Caffe on Jetson TX1

Hi,

   I just installed Caffe on my Jetson TX1. I followed the steps in this link.

http://petewarden.com/2014/10/25/how-to-run-the-caffe-deep-learning-vision-library-on-nvidias-jetson-mobile-gpu-board/

   I tested the mnist training on GPU. It takes 14 mins to finish. It seems to much slower than Intel i7 CPU result (<5min). Any one beat or have similar performance as CPU? 

http://caffe.berkeleyvision.org/gathered/examples/mnist.html

Thanks

Hmm. Can you try running this perf maximization script? https://devtalk.nvidia.com/default/topic/894945/jetson-tx1/jetson-tx1/post/4734485/#4734485