ASRock Rack BERGAMOD8-2L2T Server Board (EPYC 9004 Bergamo)

The ASRock Rack BERGAMOD8-2L2T is a high-density server motherboard engineered specifically for AMD EPYC 9004 Series Bergamo processors, supporting the full Zen 4c core architecture that enables up to 128 cores and 256 threads on a single socket. Built around the SP5 socket platform, the board leverages AMD's advanced chiplet design to deliver exceptional thread density within a compact footprint, making it a purpose-built solution for environments demanding extreme parallelism and high workload concurrency. The board supports DDR5 memory across eight DIMM slots, enabling configurations suited to memory-intensive cloud and containerized deployments.

Networking capabilities are a central strength of the BERGAMOD8-2L2T, with the model designation reflecting its dual 25GbE SFP28 and dual 10GbE BASE-T onboard LAN configuration — the "2L2T" suffix denoting two optical and two copper high-speed interfaces. This multi-port arrangement allows the board to serve as a dense compute node with sufficient bandwidth headroom for east-west traffic typical of software-defined networking, microservices architectures, and hyperconverged infrastructure. Storage connectivity is provided via multiple PCIe Gen 5 and Gen 4 expansion slots alongside SATA and M.2 NVMe interfaces, accommodating tiered storage strategies within a single-board design.

Targeted at cloud service providers, telecommunications edge deployments, and enterprise private cloud builders, the BERGAMOD8-2L2T is optimized for workloads that prioritize throughput-per-watt and core density over single-thread peak performance. The Bergamo EPYC architecture's Zen 4c cores are tuned for exactly this profile — delivering cloud-native compute efficiency at scale. IPMI-based out-of-band management via a dedicated ASPEED AST2600 BMC enables lights-out administration, firmware lifecycle management, and integration with standard data center orchestration tooling, supporting the operational demands of large-scale enterprise and service-provider environments.

Ideal for

• High-density Kubernetes and container orchestration nodes requiring maximum vCPU count per 1U or 2U chassis slot for microservices platform infrastructure
• Cloud-native private cloud buildouts using OpenStack or similar IaaS platforms where thread density and memory bandwidth per node directly translate to virtual machine consolidation ratios
• Telecommunications Network Function Virtualization (NFV) and vRAN deployments demanding simultaneous high-throughput processing across large numbers of lightweight software workloads
• Big data analytics clusters running Apache Spark, Flink, or Hadoop where per-node parallelism reduces inter-node shuffle overhead and improves job completion times
• Content delivery and media transcoding infrastructure leveraging high core counts for parallel codec and streaming pipeline processing across multiple simultaneous sessions
• Software-defined storage controller nodes where CPU thread availability supports concurrent I/O handling, erasure coding computation, and metadata management at enterprise scale

Technical specifications

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