The demo showed animation running 30 frames per second at 1,920 x 1,080. Nvidia cranked the demo up to 2,560 x 1,600 but would not reveal the frame rate. This could have huge implications for in-game graphics, although as the system currently requires 4 parallel Quadro GPUs with 1GB memory apiece, costing around $US 10,000 a pop it may be a couple of years before this hits even the most hardcore PC gamer’s desktop.

Quoth Nvidia, “the ray tracer shows linear scaling rendering of a highly complex, two-million polygon, anti-aliased automotive styling application”. Which certainly sounds impressive. What this appears to mean is “Look! A shiny car that we can move around real quick!” and, you know, that may well be enough.

Nvidia demos real-time GPU ray tracing at 1,920 x 1,080 [CustomPC]

Now it looks like NVIDIA’s plans for AGEIA’s technology aren’t just games-related.

You can head on over to FiringSquad for the entire interview, however, here’s the meatiest part of the short talk:

Second, the computer industry is moving towards a heterogeneous computing model, combining a flexible CPU and a massively parallel processor like the GPU to perform computationally intensive applications like real-time computer graphics. Physics is a natural for processing on the GPU because, like graphics, it is made up of thousands of parallel computations, and with our CUDA technology, which is rapidly becoming one of the most pervasive parallel computing programming environments in history, we can open this exciting parallel processing world to applications desperate for a giant step in computing performance—such as physics processing, computer vision, video/image processing, and a world of exciting applications we’ve not yet imagined.

What I’m getting from this is that NVIDIA isn’t so much interested in AGEIA for hardware-accelerated physics in games, but more what the massively parallel design of the hardware can be used for in certain general processing tasks. This is something AGEIA’s PPU and the GPU have in common.

Is NVIDIA looking to take on Intel and AMD? There’s definitely a place for this sort of hardware in specialised markets.

NVIDIA AGEIA PhysX Acquisition Interview [FiringSquad]

The patch, aimed at PCs running Windows Vista only, addresses a scheduling problem within the DirectX kernel. Here’s the technobabble, which I’ll attempt to decipher:

A graphics processing unit (GPU) or part of a GPU link may incur a small scheduling latency on a new Direct Memory Access (DMA) packet. This is true if the DMA packet is submitted to the GPU that has previously run out of work and if another GPU in the linked configuration is very busy. In some scenarios, significant starvation of one or more GPUs in the linked configuration reduces the expected performance of an application.

Essentially, it’s a load-balancing thing – the hotfix spreads the work evenly among the GPUs. It’s unclear whether this affects both Direct3D 10 and Direct3D 9 games – Microsoft hasn’t provided enough information to make a solid call.

The link below is for the Knowledge Base article. If you want the hotfix, I’d suggest trying elsewhere (every couple of hours) or waiting for Service Pack 1.

Graphics performance can be improved in certain multiple-GPU scenarios on a Windows Vista-based computer [Microsoft via GGMania]