New NVIDIA DGX System Software and Infrastructure Solutions Supercharge Enterprise AI
At GTC today, NVIDIA unveiled a amount of updates to its DGX portfolio to ability new breakthroughs in organization AI improvement.
NVIDIA DGX H100 systems are now offered for get. These infrastructure creating blocks assistance NVIDIA’s total-stack enterprise AI solutions.
With 32 petaflops of effectiveness at FP8 precision, NVIDIA DGX H100 provides a leap in performance for organization AI progress. It provides 3x reduce complete cost of ownership and 3.5x far more energy efficiency in contrast to the former era.
New NVIDIA Foundation Command computer software, which simplifies and speeds AI advancement, powers each DGX system — from single nodes to DGX SuperPODs.
Also unveiled was NVIDIA DGX BasePOD — the evolution of DGX POD — which can make enterprise facts-center AI deployments simpler and more quickly for IT groups to obtain, deploy and manage.
Many of the world’s AI leaders are developing technological breakthroughs — from self-driving cars to voice assistants — utilizing NVIDIA DGX techniques and software, and the rate of innovation is not slowing down.
New NVIDIA Base Command Functions
NVIDIA Base Command presents enterprise-grade orchestration and cluster administration, and it now attributes a entire software stack for maximizing AI developer productiveness, IT manageability and workload performance.
The workflow management features of Base Command now consist of assistance for on-premises DGX SuperPOD environments, enabling corporations to get centralized command of AI development tasks with simplified collaboration for project groups, and built-in monitoring and reporting dashboards.
Foundation Command is effective with the NVIDIA AI Company software package suite, which is now bundled with just about every DGX process. The NVIDIA AI software permits conclusion-to-stop AI progress and deployment with supported AI and data science instruments, optimized frameworks and pretrained models.
Additionally, it offers organization-workflow administration and MLOps integrations with DGX-Prepared Software companies Domino Facts Lab, Operate.ai, Weights & Biases and NVIDIA Inception member Rescale. It also contains libraries that optimize and accelerate compute, storage and network infrastructure — although making certain maximized method uptime, stability and trustworthiness.
New DGX BasePOD Reference Architecture
DGX BasePOD presents a reference architecture for DGX programs that incorporates layout ideal procedures for integrating compute, networking, storage and software program.
Clients are now employing NVIDIA DGX POD to electricity the enhancement of a wide vary of enterprise purposes. DGX BasePOD builds on the results of DGX POD with new field remedies concentrating on the largest AI alternatives, which include pure language processing, healthcare and daily life sciences, and fraud detection.
Sent as entirely integrated, ready-to-deploy offerings as a result of the NVIDIA Husband or wife Network, DGX BasePOD remedies vary in sizing, from two to hundreds of DGX units, with certified significant-efficiency storage from NVIDIA DGX storage know-how partners including DDN, Dell, NetApp, Pure Storage, Vast Facts and WEKA.
Leaders Electric power AI Breakthroughs With DGX Programs
Enterprises all around the earth choose NVIDIA DGX units to energy their most innovative AI workloads. Among the the AI innovators establishing mission-significant AI abilities on DGX A100 systems:
- ML investigate and products lab Adept is creating an AI teammate driven by a massive language product prototyped on NVIDIA DGX Foundry, and then scaled with NVIDIA A100 GPUs and NVIDIA Megatron on Oracle Cloud Infrastructure.
- Hyundai Motor Team is applying a 40-node DGX SuperPOD to discover hyperscale AI workloads.
- Telecom corporation KT is producing a LLM with all around 40 billion parameters for a assortment of Korean-language purposes, such as the GiGA Genie good speaker, applying the NVIDIA NeMo Megatron framework, NVIDIA DGX SuperPOD and NVIDIA Base Command software package.
- The University of Wisconsin-Madison is swiftly bringing AI to clinical imaging gadgets making use of NVIDIA DGX devices with the Flywheel analysis system and the NVIDIA Clara healthcare application framework. Using the NVIDIA Federated Understanding Application Runtime Setting, or NVIDIA FLARE, in collaboration with other hospitals, the college is securely training AI designs on DGX units for healthcare imaging, annotation and classification.
Master much more about the AI breakthroughs run by NVIDIA DGX devices by seeing NVIDIA founder and CEO Jensen Huang’s GTC keynote in replay. And sign up for the GTC session, “Creating Your AI Heart of Excellence,” with Charlie Boyle, vice president of DGX techniques at NVIDIA.
Leave a comment