As the year of AI blooms, all major tech companies focused on AI products. To go with the AI tide, the AI servers, the heart of all AI products, are more significant than ever. Inventec P8000IG6 was born to solve the eager need for AI product producers. With its extreme flexibility and comprehensively considerate design, it will be a powerful weapon for our customers to approach the future AI demand.
An Extremely Flexible 8U GPGPU server
P8000IG6’s modular architecture provides customers with the best performance platform. To fulfill diverse customers’requests, P8000IG6 uses Intel Eagle stream platform CPUs –Sapphire Rapids and Emerald Rapids, as its host node, andit adapts not only the NVIDIA H100/H200 HGX 8 GPU but also the H-Next HGX 8 GPU. Moreover, Inventec offersdifferent thermal SKUs including Air-Cooling and Liquid-Cooling solutions to apply the most efficient performance.
Comprehensively Considerate Design
Inventec P8000IG6 has the best GPU-GPU communication with Nvidia NVlink and NVswitch architecture on H100/200 8 GPU HGX architecture. In addition, to achieve the system’s best performance, the Inventec-designed switchboard bridges the Host node and the GPU node efficiently and offers maximum bandwidth for GPU direct RDMA. Moreover, to enhance power efficiency and reliability, P8000G6 equips 12V and 54V power sources to supply the Host node and GPU module individually and decouples 12V and 54V power sources to avoid ripple effect.
Great Serviceability with the Modular and Tool-less Design
A modular and tool-less design offers a comfortable maintenance experience for our customers, especially the way Inventec designs P8000IG6 with trays, modular, and tool-less design. The P8000IG6 separates the Host node and the GPU node with different trays and creates several modules for the key parts, such as fans, power supply, NIC, and others, that give our customers a much more comfortable way to maintain the specific parts.
