The NVIDIA DGX A100 is the universal system for all AI workloads, offering unprecedented compute density, performance, and flexibility in the world's first 5-petaFLOPS AI system. It integrates eight NVIDIA A100 Tensor Core GPUs, providing a unified platform for training, inference, and data analytics.
- 8x NVIDIA A100 Tensor Core GPUs with up to 640GB total GPU memory
- Dual AMD EPYC 7742 processors providing 128 cores at 2.25GHz base clock
- 2TB of DDR4 system memory for massive data processing and high-speed throughput
- 6x NVIDIA NVSwitches providing 4.8TB/s of bidirectional GPU-to-GPU bandwidth
- 9x Mellanox ConnectX-6 VPI network adapters supporting 200Gb/s HDR InfiniBand
- 30TB of internal NVMe Gen4 storage for high-speed data access and caching
- 6U rackmount form factor designed for standard enterprise data center integration
- Delivers 5 petaFLOPS of AI performance in a single node
- Optimized AI software stack including NVIDIA AI Enterprise and NGC containers
- Redundant 3000W power supplies and high-efficiency cooling systems
Related Products
A relatable product