pluskid · pluskid · Nov 1, 2016 · Oct 28, 2016 · Oct 28, 2016 · Oct 29, 2016
diff --git a/.gitignore b/.gitignore
@@ -23,3 +23,5 @@ src/cuda/kernels/*
 *~
 \#*\#
 .\#*
+
+*.DS*
diff --git a/docs/index.rst b/docs/index.rst
@@ -16,6 +16,7 @@ Tutorials
    tutorial/cifar10
    tutorial/ijulia-imagenet
    tutorial/mnist-sDA
+   tutorial/cloud
 
 ############
 User's Guide

diff --git a/docs/tutorial/cloud.rst b/docs/tutorial/cloud.rst
diff --git a/docs/tutorial/images/bitfusion.png b/docs/tutorial/images/bitfusion.png
diff --git a/docs/tutorial/images/capture.png b/docs/tutorial/images/capture.png
diff --git a/docs/tutorial/images/continue.png b/docs/tutorial/images/continue.png
diff --git a/docs/tutorial/images/create_case.png b/docs/tutorial/images/create_case.png
diff --git a/docs/tutorial/images/gpu_results.png b/docs/tutorial/images/gpu_results.png
diff --git a/docs/tutorial/images/gpu_test_output.png b/docs/tutorial/images/gpu_test_output.png
diff --git a/docs/tutorial/images/smile.png b/docs/tutorial/images/smile.png
diff --git a/docs/tutorial/images/support_request_border.png b/docs/tutorial/images/support_request_border.png
diff --git a/examples/cloud/README.md b/examples/cloud/README.md
@@ -0,0 +1,25 @@
+# Mocha in the Cloud
+
+The real power of developing deep learning networks is only realized
+when you can test your ideas and run your models on powerful compute
+platforms that complete training in a fraction of the time it took only
+a few years ago. Today, state-of-the-art machine learning algorithms
+routinely run on cloud based compute that offers access to cutting edge
+GPU technology for a tiny cost compared to buying the GPU hardware and
+running it in your personal or company servers.
+
+Amazon Web Services (AWS) is one of the most popular cloud based
+services that provide access to such powerful computers. As Mocha and
+Julia mature I'm sure that a pre-configured Amazon Machine Image (AMI)
+will emerge to run Mocha in the cloud. Now, in October 2016, such an AMI
+does not exist, but even when a pre-configured image for Mocha does
+become available I highly recommend following through this tutorial at
+least once so you understand how cloud resources are provisioned and
+configured to suport Deep Learning.
+
+We are going to show you how to take the CIFAR-10 example and get it
+running in the cloud. Along the way you will learn how to interact with
+AWS and get a broader understanding of cloud architectures in general.
+
+The example includes several images and code snapshots.  So it is best
+to follow along with the [documentation](http://mochajl.readthedocs.io/en/latest/tutorial/cloud.html). 
diff --git a/examples/cloud/cifar10.jl b/examples/cloud/cifar10.jl
@@ -0,0 +1 @@
+../cifar10/cifar10.jl
diff --git a/examples/cloud/convert.jl b/examples/cloud/convert.jl
@@ -0,0 +1 @@
+../cifar10/convert.jl
diff --git a/examples/cloud/get-cifar10.sh b/examples/cloud/get-cifar10.sh
@@ -0,0 +1 @@
+../cifar10/get-cifar10.sh
diff --git a/examples/cloud/yum_draft.rst b/examples/cloud/yum_draft.rst
@@ -0,0 +1,32 @@
+This script for yum based linux is a draft and has not been thoroughly tested
+=============================================================================
+
+Build from scratch in yum based AMIs
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+.. code-block:: bash
+
+    #! /bin/bash
+    # Install build environment
+    # The sed instruction allows the Extended Packages For
+    # Enterprise Linux (epel) repository to get added to the yum package
+    # manager. sed -i option edits in place.
+    echo "*****************Setting up the Build Environment******************"
+    sudo sed -i 's/enabled=0/enabled=1/' /etc/yum.repos.d/epel.repo
+    sudo yum -y update
+    sudo yum -y upgrade
+    # I need to check which of these are needed
+    sudo yum -y install git gcc-gfortran clang m4 patch ncurses-devel python-devel
+
+    #Set up Julia
+    echo "*****************Cloning Julia*************************************"
+    git clone https://github.com/JuliaLang/julia.git
+    cd julia
+    git checkout v0.4.7
+
+    #Determine the number of CPUs to build on
+    NUM_CPUS=$(lscpu | awk '/^CPU\(s\):/ {print $2}')
+
+    echo "*****************Making Julia on $NUM_CPUS CPUs***************************"
+    #Takes 30 minutes on a 4CPU p2.xlarge AWS instance
+    time make -j $NUM_CPUS
diff --git a/examples/mnist/mnist_learning_curve.jl b/examples/mnist/mnist_learning_curve.jl
@@ -1,3 +1,5 @@
+#!/usr/local/bin/julia
+
 using PyPlot, JLD
 # Plot the learning curve for the MNIST tutorial to verify
 # the solver is progressing toward convergence
-Original file line number
+Diff line change
@@ Expand Up / @@ -23,3 +23,5 @@ src/cuda/kernels/* @@
     *~
     \#*\#
     .\#*
+    *.DS*