FusionServer G530 V5 is a half-width heterogeneous compute node with high-density heterogeneous computing capabilities and supports GPUDirect RDMA and Peer-to-Peer (P2P) interconnection technologies. It supports multiple CPUs. The GPU ratios provide optimal heterogeneous topology configurations for various application loads. What’s more, it supports long-term evolution of CPUs and heterogeneous components based on a fully modular design. FusionServer G530 V5 adapts to application scenarios such as AI inference, HPC, intelligent cloud, and database acceleration.
Description
Technical Specifications
FusionServer G530 V5 Half-width Heterogeneous Compute Node
Half-width heterogeneous compute node
A maximum of 16 GPUs
2 x CPUs
Up to 24 x DDR4 DIMMs
16 x T4: two 2.5-inch NVMe SSDs, SAS drives, or SATA drives
4 x dual-slot GPUs or 8 x single-slot GPUs: 4 x 3.5-inch SAS/SATA HDDs and 2 x 2.5-inch NVMe SSDs/SAS/SATA HDDs
RAID 0, 1, 10, 5, 50, 6, and 60, and supercapacitor protection
Microsoft Windows Server，Red Hat Enterprise Linux，SUSE Linux Enterprise Server，Cent OS，Ubuntu
Onboard BMC management module
Management features such as IPMI, SOL, KVM over IP, and virtual media
5°C–35℃ (41°F–95°F)
