AMD EPYC 9845 Turin Dense 160-Core Processor

The AMD EPYC 9845 is a 4th Generation EPYC processor built on the Zen 5c core architecture, delivering 160 cores and 320 threads within a single socket. As part of the Turin Dense family, it leverages AMD's compact Zen 5c core design to maximize compute density per socket, making it exceptionally well-suited for cloud service providers and hyperscalers that require high vCPU counts per physical host while managing per-core software licensing overhead.

Constructed using a chiplet-based design with TSMC 3nm process technology for the compute dies and paired with an advanced I/O die, the EPYC 9845 supports up to 384GB of DDR5 memory across 12 channels, delivering substantial memory bandwidth for latency-sensitive and throughput-intensive workloads. The processor integrates PCIe 5.0 connectivity and maintains full compatibility with the SP5 platform socket, enabling deployment in existing Turin-generation server infrastructure.

The EPYC 9845 is targeted specifically at multi-tenant cloud environments, virtual desktop infrastructure, containerized microservices platforms, and high-density virtualization deployments where maximizing the number of independently schedulable compute threads per node directly translates to improved consolidation ratios. Its positioning within AMD's Turin Dense lineup reflects a deliberate architectural trade-off that prioritizes core count and power efficiency over the highest individual-core clock speeds found in standard Zen 5 variants.

Ideal for

• High-density public and private cloud virtualization requiring maximum vCPU count per physical host to optimize per-core software licensing spend
• Multi-tenant container orchestration platforms running Kubernetes clusters where thread-level isolation and density drive infrastructure economics
• Virtual Desktop Infrastructure deployments supporting large concurrent user populations on minimal physical server footprint
• Cloud-native microservices architectures demanding high parallelism across hundreds of simultaneous lightweight workloads
• Batch processing and HPC job scheduling environments where aggregate throughput across many parallel tasks exceeds the benefit of high single-thread performance
• Telecom and edge cloud infrastructure requiring dense compute nodes for NFV and vRAN workloads within constrained power and space envelopes

Technical specifications

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