High Performance for Deep Learning
- NVIDIA HGX A100 Server Architecture with:
- 4 x A100 NVlink 80GB GPU
- Direct connect PCI-E Gen4 Platform with NVIDIA® NVLink™
- Dual AMD EPYC™ 7003 Series Processors
- Up to 8TB Registered ECC DDR4 3200MHz SDRAM in 32 DIMMs
- 4 PCI-E Gen 4 x16 (LP), 1 PCI-E Gen 4 x8 (LP)
- 4 Hot-swap 2.5″ drive bays(SAS/SATA/NVMe Hybrid)
- 2x 2200W Power Supplies
- 3 Years Manufacturer’s Warranty
- Optional: Ubuntu Linux OS install
- Optional: Dihuni AI Stack Pre-loaded (Packages include NVIDIA Cuda Toolkit, NVIDIA DIGITS, NVIDIA TensorRT, Keras and NVCaffe, Caffe2, Microsoft Cognitive Toolkit (CNTK), MXNet, PyTorch, TensorFlow, Theano, Torch and others as needed