How to Launch a Ursa Major Research Workstation with and without GPUs

How to Launch a Ursa Major Research Workstation with and without GPUs

Go to the Google Cloud Console: To start, go to the Google Cloud Console, which is the web-based portal for managing your Google Cloud resources. You will need a Google account to access the console.
On the “Select a Project” drop-down in the top navigation bar select the project you want to launch in.
Navigate to Compute Engine: From the console home page, click on the hamburger menu in the top left corner and select “Compute Engine” from the list of services.
Launch a New Instance: To launch a new instance, click the “Create an instance” button on the Compute Engine dashboard.
Choose a Name and Zone: Give your instance a name and choose the zone in which you want it to be deployed.
Select a Machine Type: The next step is to choose a machine type. If you want to launch a VM with GPUs, select a machine type that includes one or more GPUs. If you do not need GPUs, select a machine type without them.
Choose a Boot Disk: You can choose the boot disk for your instance by clicking on the “Change” button under “Boot disk.”
Select an Operating System: You can choose from a variety of operating systems, including Windows and Linux. Some popular Linux distributions include Ubuntu, Debian, and CentOS.
Configure Additional Disk Space: If you need additional disk space for your instance, you can add one or more persistent disks by clicking on the “Add New” button under “Disks.”
Configure Networking: You can configure the network settings for your instance, such as firewall rules and network tags, by clicking on the “Edit” button under “Networking.”
Launch the Instance: When you have finished configuring your instance, click the “Create” button to launch it.
Connect to the Instance: Once the instance is launched, you can connect to it by clicking on the “SSH” button in the Compute Engine dashboard.

Using gcloud to Launch a Ursa Major Research Workstation:

Open a terminal or command line window.

Type in the following command to log in to your Google Cloud account:
```
gcloud auth login
```
Type in the following command to set the project you want to launch the instance in:
```
gcloud config set project [PROJECT_ID]
```

Type in the following command to launch the instance:

gcloud compute instances create [INSTANCE_NAME] --zone [ZONE] --machine-type [CPU_MEMORY] --image-family [OS] --image-project [OS_PROJECT]

(Optional) If you want to use a GPU, add the following flag to the command:
```
--accelerator type=[GPU_TYPE],count=[GPU_COUNT]
```
(Optional) If you want to add storage to your instance, add the following flag to the command:
```
--boot-disk-size=[STORAGE_SIZE] --boot-disk-type=[STORAGE_TYPE]
```
Wait for the instance to launch, which should take a few minutes.

Pointers:

When choosing the number of vCPUs and amount of memory for your instance, consider the workload you will be running on the instance. If you need more power, choose a larger instance size. If you don’t need as much power, choose a smaller instance size.
If you are using a GPU, consider the type of GPU you need based on your workload.
When choosing the operating system, consider the type of applications and services you will be running on the instance.
When choosing the storage options, consider the amount of data you will be storing and the performance requirements for your instance.
Make sure to choose a zone that is closest to your users or customers to ensure fast and low-latency access to your instance