PyMC3 Pickle Issue with GPU

[junpenglao]

@MisterRedactus

I observe that the run is significantly faster using the CPU. Is this normal behavior?

Gpu much slower than cpu

Running on GPU has a large overhead because the sampler are in Python (CPU) but the gradient and the logp function is under GPU (via theano). So unless you have very large matrix operations with expensive gradient computation that could leverage the GPU performance, using GPU is usually not worth.

@MisterRedactus

What is going on with the ‘Could not pickle model, sampling singlethreaded.’ output line? Is there something I can do to make the GPU run faster?

Not sure about this one - your model seems quite straightforward - did you try upgrading to the newest release PyMC3.5?