NVIDIA's RTX 500 and 1000 Ada GPUs bring more AI smarts to thin and light workstations
Just ahead of Mobile World Congress, NVIDIA unveiled its latest laptop GPUs and, what a surprise, they're designed largely to assist with AI processing. The RTX 500 and 1000 Ada Generation graphics cards are primarily for thin and light laptops. While they won't offer as much TOPS AI performance as current higher-end mobile GPUs, they could be a handy option for on-the-go AI processing for the likes of researchers, content creators and video editors. It's worth noting they're workstation GPUs rather than ones designed for gaming.
NVIDIA says the GPUs, which are based on the Ada Lovelace architecture, offer up to twice the ray-tracing performance of previous-gen GPUs (they employ third-gen ray-tracing cores). Fourth-gen Tensor Cores, meanwhile, deliver up to twice the throughput of previous GPUs, according to NVIDIA. The company says this helps with accelerating deep learning training, inferencing and AI-based creative workloads."
The RTX 500 has 4GB of dedicated memory, while the RTX 1000 has 6GB. NVIDIA says they deliver up to 154 and 193 TOPS of AI performance, respectively. Compared with a CPU-only AI configuration, the RTX 500 is slated to provide up to three times faster AI-powered photo editing, as much as 10 times the graphics performance for 3D rendering and up to 14 times the generative AI performance for various models.
The GPUs also support DLSS 3, the company's upscaling tech. In addition, an eighth-gen encoder includes AV1 support. NVIDIA says this video codec is up to 40 percent more efficient than H.264, enabling new possibilities for broadcasting, streaming and video calling."
If you're interested in picking up a laptop with an RTX 500 or 1000 GPU, you won't have to wait long. They'll debut this spring in laptops from the likes of Dell, HP, Lenovo and MSI.
This article originally appeared on Engadget at https://www.engadget.com/nvidias-rtx-500-and-1000-ada-gpus-bring-more-ai-smarts-to-thin-and-light-workstations-161517977.html?src=rss