Setup AWS GPU

1. Create an AWS account

Visit aws.amazon.com and click on the "Create an AWS Account" button.

2. Request limit increase

2. View Your Current Limit

View your EC2 Service Limit report at: https://console.aws.amazon.com/ec2/v2/home?#Limits

Find your "Current Limit" for the p2.xlarge instance type. By default, AWS sets a limit of 0 on the number of p2.xlarge instances a user can run, which effectively prevents you from launching this instance.

Note: Not every AWS region supports GPU instances. At the time of writing, only US West (Oregon), US East (Virginia), and EU (Ireland) offer them. You can change your region under account settings if your default region does not support GPU instances.

If your limit of p2.xlarge instances is 0, you'll need to increase the limit before you can launch an instance. In this section, you'll go through the process of submit a request to increase your limit.

• From the EC2 Service Limits page, click on “Request limit increase” next to “p2.xlarge”.
  • You will not be charged for requesting a limit increase.
• Complete all the fields.
  • For the “Region” field, select the region closest to you.
  • For the “New limit value”, enter a value of 1 (or more, if you wish).
  • For the “Use Case Description”, you can simply state: “Amazon MLU Intro to CV”

Note: If you have never launched an instance of any type on AWS, you might receive an email from AWS Support asking you to initialize your account by creating an instance before they approve the limit increase.

Wait for Approval

You must wait until AWS approves your Limit Increase Request. AWS typically approves these requests within 48 hours.

3. Launch an Instance

Once AWS approves your GPU Limit Increase Request, you can start the process of launching your instance.

• Goto the EC2 Management Console
• Click on the “Launch Instance” button.
• Click "AWS Marketplace"
• Search for "Deep Learning AMI with Source Code (CUDA 8, Ubuntu)" AMI.
• Click the "Select" button for the "Deep Learning AMI CUDA 8 Ubuntu Version" AMI.

AMI (Amazon Machine Images) contains all the environment files and drivers for you to train on a GPU. It has cuDNN, and many other packages required for this course. Any additional packages required for specific projects will be detailed in the appropriate project instructions.

Instance Type

• Filter the instance list to only show “GPU instances”
• Select the p2.xlarge instance type
• Click on the “Review and Launch” button'

Security Groups

You'll be running a few Jupyter notebooks in this course, which default to port "8888". To access this port, you'll need to open this up on AWS by editing the security group.

• Click on "Edit security groups"
• Setup a security group
  • Select "Create a new security group"
  • Set the "Security group name" to "Jupyter"
  • Click "Add Rule"
  • Set a "Custom TCP Rule"
  • Set the "Port Range" to "8888"
  • Select "Anywhere" as the "Source"
• Click "Review and Launch"

Launch

• Click on the “Launch” button to launch your GPU instance.
  • AWS will ask if you’d like to specify an authentication key pair. You'll need to do so in order to access your instance, so you select “Create a new key pair” and click the "Download Key Pair" button. This will download a .pem file, which you'll need to be able to access your instance. Move the .pem file to a secure and easily remembered location on your computer; you'll need to access your instance through the location you select.
• After the .pem file has been downloaded, click the “Launch Instances” button.
• Click the “View Instances” button to go to the EC2 Management Console and watch your instance boot.

Be Careful!

From this point on, AWS will charge you for running this EC2 instance. You can find the details on the EC2 On-Demand Pricing page.

Most importantly, remember to “stop” (i.e. shutdown) your instances when you are not using them. Otherwise, your instances might run for a day, week, month, or longer without you remembering, and you’ll wind up with a large bill!

AWS charges primarily for running instances, so most of the charges will cease once you stop the instance. However, there are smaller storage charges that continue to accrue until you “terminate” (i.e. delete) the instance.

4. Remotely connecting your EC2 instance

Make a note of the "IPv4 Public IP" address (in the format of “X.X.X.X”) on the EC2 Dashboard.

• Go to your terminal
• Navigate to the location where you stored your .pem file
  • For example, if you put your .pem file on your Desktop, cd ~/Desktop/ will move you to the correct directory.
• Type ssh -i YourKeyName.pem user@X.X.X.X
  • user could be ubuntu@ or ec2-user@
  • X.X.X.X is the IPv4 Public IP found in AWS.
  • YourKeyName.pem is the name of your .pem file

Notes: If you see a "bad permissions" or "permission denied" error message regarding your PEM file, try executing chmod 400 path/to/YourKeyName.pem, then run the ssh command again. See this page for more details to solve this issue.

Run the notebook

• jupyter notebook --ip=0.0.0.0 --no-browser
• From your local machine: You will need the token generated by your jupyter notebook to access it. On your instance terminal, there will be the following line: Copy/paste this URL into your browser when you connect for the first time, to login with a token:. Copy everything starting with the :8888/?token=.
• Access the Jupyter notebook index from your web browser by visiting: X.X.X.X:8888/?token=...(where X.X.X.X is the IP address of your EC2 instance and everything starting with :8888/?token= is what you just copied)

If you run though the TIME OUT issue: "This site can’t be reached"

Use the following commands to relay the connection and access the remote server locally.

ssh -i "YourKeyName.pem" -o 'IdentitiesOnly yes' -L 8157:127.0.0.1:8888 ubuntu@ec2-107-23-127-246.compute-1.amazonaws.com

You can access the Jupyter notebook server at:

http://127.0.0.1:8157

You can try following the wiki below:

https://docs.aws.amazon.com/dlami/latest/devguide/setup-jupyter-configure-client-mac.html