Nvidia unveils Iray VCA, an 8 GPU device for ray trace rendering

1547 0

Ray trace rendering in near real time: 19 Nvidia Iray VCA devices used together to slice through a complex Honda engineering model

Nvidia has unveiled a dedicated ray-trace rendering box designed specifically to accelerate Nvidia Iray-enabled applications, including DS Catia, Bunkspeed Drive and 3ds Max.

The Nvidia Iray Visual Computing Appliance (VCA) combines eight top-end Nvidia GPUs, CPUs and software with a view to massively accelerating the photorealistic rendering tool.

Each Iray VCA plugs into the network and can also be linked together to produce a whopping big render farm for ray tracing in near real time.

At Nvidia’s GTC event this was demonstrated by linking together 19 VCAs and slicing through a massive photorealistic engineering model of a Honda car. The results were hugely impressive, with new high definition views refining almost instantly. However, with each Iray VCA costing $50,000, a set up like this would cost close to $1 million.


The $50,000 Nvidia Iray VCA features 8 high-end Nvidia GPUs, each with 12GB of memory

Iray render farms are nothing new. At GTC 2013 Nike’s Russell McMurdie shared details of the company’s 24 Tesla C2075 GPU rack, which it uses to accelerate Bunkspeed.

But with Iray VCA it seems Nvidia is going for simplicity and scalability featuring all the hardware and software you need, and claiming little or no technical support is required. Software includes the Iray VCA Cluster Manager, which dynamically allocates Iray VCAs to meet the demands of the day’s workload.

With 8 powerful GPUs in a single box Nvidia is targeting users of Iray-enabled applications who may use one or two Nvidia GPUs inside their workstation but are looking for a shared network resource to deliver a more responsive rendering workflow.

The Iray VCA is Nvidia’s second Virtual Computing Appliance. At GTC 2013 it unveiled the GRID VCA, designed to deliver remote workstations out of a box for small engineering firms.

Full specs of the Nvidia Iray VCA
— GPU: 8 NVIDIA Top End Nvidia GPUs
— GPU Memory: 12GB per GPU
— CUDA Cores: 23,040
— CPU Cores: 20
— System Memory: 256 GB
— Storage: 2TB SSD
— Network: 2 x 1GigE, 2 x 10GigE (SFP+), 1 x InfiniBand
— Installed Software: Linux Cent OS, Iray, Iray VCA Cluster Manager

Honda car visualised on 19 Nvidia Iray VCAs. Seconds later this incredibly complex engineering model was sectioned in real time.

Honda R&D
Honda Research and Development’s styling design department is an initial user of the Nvidia Iray VCA, with a prototype cluster made up of 25 nodes to refine styling designs on future cars.

“For our styling design requirements, we developed specialised tools that run alongside our RTT global standard platform,” said Daisuke Ide, at Honda Research and Development. “Our TOPS tool, which uses Nvidia Iray on our Nvidia GPU cluster, enables us to evaluate our original design data as if it were real. This allows us to explore more designs so we can create better designs faster and more affordably.”

Leave a comment