You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2017/11/29 03:27:27 UTC
[GitHub] fanlu opened a new issue #8861: The first GPU cost more memory than the others in version 0.12.1
fanlu opened a new issue #8861: The first GPU cost more memory than the others in version 0.12.1
URL: https://github.com/apache/incubator-mxnet/issues/8861
I have 4 P40 GPU card
First I use batch_size 128 when I training deepspeech.mxnet and mxnet version is 0.10, the gpu memory cost per gpu are same.
But When I upgrade mxnet version to 0.12.1, The mxnet will throw an exception CUDA out of memory, I reduce the batch_size to 120, And I found the first gpu cost more memory than others,What is the problem?
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 21515 C python 20747MiB |
| 1 21515 C python 17149MiB |
| 2 21515 C python 18221MiB |
| 3 21515 C python 18385MiB |
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
With regards,
Apache Git Services