Intel Xeon 6700-series (Granite Rapids 6747P, 48-core)

The Intel Xeon 6747P is a 48-core mid-tier member of the Xeon 6700-series Granite Rapids family, fabricated on Intel's advanced Intel 3 process node and built on the modular tile-based architecture that defines the Granite Rapids generation. Each processor integrates four compute tiles interconnected via Intel's Embedded Multi-die Interconnect Bridge (EMIB), enabling high inter-die bandwidth while maintaining a competitive die area. The architecture features the updated Redwood Cove microarchitecture per core, delivering improved instructions-per-clock efficiency, larger per-core L2 caches, and substantially expanded on-package Last Level Cache (LLC) that reduces DRAM pressure for latency-sensitive enterprise applications. Support for DDR5 memory at high frequencies and PCIe 5.0 lanes ensures the platform remains forward-compatible with next-generation accelerators, NVMe storage fabrics, and high-speed networking infrastructure.

Targeted at mainstream enterprise virtualization servers, the Xeon 6747P occupies an optimal position in the Granite Rapids lineup by providing 48 physical cores without reaching the thermal and platform complexity of the highest-tier 60+ core SKUs. This makes it well-suited for two-socket rack servers where operators seek high virtual machine density, broad thread concurrency for cloud-native workloads, and robust memory capacity through 12-channel DDR5 per socket. The processor supports Intel Virtualization Technology for Directed I/O (VT-d), Intel Total Memory Encryption – Multi-Key (TME-MK), and Intel Software Guard Extensions (SGX), addressing the security and isolation requirements of multi-tenant enterprise and service-provider environments.

Launched in Q2 2025, the Xeon 6747P inherits the Granite Rapids platform's expanded I/O capabilities including up to 136 PCIe 5.0 lanes per socket, CXL 2.0 memory expansion support, and compatibility with Intel Advanced Matrix Extensions (AMX) for accelerated AI inference and matrix-intensive HPC workloads. The processor operates within a controlled TDP envelope that balances compute throughput with power efficiency in high-density rack deployments, making it a practical choice for enterprises modernizing aging Skylake or Ice Lake infrastructure while requiring predictable performance-per-watt characteristics.

Ideal for

• High-density VMware vSphere or Red Hat OpenShift virtualization hosts requiring broad vCPU availability across dozens of concurrent virtual machines per socket
• Kubernetes and container orchestration nodes in private cloud infrastructure demanding high thread counts and fast memory bandwidth for pod scheduling and service mesh processing
• Database server consolidation workloads including Oracle, Microsoft SQL Server, and PostgreSQL instances benefiting from large LLC capacity to reduce query latency
• AI inference and batch scoring pipelines leveraging Intel AMX instructions for accelerated INT8 and BF16 matrix operations without requiring discrete GPU acceleration
• HPC and technical computing clusters running MPI-parallel simulations in computational fluid dynamics, finite element analysis, and genomics sequencing pipelines
• Enterprise SAP HANA scale-out nodes and ERP application tiers where balanced core count, DDR5 memory capacity, and CXL expansion deliver deterministic transactional throughput

Technical specifications

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