CPU throttling

Default settings for CPU throttle on emulated mobile devices, plus how to set your own custom CPU throttle

Our test agents run on AWS EC2 c5.large instances, which provide plenty of CPU horsepower and are a good approximation for everyday desktop browsers.

Mobile devices are generally much slower than desktop, especially in developing countries where the next billion people will come online. So for our emulated mobile devices we throttle the CPU to better approximate the CPU available on actual devices.

Our CPU throttle values are based on Geekbench scores for iOS and Android devices. We used a MacBook (mid-2017) with a Geekbench single core score of 3,607 as a baseline.

Default CPU throttle

Higher values are slower. A CPU throttle of 0 means no throttling, and 2 means half the CPU.

DeviceCPU Slowdown Factor
Apple iPad1.8
Apple iPhone 53.0
Apple iPhone 62.6
Apple iPhone 71.2
Apple iPhone 80
Apple iPhone X0
Google Nexus 62.9
Google Nexus 73.1
Google Nexus 103
Google Pixel2.4
Google Pixel 22.2
Samsung Galaxy S III3.2
Samsung Galaxy S43.1
Samsung Galaxy S72.3
Samsung Galaxy S82.2

Set your own CPU throttle

If you want to change your CPU throttle or emulate any other older slower devices, you can create a custom browser in your Settings.

  1. Go to your Settings page and hit the green Edit Browsers button.

Edit Browsers

  1. Scroll down to the Custom Browser Profiles section and click the link to add a custom profile.

Add Custom Browser Link

  1. The Add a Custom Browser page gives you a bunch of customization options, including Throttle CPU. In the example below, we've created a 2X CPU slowdown for Safari over 3G. We've named this profile Safari 3G throttled.

Adding a Custom Browser

  1. When you've created your custom browser profile, hit the green Add button. This will take you back to your main Settings page, where you'll see your new custom profile grouped with your other browsers. This browser profile is now available for you to use in your Sites settings.

Choosing a Custom Browser