Thank you Pascal and Fred, Fred I'll revise, check and post again. The file that's probably missing from the example is the one that loads the mnist dataset and could also be found in the theano tutorials.
I think Pascal has nailed the point correctly. Just to clarify I was not referring to the slow down speed between the new and old conv2d (that's bad on it's own if it happens but could be reasonable if we consider Fred's answer). In this thread I was just addressing the difference between the two when they utilize the cores in my laptop. 4 cores + openblas + openmp I use top/htop during the training and see the load in the different cores. During that time nothing else runs in my machine except from that particular training example. I even killed background daemons such as sshd in order to preserve better status of the cpu. Old conv2d shows all 4 cores 100% New conv2d shows minimal usage and fluctuations e.g. core1 = 40% core2 = 10% core3 = 60% core4 = 90% And all drop down to 0% and up again to different percentage values during training but definitely not to 100% as the old conv2d used to show. Cheers. On Tuesday, September 13, 2016 at 7:24:46 PM UTC+1, Pascal Lamblin wrote: > > On Sun, Sep 04, 2016, [email protected] <javascript:> wrote: > > Thank you nouiz! Great to know. Does that imply that the new conv2d > doesn't > > use full parallelism capabilities as the old one? > > BTW in both the experiments that I've run I've tested the new conv2d > > against the old with openblas and mkl-intel libraries. > > The results were the same the old one utilizes better the available > cores. > > Just to clarify: the cores being used 100% and the computation being > completed faster are two different things. > > The new conv2d code (using corrMM on the CPU, which uses BLAS) should be > much more efficient overall, and may _need_ less CPU usage, making the > CPU less busy, but it may still be faster. > > Did you time or profile the actual execution time? > > > The new one fluctuates a lot. But the down side is that > > the old conv2d doesn't support arguments for strides and padding besides > > the pre-defined ones. > > > > Thanks again. > > > > > > On Thursday, September 1, 2016 at 4:55:05 PM UTC+1, nouiz wrote: > > > > > > They use different implementation. Make sure Theano use a parallel > blas > > > library. The new conv2d use it for part of the parallelism. > > > > > > Le 25 août 2016 21:53, <[email protected] <javascript:>> a écrit > : > > > > > >> Hi everyone, > > >> > > >> I've recently come across some weird behaviour regarding the new > > >> theano.tensor.nnet.conv2d > > >> and the old > > >> theano.tensor.nnet.conv.conv2d > > >> convolution functions. > > >> > > >> I have 2 different models one uses the old the other the new conv2d > > >> method. > > >> The difference between the two is that the the model that uses the > new > > >> conv2d methods has more layers than the other one, plus that I've > > >> explicitly defined padding and stride. > > >> > > >> Other than that everything else is the same. Number of data, training > > >> algorithm, batchSize .... etc. pretty much the same. > > >> > > >> Once I execute them, the smaller model with the old conv2d method > > >> utilizes all the cores in my system ;) great. > > >> The bigger model with the new conv2d method doesn't, which is strange > > >> because in this case the bigger the model the more resources would > need. > > >> > > >> Are there any differences in the way the two conv2d methods utilize > > >> openmp? > > >> > > >> -- > > >> > > >> --- > > >> 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] <javascript:>. > > >> 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] <javascript:>. > > For more options, visit https://groups.google.com/d/optout. > > > -- > Pascal > -- --- 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.
