Any way of using multiple CPU cores?

[jmurth1234]

Hi,

I'd like to be able to run this on my server, however the model seems to only use one core, and as I don't have a GPU it makes the training extremely slow.

Is there a way of using multiple cores?

[vasyarv]

Maybe you need to turn it with OpenMP http://deeplearning.net/software/theano/tutorial/multi_cores.html

You can switch your lineal computation lib to OpenBlas It is described here https://habrahabr.ru/post/274331/ (sorry , article is in russian)

[drscotthawley]

Since my Mac laptop has an AMD GPU and I haven't figured out how to OpenCL working (any tips?), just adding "OMP_NUM_THREADS=" got me an instant factor of 2 in speedup without doing any other work.

[piteight]

mattpearson made version compatible with newest keras, please look at https://github.com/mattpearson/GRUV I have noticed that in that version, it switched to tensorflow beckend and used all CPU cores.