Many hardware enthusiasts have been waiting for AMD’s response to Nvidia’s GeForce 8800 family of graphics processors before deciding which company would get their hard earned cash. However, that wait has been a long one – well over six months in fact – and a lot of enthusiasts got fed up and bought a GeForce 8800-series card.
On Monday, AMD launched its counter-attack on Nvidia with its eagerly anticipated R600 graphics processing unit, but in many ways it’s a strange one – we’ll come back to why this is the case shortly though.
An (almost) United Front:
If you can remember back to when we interviewed Nvidia’s David Kirk, he skirted around on his company’s direction with unified shader architectures. During the interview, he down-played unified shaders, stating that “another word for ‘unified’ is ‘shared’, and another word for ‘shared’ is ‘competing’.”
Kirk finished off the interview with some carefully selected words that many people didn’t seem to pick up on – essentially rubbishing unified shader architectures was juicy enough for most hardware enthusiasts.
“We will do a unified architecture in hardware when it makes sense. When it's possible to make the hardware work faster unified, then of course we will. It will be easier to build in the future, but for the meantime, there's plenty of mileage left in this architecture.” The time was obviously right with the GeForce 8-series and we certainly weren’t expecting G80 to be unified.
On the other hand, ATI's situation was much easier to predict. The company's direction became pretty clear when it unveiled its Xenos chip, which is the GPU behind the awesome graphics we're witnessing on Microsoft’s Xbox 360 console. When we spoke to ATI's Richard Huddy, he made it pretty clear that unified shader architectures were the way forward for desktop graphics.
By that rationale, Microsoft's stance was also pretty clear, as Richard pointed out. “Talking to the guys at Microsoft, it’s impossible to escape the conclusion that the future is for unified pipelines, there’s no doubt.”
For various reasons though, this has taken a couple of generations to arrive. From ATI's perspective, its R520 GPU was already nearing completion (after some heavy delays) by the time the Xenos design was finished and delivered to Microsoft. On the software front, DirectX 9.0 and previous versions of the API really weren’t tailored to unified shaders. And thus, a unified graphics pipeline on the PC couldn't happen until after Microsoft chose to made a big step forwards when it started DirectX 10 from a clean slate.
So, if we fast forward to today, we have a unified front – Microsoft, ATI and Nvidia are all pulling in the same direction when it comes to the future of gaming on the PC. This was also A Good Thing for programmers too, as it would allow them to express themselves more freely in their shader programmes – the shackles were removed just as quickly as you can say HDR-plus-AA.
Unified Shader Architecture – no more idle silicon!
bit-tech's R600 Coverage:
Because there is a lot of technology to cover in AMD's R600 architecture, here's a handy break-down of the bit-tech coverage:
Before we delve into ATI’s new architecture, there are a number of things we’re taking for granted. Thus, if you’re not familiar with ATI's Radeon X1900-series architecture, Nvidia’s G80 graphics processing unit and what DirectX 10 brings to the table, we recommend you give the articles listed below a read: