Training ImageNet with 2 GPUs

[caffecuda]

Hello,
I have two identical GPUs in my machine but when training ImageNet it seems only one of them is used, is this normal? If not how should I change the configurations to use both GPUs? Does using one or two GPU(s) produce identical/similar nets? Thanks!

[anguyen8]

I seem to be having the same issue.. only 1 GPU is used while 2 are available. I greatly appreciate some advice on how to configure caffe to use 2 GPUs.

[sguada]

Currently Caffe uses one GPU at the time, but we working in MultiGPU
training.

On Sunday, July 6, 2014, Anh Nguyen notifications@github.com wrote:

I seem to be having the same issue.. only 1 GPU is used while 2 are
available. I greatly appreciate some advice on how to configure caffe to
use 2 GPUs.

—
Reply to this email directly or view it on GitHub
#630 (comment).

Sergio

[kloudkl]

NVBLAS may provide the convenience.
#547 (comment)

[anguyen8]

@sguada : Thank you for your confirmation.
@kloudkl 👍 Oh nice. Thanks for the pointer, I will definitely give it a try in a near future.

[shelhamer]

Duplicate of #653.