newsonline24

At its core is an AMD® EPYC™ 9005 Series Turin server processor with up to 160 cores and 320 threads. For massively parallel AI, inference and analytics workloads, the RGS-9805GC supports two dual-slot NVIDIA® FHFL GPUs rated at up to 400 W each. This makes the system suitable for applications that require large data volumes to be processed at the point of origin and low-latency decisions to be made in real time.

Seven PCIe Gen5 expansion slots provide substantial capacity for project-specific configurations. In addition to two PCIe x16 slots for GPUs, five further slots can accommodate frame grabbers, high-speed network adapters or interface cards for GMSL2, CoaXPress and USB cameras. The server can therefore be tailored flexibly to different sensor, camera and network topologies.

For memory- and data-intensive real-time processing, the RGS-9805GC supports up to 384 GB of DDR5-6000 ECC RDIMM memory. Two M.2 NVMe sockets, a front-accessible U.2 tray and an internal SlimSAS connector enable high-performance storage and data-logging configurations. Network connectivity includes two 10GBASE-T ports with RDMA support and an additional Gigabit Ethernet port.

An integrated automotive MCU adds CAN 2.0, ignition power control, isolated digital output and continuous system monitoring. The 9–32 VDC wide-range input supports deployment in vehicles and mobile platforms. For harsh operating environments, the server is designed to meet MIL-STD-810H shock and vibration requirements and is specified for operating temperatures from -25 °C to +60 °C.

With compact dimensions of 444 × 388 × 130 mm, the RGS-9805GC can be installed in a 19-inch rack or integrated using wall-mount or damping brackets. Its combination of AMD® EPYC™ processing, dual-GPU acceleration and versatile PCIe expansion makes it a powerful platform for autonomous driving, HD mapping, industrial machine vision, physical AI and other mission-critical edge HPC applications.