Inventec, a supplier of servers based in Taiwan, is set to unveil its P8000IG6 server, engineered to swiftly and flexibly scale up data center capacity, catering to the most demanding AI and high-performance computing (HPC) workloads.
In an era marked by the proliferation of AI technologies, the demand for data centers is surging at an unprecedented pace. According to McKinsey, data center operators grapple with the pressing need for increased capacity amidst escalating capital costs, driving a shift towards prefabricated and modular solutions to expedite expansion while curtailing setup time.
George Lin, VP of Business Unit VI at Inventec Enterprise Business Group (Inventec EBG), underscores the significance of the P8000IG6 server in addressing the burgeoning demand for scalable capacity in the AI age.
The P8000IG6 server boasts top-tier performance coupled with unparalleled flexibility:
Powered by up to eight GPUs, the P8000IG6 leverages the NVIDIA HGX AI supercomputing platform, meticulously optimized for the most intricate AI and HPC workloads. Customers can choose from an array of NVIDIA H100, H200, or B100 GPUs.
Seamless interconnectivity is ensured through NVIDIA NVLink and NVSwitch, expediting communication among all eight GPUs onboard. Inventec’s switchboard complements this setup, facilitating seamless connectivity with other GPUs within the data center and offering maximal bandwidth for GPU Direct RDMA.
The motherboard, housed in a separate tray from the GPUs, is equipped with Intel Xeon Scalable Processors (Eagle Stream), providing customers with options between 4th Gen and 5th Gen processors.
Furthermore, the server’s modular architecture streamlines deployment and installation processes, facilitates infrastructure expansion and upgrades, and ensures compatibility with various server configurations. Inventec’s team of experts extends white-glove customization services to cater to data centers seeking specialized solutions.
The P8000IG6 server encompasses leading-edge connectivity, power, and cooling features:
It integrates the latest DDR5 SDRAM and PCIe Gen. 5.0 connectivity to enhance operational efficiency.
Decoupled power sources for the CPU and GPU modules — 12V and 54V, respectively — enhance power efficiency and reliability. The server offers both air and liquid cooling options, augmenting its versatility.
Maintenance of the P8000IG6 server is streamlined through its modular design. With separate trays for the CPU and GPU modules and various service-friendly modules — including fans, power supply, and NIC — combined with a tool-less design, maintenance procedures are simplified for enhanced operational efficiency.
InfotechLead.com News Desk
