Your most demanding projects can be rendered with the next level of cloud GPU performance from NVIDIA L4 GPUs and Google Cloud, which also offers capacity autoscaling, which automatically adjusts resources to match traffic surges and daily peaks and troughs. This guarantees that you always have the capacity you require. Google is pleased to announce today that Immersive Stream for XR, Google Cloud’s solution for hosting, rendering, and streaming 3D and eXtended reality (XR) experiences, has improved in both performance and cost. With support for NVIDIA L4 Tensor Core GPUs and capacity autoscaling, you will be able to provide more immersive and affordable experiences.
NVIDIA L4 Tensor Core GPU
The ground-breaking universal accelerator for effective graphics, AI, and video.
As of right now, Immersive Stream for XR supports the NVIDIA L4 GPU as a GPU class for 3D-only usage. When compared to the current NVIDIA T4 Tensor Core GPUs, the performance of NVIDIA L4 GPUs is doubled. For immersive experience providers, the price-performance ratio is redefined with the release of the NVIDIA L4 GPU class on Immersive Stream for XR.
“To power the next generation of immersive experiences, Google can offer top performance at a lower cost by partnering with Google Cloud once again for its Immersive Stream for XR, now powered by NVIDIA L4 GPUs.” Senior Director of Enterprise Software Products at NVIDIA, Anne Hecht
Additionally, a brand-new capacity autoscaling feature dynamically adjusts capacity to traffic demands in real-time. All that is required is the configuration of a target buffer of standby available capacity with a range of regional capacities. By eliminating the need for labor-intensive manual capacity planning and ensuring adequate server capacity, this lets you concentrate on creating content rather than managing infrastructure and only pay for the resources you use. All service instances can use the new autoscaling feature, which works with either the NVIDIA T4 or NVIDIA L4 GPUs.
The aforementioned new features are currently accessible. Use the quickstart guide and choose theNVIDIA L4 Tensor Core GPU class and 3D-only experience mode to get started. In general, the 3D-only mode is more affordable than the AR mode, even though it still allows for augmented reality experiences. View this comparison of the AR and 3D modes. To immediately reduce the cost of your immersive experiences, you can also find out more in this guide about how to enable and configure autoscaling for your service instances. Boost Workloads in AI, Video, and Graphics
Acceleration for visual computing, virtualization, AI, video, and graphics is provided universally and energy-efficiently by the NVIDIA L4 Tensor Core GPU, which is powered by the NVIDIA Ada Lovelace architecture. L4, which comes in a low-profile form factor, is an affordable and energy-efficient solution that offers high throughput and low latency in all servers, be it cloud, data center, or edge.
Experience AI Video Pipeline Performance in Real Time
Utilize NVIDIA L4’s power to transform video applications. Servers with NVIDIA L4 Tensor Core GPU support up to 1,040 concurrent AV1 video streams at 720p30 for mobile users, whether delivering immersive augmented and virtual reality (AR/VR) experiences, allowing users to create creative stories, or streaming live to millions of viewers.
When combined with the CV-CUDA library, NVIDIA L4 GPUs offer a 1.5X increase in GPU memory and fourth-generation Tensor Cores, significantly improving the understanding of video content. Compared to CPU-based solutions, L4 offers 120X better AI video performance. This gives businesses real-time insights into how to personalize content, enhance search relevance, identify objectionable content, and apply smart-space solutions.
Use L4 to Use Less Energy and Space
There is a greater need than ever for effective, affordable computing as AI and video become more commonplace. Compared to conventional CPU-based infrastructure, NVIDIA L4 Tensor Core GPUs offer up to 120X better AI video performance, up to 99 percent better energy efficiency, and a lower total cost of ownership. This enables businesses to scale their data centers to many more users while reducing rack space and drastically reducing their carbon footprint. A 2 megawatt (MW) data center’s energy savings from moving from CPUs to NVIDIA L4s can power almost 2,000 homes for a year or equal the carbon offset from growing 172,000 trees over ten years.
Boost the Performance of Generative AI Quickly
In all industries, generative AI for text and images improves customer experiences by making lives easier and more immersive. Up to 2.5 times faster than the previous generation of GPUs, NVIDIA L4 turbocharges compute-intensive generative AI inference. Additionally, the L4 GPU generation can create larger images up to 1024×768, something that was not possible with the previous GPU generation due to its 50% greater memory capacity.
Enhance Graphics Performance
For AI-based avatars, NVIDIA Omniverse virtual worlds, cloud gaming, and virtual workstations, NVIDIA L4 offers over 4X better performance thanks to its third-generation RT Cores and AI-powered NVIDIA Deep Learning Super Sampling 3 (DLSS 3). These features let artists create immersive visual experiences that aren’t achievable with CPUs by allowing them to create real-time, cinematic-quality graphics and scenes.
Increase Tasks in an Effective and Long-Term Way
An essential component of the NVIDIA data center platform is NVIDIA L4. The platform accelerates over 3,000 applications and is available everywhere at scale, from data center to edge to cloud, offering both dramatic performance gains and energy-efficiency opportunities. It is designed for video, AI, graphics, simulation, data science, and data analytics.
L4 is a low-profile form factor that operates in a 72W low-power envelope. It is optimized for mainstream deployments and is a cost-effective and efficient solution for any server or cloud instance from NVIDIA’s partner ecosystem.
Optimize Development and Implementation The NVIDIA AI Enterprise software suite includes AI solution workflows, frameworks, pre-trained models, and infrastructure optimization that are certified to run on popular data center platforms and mainstream NVIDIA-certified systems with NVIDIA L4 GPUs. This enterprise-ready software is designed to simplify AI development and deployment.
With the best performance in training, inference, and data science, NVIDIA AI Enterprise is a license addition for NVIDIA L4 GPUs that makes AI available to almost any kind of enterprise. When used in conjunction with NVIDIA L4, NVIDIA AI Enterprise makes it easier to construct a platform that is ready for AI, speeds up the development and implementation of AI, and provides performance, security, and scalability to enable quicker insight gathering and the sooner realization of business value.
