ASUS RS520QA-E13-RS8U Multi-Node Server

The ASUS RS520QA-E13-RS8U is a 2U quad-node server platform built on the AMD EPYC 9005 series (Zen 5 architecture), delivering up to four independent compute nodes within a single 2U chassis. Each node supports dual AMD EPYC 9005 processors, enabling organizations to consolidate massive parallel workloads into a highly space-efficient form factor. The platform is purpose-designed for high-performance computing, AI inferencing at scale, and dense virtualization environments where rack space, power efficiency, and thermal management are primary operational constraints. With PCIe 5.0 connectivity, DDR5 memory subsystems, and high-speed fabric interconnect readiness, each node operates as a fully autonomous server instance while sharing chassis-level power and cooling infrastructure to reduce total cost of ownership at scale.

Each of the four compute nodes accommodates up to 12 DDR5 DIMMs per node (24 per dual-socket configuration), providing substantial memory bandwidth to feed the wide core counts available in the EPYC 9005 family, which scales up to 192 cores per socket. The RS520QA-E13-RS8U integrates shared hot-swap NVMe and SATA storage bays accessible per node, along with dedicated PCIe 5.0 expansion slots for GPU accelerators, high-speed NICs, or NVMe expansion cards. The chassis incorporates a shared redundant power supply subsystem with platinum-level efficiency, contributing to measurable PUE improvements in dense deployments. Out-of-band management is handled per node via ASUS ASMB11 baseboard management controllers compatible with IPMI 2.0, Redfish, and ASUS Control Center Enterprise for unified fleet orchestration.

The RS520QA-E13-RS8U is particularly compelling for operators building next-generation HPC clusters, AI/ML training and inference farms, cloud service provider infrastructure, and financial analytics platforms. Its multi-node architecture achieves server density of up to four independent compute nodes per 2U of rack space, effectively quadrupling node density compared to traditional 1U single-node designs and dramatically reducing cabling complexity, power distribution overhead, and rack unit consumption. The platform supports mainstream Linux distributions and major hypervisors, and is validated for leading HPC workload managers including SLURM and PBS Pro, making it immediately deployable in production scientific and enterprise environments.

Ideal for

• High-performance computing (HPC) clusters requiring extreme node density per rack unit for simulation, molecular dynamics, and computational fluid dynamics workloads
• AI and machine learning inference farms where multiple independent GPU-accelerated nodes must be co-located within constrained rack environments
• Cloud service provider and hyperscale data center buildouts demanding maximum compute density with shared power and cooling infrastructure
• Financial services firms running low-latency quantitative analytics and Monte Carlo simulations across many parallel compute threads
• Enterprise virtualization consolidation projects leveraging high EPYC core counts and large DDR5 memory pools to maximize VM density per rack unit
• Research institutions and national laboratories deploying scale-out compute fabrics for genome sequencing, climate modeling, and physics simulation pipelines

Technical specifications

Available from Omnixon Global. Submit an RFQ and our team will confirm configuration and availability for your order.