Overview

The following quick start guide offers a general pipeline for the steps of running AutoLFADS or RADICaL over Google Cloud Platform, and each step can be expanded to offer abbreviated instructions including code and links to detailed sections. Brand new users are recommended to use the more detailed versions of instructions, beginning with First Time Set-up.

(First run begin here)

STEP 1 - Set Up Project 10 Minutes

1.1: Create project, set up billing

1.2: Request additional quota (if needed)

Up to 48 Hour Wait

STEP 2 - Set-up Infrastructure 45 Minutes

2.1: Clone SNEL repository into cloud shell

(Code in red border should be executed in cloud shell)

git clone -b GCP https://github.com/snel-repo/autolfads.git

2.2: Create server machine

 sh server_set_up.sh tutserver us-central1-c

Up to 10 Minute Wait

2.3: Create client machines

 sh machine_setup.sh tutcliente 4 us-east1-c

 sh machine_setup.sh tutclientc 4 us-central1-c 

Up to 20 Minute Wait

2.4: Check if Docker is finished installing sucessfully

 sh check.sh pbtclient 

2.5: Create bucket with folders data and run

(New users to already existing project begin here)

STEP 3 - Add User 5 Minutes

(This step can only be done after Docker is finished installing)

3.1: Add user to Docker group

 sh add_docker_user.sh pbtclient

3.2: Clone SNEL repository on server

(Code with blue border should be executed in server shell)

git clone -b GCP https://github.com/snel-repo/autolfads.git

(Additional run, new dataset begin here)

STEP 4 - Upload Data 5 Minutes

4.1: Upload data with prefix lfads

(Additional run, same dataset begin here)

STEP 5 - Start Run 10 Minutes

5.1: Link pbt_script_multiVM.py to your data and
edit parameters

5.2: Begin AutoLFADS run in tmux

 tmux

 python2 pbt_script_multiVM.py

Up to 2 Hour Wait

STEP 6 - Analyze Output 30 Minutes

(You should stop all machines now)

6.1: Download output zip file from bucket

6.2: Run pbt_plot.m