# Bandwidth test result for A800 does not meet expectations

**URL:** <https://forums.developer.nvidia.com/t/bandwidth-test-result-for-a800-does-not-meet-expectations/263948>\
**Category:** General\
**Tags:** a100\
**Created:** [August 22, 2023, 6:19am UTC](https://forums.developer.nvidia.com/t/bandwidth-test-result-for-a800-does-not-meet-expectations/263948 "2023-08-22T06:19:11Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![peterzhang](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@peterzhang](https://forums.developer.nvidia.com/u/peterzhang)\
**Post date:** [August 22, 2023, 6:19am UTC](https://forums.developer.nvidia.com/t/bandwidth-test-result-for-a800-does-not-meet-expectations/263948/1 "2023-08-22T06:19:11Z")

</div>

We have physical computer node with A800 GPU\*8, and its output of “nvidia-smi topo -m” is shown below:

 ![image](https://global.discourse-cdn.com/nvidia/original/4X/f/9/e/f9e5ba5a46a47b8f519c7a1f5bfc3c387728dc6c.jpeg)

However, when a virtual machine with A800 \* 4 GPU created on the node, and we ran benchmark test on it. As shown in the above figure, the bidirection bandwidth was closed to 100GB/s and the single bandwidth was closed to 50GB/s.

It makes me confused that in 4GPU virtual machine, each GPU has 4 nvlinks. The nvlinke supports 50 GB/s bidirectional bandwidth, so the read&write bidirectional bandwidth on single GPU should be closed to 50GB/s \* 4 = 200GB/s and single direction bandwidth should be closed to 25GB/s \* 4 = 100GB/s.

Is there something wrong in my interpretation of aggregate bandwidth in A800, or in my understanding of NVLink?