ASUS has introduced its brand-new HGX H100 eight-GPU AI server: the ESC N8-E11, as well as an extensive selection of PCIe GPU servers known as the ESC8000 and ESC4000 series, powered by Intel and AMD platforms to support higher CPU and GPU TDPs.
To begin with, the new ESC N8-E11 is equipped with eight NVIDIA H100 Tensor Core GPUs and is meticulously developed to minimize the time taken for large-scale AI training models and HPC. Powered by 4th Gen Intel Xeon Scalable processors, this 7U dual-socket server showcases a one-GPU-to-one-NIC topology that can accommodate up to eight NICs.
With its modular design, the server significantly decreases cable usage which results in shorter assembly time, elimination of cable routing complexities, and also lower risk of airflow obstruction
Moreover, the ESC N8-E11 includes cutting-edge technologies such as fourth-generation NVLink and NVSwitch, NVIDIA ConnectX-7 SmartNIC, as well as the software layer of the NVIDIA AI platform which boosts the development of AI and data science.
Not only is the server engineered with a two-level GPU and CPU sled to optimize thermal efficiency and scalability, but it is also compatible with direct-to-chip liquid cooling, which greatly reduces the overall power-usage effectiveness of a data center.
Kindly refer to the link below for more details on the new ASUS ESC N8-E11: