Your Browser is Out of Date

Nytro.ai uses technology that works best in other browsers.
For a full experience use one of the browsers below
Implementation Guide—Virtualizing GPUs for AI with VMware and NVIDIA Based on Dell Infrastrucutre
Introduction Solution Components
Hardware components
Software components
Deploying the Solution Verification Summary References

Hardware components

Hardware components

  • The following table lists the hardware components that this validated design supports: 

    Table 1.      Hardware components

    Category

    Hardware components

    Compute server

    PowerEdge R750xa, R750, R740, R740xd, or R7525 servers

    VxRail V670 or VxRail V570F appliances

    Number of nodes in a cluster (for vSphere with Tanzu and vSAN)

    Minimum of 3 nodes

    We recommend 4 ESXi hosts when using vSAN for resiliency during updating and upgrading.

    Network switch

    2 x Dell S5248F-ON 

    For multinode training: 2 x Dell S5232F-ON or 2 x Mellanox SN3700

    OOB switch

    1 x Dell N3248TE-ON

    The following table describes the PowerEdge server hardware configuration:

    Table 2.      PowerEdge Server configuration

    Category

    Hardware components

    Processor

    Intel Xeon Platinum or Gold processors for PowerEdge R750, PowerEdge R750xa, and VxRail 670F 

    AMD EPYC processor for PowerEdge R7525

    Memory

    512 GB memory (recommended)

    GPUs

    NVIDIA A100 or A30 GPUs

    Network adapters

    ConnectX-5 DX 25 GbE

    For multinode training: ConnextX-6 DX Dual Port 100 GbE

    Storage adapters

    BOSS controller card + with 2 M.2 Sticks 480 GB (RAID 1) – For ESXi

    HBA 330 – For vSAN 

    Internal storage

    Cache tier: 2 x 800 GB SSD SAS Write Intensive 

    Capacity tier: 12 x 960 GB SSD SAS Read Intensive

    OOB Management

    iDRAC Enterprise