Overview

This is a setup guide aimed at people new to GCP and a general cloud workflow with GPUs.

Table of contents

---

GCP VM setup

login and basics

For this we are going to use GCP compute engine VMs specifically with GPUs, starting off here and activate the compute engine API.

You might have to input your payment information but you will most likely get around 300$ in credits which equates to roughly 300 hours of VM usage, so assuming this is your first time using GCP you shouldn't be too worried.

Once you set up everything you should be greeted with a screen similar to this.

It's a bit overwhelming but the two most important things for now are CREATE INSTANCE and the External IP ( which you use to connect to the VM ).

quota check

Before we the instance - since we want to use GPUs - we have to check if our GPU quota (number of GPUs allowed) is >0, since for some reason an issue is that the default is sometimes 0. To check this

1. Click on the search bar above CREATE INSTANCE
2. Look for All quotas
3. In the field Enter property name or value look for Name:GPUs (all regions)
4. Click the little checkbox next to the name
5. Click EDIT QUOTAS (top right of the screen)
6. Type 1 (if you want multiple GPUs then type more)
7. Hit SUBMIT REQUEST

Visually this looks something like this

creating an instance

Click CREATE INSTANCE

Machine configuration

1. For Name type whatever you want (I usually just leave the default)
2. For Region I'd recommend one of the US ones since they tend to have the most availability for GPUs
3. For Zone it doesn't matter too much (you can leave the default)
4. For Machine family select GPUs
5. For GPU type select NVIDIA T4 (this is the cheapest option)
6. For Number of GPUs select 1 (you can select more if you want but this is the cheapest option)
7. For Machine type you can either select a preset or customize it, I'd recommend at least 4 cores and 15GB of RAM

Boot disk

1. Click CHANGE
2. For Operating system select Deep Learning on Linux
3. For Version select the latest one with CUDA
4. For Size (GB) I'd recommend at least 100GB

Identity and API access

1. For Service account select Compute Engine default service account
2. For Access scopes select Allow full access to all Cloud APIs

Firewall

1. Check Allow HTTP traffic
2. Check Allow HTTPS traffic

Then click CREATE

connecting to your VM

Once your VM is created you can connect to it by clicking the SSH button next to the VM name. This will open a new window with a terminal connected to your VM.

Alternatively you can use the gcloud CLI to connect to your VM. First install the CLI by following the instructions here. Then run the following command:

gcloud compute ssh --zone "YOUR_ZONE" "YOUR_VM_NAME" --project "YOUR_PROJECT_ID"

setting up VSCode remote

If you want to use VSCode to connect to your VM you can use the Remote - SSH extension. First install the extension by clicking here. Then follow these steps:

1. Press Ctrl+Shift+P and type Remote-SSH: Connect to Host...
2. Click + Add New SSH Host...
3. Type YOUR_USERNAME@EXTERNAL_IP where YOUR_USERNAME is your username on the VM and EXTERNAL_IP is the external IP of your VM
4. Select the SSH config file you want to save the host to
5. Click Connect

You might need to set up SSH keys first. You can do this by running the following command:

gcloud compute config-ssh

This will add the necessary entries to your SSH config file.