Wednesday, July 18, 2018

Is the MacBook Pro 2018 subject to excessive thermal throttling?

Video claims it is

Apple’s top-of-the-range 15-inch MacBook Pro 2018 comes with a powerful 6-core Intel Core i9 processor, but a YouTube reviewer has suggested that the device excessively throttles the CPU’s performance when it gets too hot.

Thermal throttling is a rather common practice – it limits the performance of components when they get too hot, protecting the device from damage. The harder a component like a processor works, the more heat it produces.

YouTube reviewer Dave Lee, has posted a video (which you can view below), in which he claims the MacBook Pro 2015 Core i9 edition heats up very quickly, causing the throttling to kick in and lower the clock speed of the processor. The lower the clock speed, the slower the processor performs.

Slow and steady

As we mentioned, thermal throttling is a common practice, but what worried Lee was that the throttling kicked in so soon – he claims after a few seconds of video editing in Adobe Premiere – and that the throttling slows the Core i9 chip to 2.2GHz, which is below the 2.9GHz base clock the processor should normally work at.

This means the performance upgrade you get from the more expensive Core i9 version is limited compared to an Intel Core i7 version, which Lee claims can manage average clock speeds of 3.1GHz with the Turbo Boost feature.

Lee also runs the i9 MacBook Pro in a freezer, and in the colder climate, the notebook performs faster, with the time it takes to render a video dropping from 40 minutes to 27 minutes. As Lee points out, this suggests that the thermal solution inside the MacBook Pro, which is designed to keep its components cool despite the thin and light design of the notebook, isn’t doing a terribly good job.

Of course, this is just one person’s findings, using a single app and benchmark test. We’ll be reviewing the 15-inch MacBook Pro 2018 ourselves soon, and we'll keep a close eye on its temperatures and performance.

READ MORE TechRadar: