Questions about nvidia-smi

I am using nvidia-smi for collecting CUDA application runtime statistics. The utility nvidia-smi gives the utilization of GPU core and memory like following:

==============NVSMI LOG==============

Timestamp                       : Mon Feb 21 17:43:56 2011

Driver Version                  : 260.19.26

GPU 0:

        Product Name            : GeForce 8800 GTX

        PCI Device/Vendor ID    : 19110de

        PCI Location ID         : 0:4:0

        Board Serial            : 211561763875

        Display                 : Connected

        Temperature             : 57 C

        Fan Speed               : 60%

        Utilization

            GPU                 : 47%

            Memory              : 1%

My question is how are the ‘Utilization of GPU’ and ‘Utilization of memory’ defined? For example, GPU core’s utilization is 47%. It means there are 47% of SMs active working? Or all the GPU cores are busy in 47% time while idle other 53% time? For memory, the utilization stands for the ratio between current bandwidth and max bandwidth, or the busy time ratio in last time unit?

Thanks for any suggestion.

Tim Murray explained the current nvidia-smi utilization counters in this post

Thank you for your response!

But based on my several little experiments, for some applications with same or similar bandwidth usage, they may have different memory utilization rate. Do you have any clue about it?

Really appreciate your help :thanks: