Multi GPU not being utilized

[jalexhurt]

System information

• What is the top-level directory of the model you are using: object-detection
• Have I written custom code (as opposed to using a stock example script provided in TensorFlow): No
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): nvidia-docker with nvidia/cuda 10 and cuDNN 7
• TensorFlow installed from (source or binary): binary
• TensorFlow version (use command below): 1.14.0
• Bazel version (if compiling from source): N/A
• CUDA/cuDNN version: 7
• GPU model and memory: GTX 1080 8GB
• Exact command to reproduce: ?

Describe the problem

When training with multiple GPUs, GPU 0 is being utilized while utilization of all other GPUs sit at 0%. I believe the desired behavior is that more GPUs would offer some performance advantage over a single GPU

Source code / logs

models/research/object-detection/legacy/train.py

[jalexhurt]

Any updates on this? I have more hardware available that sits idle when trying to train.

[kun-dao]

I also met same problem，I have four TITAN XP cards , but only one has been used. did you solve it ? can you tell me about it ?