Lenovo ThinkSystem SR670 V2 Rack Server

From Exascale to Everyscale™

From single node enterprise deployments to the largest supercomputers in the world, the SR670 V2 can scale to meet any performance demand.

Buy Now

8x NVIDIA H100 GPUs (640GB of total GPU memory at 3.3TB/s) or 8x NVIDIA H200 GPUs (1,128GB of total GPU memory at 4.8TB/s), utilizing 4x NVIDIA NVSwitches for a total of 7.2TB/s of bi-directional GPU-to-GPU bandwidth
8x AMD Instinct™ MI300X (1,536GB of total GPU memory at 5.2TB/s) for a total of 6.4TB/s of bi-directional GPU-to-GPU bandwidth
Two 5^th Gen Intel^® Xeon^® Scalable processors with 32 DDR5 DIMM sockets
Up to 10x PCIe Gen5 x16 slots with 8x directly connected to the GPU complex
Up to 16x 2.5″ NVMe drives for maximum storage capacity and performance

Features

GPU Rich Platform

As more workloads leverage the capabilities of accelerators, the demand for GPU´s increases. The ThinkSystem SR670 V2 delivers optimal performance across industry verticals including retail, manufacturing, financial services and healthcare allowing for greater extraction of insights to drive innovation using machine learning and deep learning.

Lenovo Neptune™ Technology

Some models feature Lenovo Neptune™ hybrid cooling module which quickly dissipates heat in a closed loop liquid-to-air heat exchanger, delivering the benefits of liquid cooling without adding plumbing.

Accelerated Compute Platform

The NVIDIA^® A100 Tensor Core GPU delivers unprecedented acceleration—at every scale—to power the world’s highest performing elastic data centers for AI, data analytics, and high-performance computing (HPC) applications. The A100 can efficiently scale up or be partitioned into seven isolated GPU instances, with Multi-Instance GPU (MIG) providing a unified platform that enables elastic data centers to dynamically adjust to shifting workload demands.

The ThinkSystem SR670 V2 is designed to support the vast NVIDIA Ampere datacenter portfolio including NVIDIA HGX A100 4-GPU with NVLink, up to 8 NVIDIA A100 Tensor Core GPU with NVLink Bridge, and NVIDIA A40 Tensor Core GPU with NVLink bridge. Interested in other NVIDIA GPUs? See our full portfolio in the ThinkSystem and ThinkAgile GPU Summary.

See Standard specifications

Lenovo ThinkSystem SR680a V3 Rack Server

Specifications

Form Factor/Height

3U Rack-mount with three module

Processors

2x 3rd Gen Intel Xeon Scalable processors per node

Memory

Up to 2.0TB using 32x 64GB 3200MHz TruDDR4 3DS RDIMMs per node
Intel® Optane™ Persistent Memory 200 Series

Base Module

Up to 4x double-wide, full-height, full-length FHFL GPUs each PCIe Gen4 x16
Up to 8x 2.5” Hot Swap SAS/SATA/NVMe, or 4x 3.5” Hot Swap SATA (selected configurations)

Dense Module

Up to 8x double-wide, full-height, full-length GPUs each PCIe Gen4 x16 on PCIe switch
Up to 6x EDSFF E.1S NVMe SSDs

HGX Module

NVIDIA HGX A100 4-GPU with 4x NVLink connected SXM4 GPUs
Up to 8x 2.5” Hot Swap NVMe SSDs

RAID Support

SW RAID standard; Intel® Virtual RAID on CPU (VROC), HBA or HW RAID with flash cache options

I/O Expansion

Up to 4x PCIe Gen4 x16 adapters (2 front or 2-4 rear) and 1x PCIe Gen4 x16 OCP 3.0 mezz adapter (rear) depending on the configuration

Power and Cooling

Four N+N redundant hot-swap PSUs (up to 2400W Platinum)
Full ASHRAE A2 support with internal fans and Lenovo Neptune™ liquid-to-air hybrid cooling on HGX A100

Management

Lenovo XClarity Controller (XCC) and Lenovo Intelligent Computing Orchestration (LiCO)

OS Support