4U AI Training System with 8 Habana® Gaudi® Deep Learning Processors and dual 3rd Gen Intel® Xeon® Scalable Processors
The Supermicro X12 Gaudi AI Training System, powered by Habana Gaudi Deep Learning Processors, pushes the boundaries of deep learning training and can scale up to hundreds of Gaudi processors in one AI cluster. Gaudi is the first DL training processor with integrated RDMA over Converged Ethernet (RoCE v2) engines on-chip. With bi-directional throughput of up to 2 TB/s, these engines play a critical role in the inter-processor communication needed during the training process. This native integration of RoCE allows customers to use the same scaling technology, both inside the server and rack (scale-up) and across racks (scale-out). These can be connected directly between Gaudi processors or through any number of standard Ethernet switches.
Key Applications
- Purpose Built for AI/Deep Learning Training
- Computer Vision
- Natural Language Processing
- Recommendation
Key Features
- High Density 4U System supporting 8 Habana Gaudi HL-205 AI Processors (supported via included carrier board HGI-MEZZ)
- 24 on-board 100GbE RDMA (via 6x QSFP-DDs) for scale-out
- 1 PCI-E 4.0 x16 FHHL slot + 2 PCI-E 4.0 AIOM (OCP 3.0 superset) slots