AMD Processor

AMD EPYC 9005 Series (Turin)

Up to 192 cores / 384 threads (dual-socket: 384C/768T) Cores • AMD • Socket SP5 (LGA 6096)

AI Workstation Builders 24+ Years IT Infrastructure Custom Configurations Available
Cores
Up to 192 cores / 384 threads (dual-socket: 384C/768T)
TDP
Up to 500W per socket
PCIe Lanes
128 PCIe Gen 5 lanes per socket
Memory
12-channel DDR5 ECC per socket (up to 6TB dual-socket)
Socket
SP5 (LGA 6096)

Detailed Specifications

  • Cores: Up to 192 cores / 384 threads (dual-socket: 384C/768T)
  • Architecture: Zen 5 and Zen 5c
  • Socket: SP5 (LGA 6096)
  • Memory: 12-channel DDR5 ECC per socket (up to 6TB dual-socket)
  • Pcie Lanes: 128 PCIe Gen 5 lanes per socket
  • Tdp: Up to 500W per socket
  • Models: EPYC 9115 (16C), EPYC 9175F (16C high-freq), EPYC 9575F (64C), EPYC 9755 (128C), EPYC 9965 (192C)

Best For

  • Multi-GPU AI training servers (up to 8-10x GPUs per server)
  • Massive data preprocessing with 384 threads
  • High-memory AI workloads (up to 6TB RAM)
  • Dual-socket GPU servers for H200 NVL and RTX PRO 6000
  • HPC and scientific computing
  • Large-scale AI inference serving

Available in PTG Systems

The AMD EPYC 9005 Series (Turin) is available in Petronella-built AI workstations and servers, configured and optimized for your workload.

  • Up to 10x RTX PRO 6000 Blackwell or H200 NVL
  • Up to 6TB DDR5 ECC RAM
  • NVMe storage arrays

Why Buy from Petronella

PTG builds EPYC 9005 AI servers as the backbone for multi-GPU training and inference infrastructure. Dual-socket configurations with 256 PCIe 5 lanes feed up to 10 GPUs at full bandwidth.

  • Custom-built systems with tested configurations
  • GPU pairing recommendations for AI workloads
  • Enterprise support and warranty
  • On-site deployment available (Raleigh-Durham area)
  • Remote management and monitoring

Order a AMD EPYC 9005 Series (Turin) System

Talk to our hardware team for custom configurations, pricing, and availability.

Call (919) 348-4912 Request a Quote