NVIDIA DGX™ A100
First AI System Built On New NVIDIA A100 GPU
- 8X NVIDIA A100 GPUS WITH 320 GB TOTAL GPU MEMORY
12 NVLinks/GPU, 600 GB/s GPU-to-GPU Bi-directonal Bandwidth
- 6X NVIDIA NVSWITCHES
4.8 TB/s Bi-directional Bandwidth, 2X More than Previous Generation NVSwitch
- DUAL 64-CORE AMD CPUs AND 1 TB SYSTEM MEMORY
3.2X More Cores to Power the Most Intensive AI Jobs
- 15 TB GEN4 NVME SSD
25GB/s Peak Bandwidth, 2X Faster than Gen3 NVME SSDs
The Universal System for All AI Workloads
NVIDIA DGX™ A100 offers unprecedented compute density, performance, and flexibility. The system features the world’s most advanced accelerator, the NVIDIA A100 Tensor Core GPU, enabling enterprises to consolidate training, inference, and analytics into a unified, easy-to-deploy AI infrastructure that includes direct access to NVIDIA AI experts and multi-layered built-in security.
The DGX A100 packs 5 petaFLOPS of AI performance into a 6U form factor, replacing legacy compute infrastructure with a single, unified system. DGX A100 also offers the unprecedented ability to deliver fine-grained allocation of computing power, using the Multi-Instance GPU capability in the NVIDIA A100 Tensor Core GPU, which enables administrators to assign resources that are right-sized for specific workloads. This ensures that the largest and most complex jobs are supported, along with the simplest and smallest. Running the DGX software stack with optimized software from NGC, the combination of dense compute power and complete workload flexibility make DGX A100 an ideal choice for both single node or large scale clusters deployments.
Scale Up with NVIDIA DGX SuperPOD
The NVIDIA DGX SuperPOD™ with NVIDIA DGX™ A100 systems is the next generation artificial intelligence (AI) supercomputing infrastructure, providing the computational power necessary to training today’s state-of-the-art deep learning (DL) models and to fuel innovation well into the future. It is designed to tackle the most important challenges of AI at scale, delivering unmatched levels of multi-system training.
Traditional large compute clusters are constrained by the increasing impact to performance associated with inter-GPU communications as configurations become larger and computation is parallelized over more and more nodes. This results in diminishing returns in terms of performance gained by incremental compute nodes. The DGX SuperPOD has demonstrated world-record-breaking performance and versatility in MLPerf 0.61, setting eight records in AI performance. Deploy in weeks and solve the world’s most challenging computational problems with the DGX SuperPOD now!
NVIDIA A100 Tensor Core GPU
Powerful Components of DGX A100
- 8X NVIDIA A100 GPUS WITH 320 GB TOTAL GPU MEMORY — 12 NVLinks/GPU, 600 GB/s GPU-to-GPU Bi-directonal Bandwidth
- 2 6X NVIDIA NVSWITCHES — 4.8 TB/s Bi-directional Bandwidth, 2X More than Previous Generation NVSwitch
- 3 9x MELLANOX CONNECTX-6 200Gb/S NETWORK INTERFACE — 450 GB/s Peak Bi-directional Bandwidth
- 4 DUAL 64-CORE AMD CPUS AND 1 TB SYSTEM MEMORY — 3.2X More Cores to Power the Most Intensive AI Jobs
- 5 15 TB GEN4 NVME SSD — 25GB/s Peak Bandwidth, 2X Faster than Gen3 NVME SSDs
Technology Inside NVIDIA DGX A100
Multi-Instance GPU (MIG)
The eight A100 GPUs in DGX A100 can be configured into as many as 56 GPU instances, each fully isolated with their own high-bandwidth memory, cache, and compute cores.
NVLink & NVSwitch
Third generation of NVIDIA® NVLink™ doubles the GPU-to-GPU direct bandwidth to 600 GB/s, almost 10X higher than PCIe Gen4 and next-generation NVIDIA NVSwitch™, which is 2X times faster than the previous generation.
New Mellanox ConnectX-6 VPI HDR InfiniBand/Ethernet adapters running at 200 gigabits per second (Gb/s) to create a high-speed fabric for large-scale AI workloads.
Optimized Software Stack
Integrated DGX software stack, including an AI-tuned base operating system, all necessary system software, and GPU-accelerated applications, pre-trained models, and more.
Most robust security for AI deployments, with a multi-layered approach stretching across the baseboard management controller (BMC), CPU board, GPU board, self-encrypted drives, and secure boot.
Essential Building Block of the AI Data Center
The Universal System for Every AI Workload
One universal building block to run any workload anytime – from analytics, AI training to inference.
Integrated Access to AI Expertise
Fast-track AI transformation with NVIDIA DGXperts to help maximize the value of DGX investment.
Game-changing Performance for Innovators
Provides unprecedented acceleration with eight A100 GPUs and is fully optimized for NVIDIA CUDA-X™ software and end-to-end NVIDIA data center solution stack.
Unmatched Data Center Scalability
Build leadership-class AI infrastructure that scales to keep ahead of demand.
Faster Analytics Means Deeper Insights to Fuel AI Development
Faster Training Enables the Most Advanced AI Models
Faster Inference Increases ROI Through Maximized System Utilization
8x NVIDIA A100 Tensor Core GPUs
320 GB total
- 5 petaFLOPS AI
- 10 petaOPS INT8
- Dual 20-Core Intel Xeon
- E5-2698 v4 2.2 GHz
System Power Usage
Dual AMD Rome 7742, 128 cores total, 2.25 GHz(base), 3.4 GHz (max boost)
- 8x Single-Port Mellanox ConnectX-6 VPI
- 200Gb/s HDR Infiniband
- 1x Dual-Port Mellanox
- ConnectX-6 VPI 10/25/50/100/200Gb/s Ethernet
- OS: 2x 1.92TB M.2 NVME drives
- Internal Storage: 15TB
- (4x 3.84TB) U.2 NVME drives
Ubuntu Linux OS
271 lbs (123 kgs)
- Height: 10.4 in (264.0 mm)
- Width: 19.0 in (482.3 mm) MAX
- Length: 35.3 in (897.1 mm) MAX
Packing System Weight
315 lbs (143kgs)