NVIDIA RTX 6000 Workstations for AI and LLM Development

computers feature NVIDIA Omniverse , NVIDIA RTX 6000 Ada GPUs, and NVIDIA AI Enterprise Software

Powerful new NVIDIA RTX 6000 workstations created for development and content production in the age of generative AI and digitization were today unveiled by NVIDIA and international manufacturers.

The systems, which come from manufacturers including BOXX, Dell Technologies, HP, and Lenovo, are built with NVIDIA RTX 6000 Ada Generation GPUs and include NVIDIA AI Enterprise and NVIDIA OmniverseTM Enterprise software.

Separately, NVIDIA unveiled three new desktop workstation Ada Generation GPUs – the NVIDIA RTX 5000, RTX 4500, and RTX 4000 — to give professionals around the world access to the most recent AI, graphics, and real-time rendering technology.

According to Bob Pette, vice president of professional visualization at NVIDIA, “few workloads are as challenging as generative AI and digitalization applications, which require a full-stack approach to computing.” With the most recent RTX workstations powered by NVIDIA, professionals may now take on these on a desktop, giving them the ability to create massive, digitalized worlds in the generative AI era.

Up to four NVIDIA RTX 6000 Ada GPUs with 48GB of memory each are available in the new RTX workstations, and a single desktop workstation may deliver up to 5,828 TFLOPS of AI capability and 192GB of GPU memory. Systems can be configured with NVIDIA AI Enterprise or Omniverse Enterprise, depending on the needs of the customer, to power a variety of demanding generative AI and graphics-intensive tasks.

NVIDIA NeMoTM, an end-to-end framework for developing and customizing foundation models for generative AI, NVIDIA RAPIDSTM libraries for data science, as well as frameworks, pretrained models, and tools for developing typical enterprise AI use cases, such as recommenders, virtual assistants, and cybersecurity solutions, are now included in NVIDIA AI Enterprise 4.0, which was separately announced today.

A platform for industrial digitalization called Omniverse Enterprise lets teams to create OpenUSD applications and interoperable 3D workflows. Omniverse, an OpenUSD-native platform, allowing geographically dispersed teams to work together on datasets with full design fidelity from hundreds of 3D apps.

“Yurts offers a full-stack generative AI solution that fits our customers’ various form factors, deployment methodologies, and financial constraints. This has been accomplished using the RTX 6000 Ada and LLMs for various natural language processing applications. According to Jason Schnitzer, chief technology officer at Yurts, “from private data centers to workstation-sized solutions that fit under a desk, Yurts remains committed to extending our platform and offering with NVIDIA.

Workstation users can also benefit from the new NVIDIA AI Workbench, which will soon be made available in early access and gives developers a unified, user-friendly toolbox for quickly building, optimizing, and running generative AI models. Pretrained generative AI models can be swiftly created, tested, and customized by users of any skill level on a PC or workstation before being scaled to almost any data center, public cloud, or NVIDIA DGX Cloud.

RTX technology of the future

The newest NVIDIA Ada Lovelace architecture technologies are included in the new RTX 5000, RTX 4500, and RTX 4000 desktop GPUs, including:

Up to two times as much single-precision floating point throughput with NVIDIA CUDA® cores as compared to the previous generation.

Third-generation RT Cores: Able to conduct ray tracing with either shading or denoising capabilities while providing up to 2x the throughput of the previous generation.

Fourth-generation Tensor Cores: Enhanced support for the FP8 data format and up to 2x quicker AI training performance than the prior generation.

DLSS 3: Using the power of AI, real-time graphics now have new degrees of realism and interactivity.

More memory for the GPU: The RTX 4000 delivers 20GB of GDDR6 memory, the RTX 4500 gives 24GB of GDDR6 memory, and the RTX 5000 boasts 32GB of GDDR6 memory, all of which enable error-code correction for error-free processing with big 3D models, rendered images, simulations, and AI datasets.

Extended reality capabilities: Support for high-resolution augmented reality and virtual reality devices to deliver the high-performance graphics necessary for producing beautiful augmented reality, virtual reality, and mixed reality content.

“The NVIDIA RTX 5000 Ada GPU demonstrates NVIDIA’s impressive generational performance improvements—it has significantly increased our efficiency in creating stereo panoramas using Enscape,” stated Dan Stine, director of design technology at architectural firm Lake|Flato. Our massive, complicated models look excellent in virtual reality thanks to the performance boost and large frame buffer of RTX 5000 GPUs, providing a more cozy and contextual experience for our clients.

Availability

Starting in the fall, system builders will provide RTX workstations with up to four RTX 6000 Ada GPUs, NVIDIA AI Enterprise, and NVIDIA Omniverse Enterprise.

Starting today, HP and various international distribution partners, including Leadtek, PNY, and Ryoyo Electro, will begin selling the new NVIDIA RTX 5000 GPU. Fall availability for the NVIDIA RTX 4500 and RTX 4000 GPUs is expected through BOXX, Dell Technologies, HP, Lenovo, and other international distribution partners.