- KH590S3
- 2-socket Korean's GPU server with the latest 4th generation Intel® Xeon® Scalable processors and x86 motherboards self-designed and built with pure proprietary technology (4U rack-mounted form factor)
This is High-performance, highly scalable GPU server systems optimized for deep learning and machine learning-based artificial intelligence systems with up to 11 or 20 GPU accelerators
1. Key Technical Features
· 2-socket server with the latest 4th Generation Intel® Xeon® Scalable Processors
· Supports up to 120 Cores per system (up to 350W TDP)
· Supports up to 32 DDR5 DIMM memory slots (8TB and 4,800 MT/s)
2. Flexible expansion according to customer needs
· 8/12/24 Bay Hot swap drive bay configurable with regular SAS and SATA drives
· Supports 2 x embedded NVMe SSDs (M.2) for fast booting
· Supports up to 11 or 20 PCIe expansion slots for multiple GPU accelerators
3. High availability and Operational efficiency
· Redundancy design for key devices such as '3+1' power and FAN enables online status replacement in the event of a failure
· Compute module layout and cooling design for GPU heat generation and Air Flow
· BMC enables remote support, FAN control with GPU temperature monitoring and system monitoring
· Enable efficient server system management with IPMI 2.0 and RESTful APIs (including Redfish)