Jump to Content
Telecommunications

Immersive Stream for XR: Performance and savings with NVIDIA L4 GPU and autoscaling

February 15, 2024
Qingsi Wang

Software Engineer, Immersive Stream for XR

Try Gemini 1.5 Pro

Google's most advanced multimodal model in Vertex AI

Try it

NVIDIA L4 GPUs and Google Cloud deliver the next level of cloud GPU performance to handle the rendering of your most demanding projects, along with capacity autoscaling that automatically scales resources to match traffic surges and daily peaks and troughs. This ensures that you have only the capacity you need at all times. Today, we are excited to announce improvements in both performance and cost savings to Immersive Stream for XR, Google Cloud’s solution for hosting, rendering, and streaming 3D and eXtended reality (XR) experiences, which will help you deliver more immersive and cost-effective experiences via NVIDIA L4 Tensor Core GPU support and capacity autoscaling. 

The NVIDIA L4 GPU is now generally available as a GPU class on Immersive Stream for XR for the 3D-only mode. NVIDIA L4 GPUs offer twice the performance compared with the existing NVIDIA T4 Tensor Core GPUs. The NVIDIA L4 GPU class on Immersive Stream for XR redefines the price-performance ratio for immersive experience providers. 

“By once again collaborating with Google Cloud for its Immersive Stream for XR, now powered by NVIDIA L4 GPUs, we’re able to offer top performance at a lower cost to power the next-generation of immersive experiences.” - Anne Hecht, senior director of enterprise software products, NVIDIA

Further, a new capacity autoscaling feature dynamically matches capacity to real-time traffic needs. You only need to configure a range of regional capacities with a target buffer of standby available capacity. This helps ensure sufficient server capacity without the overhead of manual capacity planning, allowing you to focus on content creation, not infrastructure, and only pay for the resources you need. This new autoscaling feature is available for all service instances, using either the NVIDIA T4 or NVIDIA L4 GPUs. 

The above new capabilities are available today. To get started, you can follow the quickstart guide to get started, select the NVIDIA L4 GPU class and 3D-only experience mode. Generally speaking, 3D-only is a cost-effective mode, priced significantly less than AR mode while still supporting augmented reality experiences. Please see this comparison between 3D and AR modes. You can also learn more about how to enable and configure autoscaling for your service instances in this guide to lower costs for your immersive experiences, immediately.

Posted in