Hello everyone, I’m doing deep learning research, and I have already installed 2 titan rtx GPU cards and 1 nvlink bridge. I could set batch size as 8 before I installed nvlink and I expected that I could set batch size as 16 after I install nvlink bridge, but it turns out that it doesn’t work…Please help me solve this issue, thank you in advance!
here is the information of my machine:
OS: ubuntu16.04
Driver: 410.79
GPU card: titan rtx x 2
Container: nvcr.io/nvidia/tensorflow:18.09-py3