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
Computer Vision for Sports and Entertainment
Introduction Solution overview Lab architecture overview Scaling and sizing guidance
Scaling and sizing guidance
A40 vs A16 in the computer vision lab
Sizing the system
System tuning
Testing maximum streams per node
VxRail scaling guidance
Ipsotek VISuite application scaling
Validation Conclusions References

A40 vs A16 in the computer vision lab

Home > Workload Solutions > Computer Vision > White Papers > Computer Vision for Sports and Entertainment > A40 vs A16 in the computer vision lab

A40 vs A16 in the computer vision lab

  • The following table shows the comparison between the two GPU types available in our lab for computer vision validations:

    Table 1. A40 and A16 comparison
    Model Memory Power CUDA cores Encoders Decoders
    A40 48 GB 300 w 10,752 1 2
    A16 64 GB

    (16 GB x 4)

    		250 w 5,120 4 8
    Note: The A16 quad-board design becomes apparent after installing the GPU in the system. The operating system reports the existence of 4 x A2 GPUs instead of a single A16 device. An A40 card appears as a single device from the operating system tools.