Lenovo ThinkSystem SR670 Rack Server

Lenovo ThinkSystem SR670 Rack Server (CTO) - 3 Years Local Warranty

High-Performance 2U AI and GPU Server Platform

Lenovo ThinkSystem SR670 Rack Server (CTO) is a powerful 2U dual-socket server engineered for GPU-accelerated workloads, artificial intelligence, deep learning, and high-performance computing deployments. Designed for enterprises, research institutions, and AI infrastructure teams, the SR670 supports multiple GPU accelerators per node, making it ideal for intensive AI training, inference, and simulation workloads demanding high compute density.

Dual Intel Xeon Scalable Processor Performance

The ThinkSystem SR670 is powered by dual Intel Xeon Scalable processors, providing substantial CPU computing capacity for AI and HPC workloads. The dual-socket architecture enables organizations to scale CPU cores and memory capacity, balancing compute resources needed for data ingestion, preprocessing, and model serving alongside GPU processing.

GPU Acceleration for AI and HPC Workloads

The SR670 supports up to 8 double-wide GPUs, enabling organizations to build high-density GPU compute clusters within a compact 2U form factor. It is compatible with NVIDIA GPU accelerators, supporting AI model training, deep learning inference, simulation, and rendering workloads. The server's GPU interconnect design enables high bandwidth communication between GPU accelerators.

Enterprise-Ready Reliability and Management Features

The ThinkSystem SR670 includes enterprise reliability features such as redundant power supplies, hot-swap fans, and Lenovo XClarity Controller for out-of-band management and monitoring. These features support continuous operations in demanding data center environments and enable efficient lifecycle management.

Technical Specifications

General
Product Name: Lenovo ThinkSystem SR670 Rack Server (CTO)
Product Type: 2U Dual-Socket GPU Server
Form Factor: 2U Rack
Processor Architecture: Intel Xeon Scalable
Deployment: AI Training, HPC, Deep Learning, GPU Computing, Data Center Workloads

Processor Specifications
Processor Sockets: 2 (Dual Socket)
Processor Support: Intel Xeon Scalable Family (2nd or 3rd Generation)
Cores per Processor: Up to 28 cores per processor (model-dependent)
Max Threads: Up to 56 threads per processor

Memory Specifications
Memory Type: DDR4 RDIMM / LRDIMM
Memory Slots: 24 DIMM Slots
Max Memory Capacity: Up to 3 TB DDR4
Memory Speed: Up to 3200 MT/s
ECC Memory Support: Supported

GPU Specifications
GPU Support: Up to 8 × Double-Width GPUs
Compatible GPU Models: NVIDIA Tesla / A-Series / H-Series Accelerators
GPU Interconnect: NVLink / PCIe (model-dependent)
GPU Power Support: High-power GPU configurations supported

Storage Specifications
Drive Bays: Up to 8 × 2.5-inch SFF hot-swap bays
Storage Interfaces: SAS, SATA, NVMe
Drive Types: HDDs, SSDs, NVMe drives
M.2 Boot Support: Supported (2 M.2 slots)

Expansion & PCIe Slots
PCIe Expansion Slots: Multiple PCIe Gen 3 slots
Storage Controller Options: ThinkSystem RAID Adapters (RAID 0, 1, 5, 6, 10)
Network Expansion: PCIe Network Adapters

Power & Redundancy Specifications
Power Supply Options: High-capacity Hot-swap Redundant PSU (configuration-dependent)
Power Efficiency: 80 PLUS Platinum

Cooling & Chassis
Cooling: Redundant Hot-Swap Fans
Fan Type: Hot-swap replaceable
Chassis Size: 2U Rack
GPU Airflow Optimization: Designed for GPU thermal management

Management & Security
Management Controller: Lenovo XClarity Controller (XCC)
Out-of-Band Management: Supported
Trusted Platform Module (TPM): Supported
UEFI Firmware: Supported

Operating System Compatibility
Windows Server 2019
Windows Server 2022
Red Hat Enterprise Linux (RHEL)
SUSE Linux Enterprise Server (SLES)
Ubuntu Server
VMware vSphere

Physical Specifications
Form Factor: 2U Rack
Width: 17.6 inches (Rack standard)
Weight: Configuration-dependent

Key Features
2U Dual-Socket AI GPU Server
Up to 8 × double-wide GPU support
Intel Xeon Scalable processors
Up to 24 DIMM slots
NVMe and SAS/SATA storage support
Lenovo XClarity management

Typical Use Cases
AI model training
Deep learning inference
HPC simulation
Scientific research computing
GPU computing clusters
Rendering workloads
Data analytics
Enterprise AI infrastructure