ASRock Industrial's Xeon 600 Edge AI Platform

[Erkan Teskancan]

## ASRock Industrial's Xeon 600 Edge AI Platform

ASRock Industrial has introduced its new 4U Edge AI platform, optimized for industrial environments with multi-GPU support and high memory capacity.

Edge AI applications such as industrial visual inspection, real-time analytics, and generative AI require data center-class computing power outside of central data centers. In response to this need, ASRock Industrial has announced the iEPF-11000S Series Edge AIoT platform, developed with Intel® Xeon® 600 processors and the W890 chipset.

### iEPF-11000S Series Features

• Designed for compute-intensive Edge AI tasks with support for four GPUs
• Intel® Xeon® 600 series processors
• W890 chipset
• Up to 2 TB of quad-channel DDR5 ECC RDIMM or RDIMM-3DS memory
• Ideal for multiple AI model training, inference, predictive maintenance, and large language model (LLM) applications
• Six PCIe Gen5 x16 and one PCIe Gen5 x8 expansion slots
• Gen5 bandwidth for NVMe storage and high-performance GPUs
• Data integrity ensured with ECC-supported memory
• Eight SATA3 ports and Intel® VMD RAID 0/1/5/10 data management
• Storage expansion with two M.2 Key M slots
• Dual 1 GbE LAN and optional dual 10 GbE network connectivity
• 602.6 × 175 × 438 mm dimensions with 4U rack-mount and tower configuration support
• 1600 W ATX power supply for high-performance GPU support
• Security and remote management with TPM 2.0 and Intel® vPro

### iEPF-10000S Series Summary

• Energy and space savings with less GPU capacity
• Intel® Core™ Ultra (Arrow Lake-S) and Intel® Core™ Series 2 (Bartlett Lake-S) processors
• Supports two GPUs
• Suitable for machine vision, AI-assisted inspection, and industrial analytics tasks
• Ideal for production lines and equipment integration due to its small footprint and low power consumption

### Remote Management and Edge AI Approach
Both series support out-of-band (OOB) management modules to enable remote monitoring and control in distributed Edge AI applications. The iEPF-11000S Series uses the AI-M2-OOB-1G module, while the iEPF-10000S Series uses the AI-OOB-LITE module.

These platforms combine multi-GPU scalability, high-capacity ECC memory, and remote management, enabling AI training and inference operations to be performed in industrial environments without relying on central data centers.