, but the hardcore audience still want to pick it up because it is a deeper experience: it looks good and it pushes the hardware.” As Mike says, “The ideal situation is to have a game that appeals on multiple levels. Experience shows that different gamers have different levels of expectation – so generating games that appeal to all these disparate groups is probably the biggest challenge of all.”