The new back-end do compilation during the first call to the Theano function. Can you redo your timing while ignoring the first call to the Theano function?
If that don't fix the problem, can you enable Theano profiler and send us both profile? Use those Theano flags: profile=True profiling.ignore_first_call=True To know what the new back-end provide: https://github.com/Theano/Theano/wiki/Converting-to-the-new-gpu-back-end%28gpuarray%29 I don't expect a slow down in the new back-end (except for the first call). If there is one, we should fix it. Frédéric On Fri, Feb 17, 2017 at 7:53 AM Adam Becker <[email protected]> wrote: > I noticed this problem late 2016 while investigating scan overhead. Check > here > <https://github.com/Theano/libgpuarray/issues/292#issuecomment-262696507>. > The comment also have a link to a script help reproducing the problem. > > > On Friday, February 17, 2017 at 7:43:39 PM UTC+8, Ozan Çağlayan wrote: > > Hi, > > With the current Theano HEAD + libgpuarray, I launched two RNN-based MT > systems with the old backend and the new backend and apparently the new > backend is a little bit slower than the old one. > > This is on CUDA 7.5, CUDNN5.1 and Tesla K40 GPU, batchsize=64: > old backend: 166ms / batch > new backend: 186ms / batch > > On a moderate 4M sample dataset, this would bring an overhead of ~20 > minutes per epoch. > > Is this expected? If yes, why would I prefer the new backend? > > Thanks. > > -- > Ozan Çağlayan > Research Assistant > Galatasaray University - Computer Engineering Dept. > http://www.ozancaglayan.com > > #HayırdaHayırVar > > -- > > --- > You received this message because you are subscribed to the Google Groups > "theano-users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- --- You received this message because you are subscribed to the Google Groups "theano-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
