Google Compute Engine Machine Type Comparison

This webapp helps to find the optimal Google Compute Engine (GCE) machine type or instance in the many Google Cloud Platform (GCP) regions. A lot of information has been collected from various Google Cloud websites and different sources.

I was tired of switching between different Google documentations to get all the information to find the best Google Compute Engine machine for my needs or those of my clients. I have therefore merged all important information into one place. Hope it helps you too.

Machines

Here you can find an overview of all 159 different Google Compute Engine machine types. You can also find the average costs per GCE machine type across all Google Cloud regions in this overview.

You have the choice between 92 machine types with Intel CPU, 77 CPU machine types with AMD CPU and 7 CPU machine types with Ampere Altra Arm-based processors. 86 machine types are certified for SAP application and 17 machine types are certified for SAP HANA on Google Cloud.

Disks

Here you can find an overview of 5 different disk types for your instances.

Images

Here you can find an overview of many preconfigured public images that have compatible Linux, UNIX or Windows operating systems.

Regions

Here you can find an overview of all 35 Google Cloud regions and the corresponding zones. A total of 10 million public IPv4 addresses are assigned to all Google Cloud Platform regions. There is also a map view with all Google Cloud Platform regions. You can also find the available Google Compute Engine machine types per GCE machine series (CPU type and CPU plattform) in this overview. A cost overview of the persistent disk types in the regions can also be found in this section.

Of the 35 regions, 10 regions have a low carbon impact. This means that this regions have a Google CFE% of at least 75%, or, if CFE% information is not available, a grid carbon intensity of maximum 200 gCO2eq/kWh.

In total, you have the freedom to choose between 3664 different combinations of Google Compute Engine machine types and Google Cloud regions. For each combination the cost per hour, month with sustained use discounts (SUD), month with 1 year commitment (CUD) and price per month with 3 year commitment (CUD) was calculated. The cost of paid "premium" operating system licenses, such as SUSE Linux Enterprise Server (SLES), Red Hat Enterprise Linux (RHEL) and Microsoft Windows Server, was also calculated.

Instance Picker

With the Instance Picker you can quickly and easily compare all possible Google Compute Engine machine types and instances in all Google Cloud Platform regions.

You can set filters for each column and sort them. Filters can be combined.

A good first filter is to set the preferred region. You can also filter by parts of the region. Example europe to filter all machine types in Europe.

Filter

You can filter and sort all fields in the Instance Picker.

Filter on region europe-west4 and CPU base clock frequency greater than 3 GHz: Screenshot: Filter

Filter on region europe-west4 and Memory greater than 30 GB but also less than 60 GB: Screenshot: Filter combined

Sort and more

Click header title to sort column: Screenshot: Sort colums

Click group header title icon to show more colums: Screenshot: Show more colums

Keyboard shortcuts

You can navigate the Instance Picker with your keyboard.

Press Ctrl + / to focus the vCPU filter.

Use the arrow keys ( ) to move focus up, down, left and right. Tab will move the focus horizontally until the last cell and then move on to the next row. Shift + Tab will move the focus horizontal backward until the first cell and then move back to the previous row.

Pressing the Space key on a cell will select the cells row, or deselect the row if already selected.

With the following keyboard shortcuts you can copy selected row from the table in CSV format to your clipboard:

The grid header also supports full keyboard navigation:

Press ? to see the help with the keyboard shortcuts.

Download

You can download a CSV file (1.74 MB) and SQL export (0.17 MB, gzip compressed) with all machine types, disk types, operating system images, Google Cloud regions and zones. You can import and edit the CSV file to your favorite spreadsheet program (MS Excel, LibreOffice or Google Sheets). You can import the SQL dump into your favorite database (like SQLite).

Imported CSV file in LibreOffice spreadsheet program: Screenshot: LibreOffice

Imported SQL export in DB Browser for SQLite: Screenshot: DB Browser for SQLite

With the SQL export you can also do great queries using the command line (CLI).

Command line:

# Download SQL export
curl -O https://gcloud-compute.com/machine-types-regions.sql.gz
# Decompress SQL export
gzip -d machine-types-regions.sql.gz
# Import SQL dump
sqlite3 gce.db < machine-types-regions.sql
# SQL query
sqlite3 gce.db 'SELECT name, vCpus, hour FROM instances WHERE vCpus >= 8 AND region LIKE "europe-west4" ORDER BY hour ASC LIMIT 5'

Example SQL output:

e2-highcpu-8|8.0|0.21785232
n2d-highcpu-8|8.0|0.274688
e2-standard-8|8.0|0.29508816
n1-highcpu-8|8.0|0.3119968
n2-highcpu-8|8.0|0.315728

Resources of the Information

The information was obtained from the following different Google websites:

Open Source

Everything you see here is published as open source software and licensed under the Apache License (version 2.0). You can find the source code on GitHub. Any improvements and pull requests are welcome.