Time to talk about performance, because build quality issues aside, this is the star of the show. Again, we’re looking at an Intel Core i7-8550U quad-core, configured to use the standard 15W TDP, along with an Nvidia GeForce MX150 discrete GPU. We already know how the i7-8550U should perform, however its pairing with the MX150 does make it considerably more capable at graphics and compute workloads. Getting both these chips in a fanless chassis is a huge feat of engineering.
I’m not going to spend a lot of time discussing the pure CPU performance of the i7-8550U as I’ve already covered that a couple of times in the past. The key observation here is despite the use of just passive cooling, the Switch 7 has no problems matching other i7-8550U devices; it doesn’t thermal throttle significantly or underperform relative to the active-cooled, standard-performance Razer Blade Stealth.
Amazingly, it’s actually faster than the HP Spectre x360 in most tests, sometimes by significant amounts. I criticized the Spectre x360 in my review for having a weak cooler; maybe they should have ditched their mediocre fan for the fanless design Acer uses with the Switch 7, because it’s much more effective.
However there are some limits to the Switch 7’s thermal solution. The Switch 7 doesn’t feature as high CPU boost clocks, which leads to lower performance in very short workloads like MATLAB and Excel compared to the same 15W CPU in the Razer Blade Stealth. Clock speeds throughout a Cinebench R15 multi-threaded test show exactly this phenomenon: lower boost clocks but decent sustained clocks, as Acer has prioritized long-term performance rather than short, hot bursts.
Performance does steadily decline during lengthy tests or back-to-back short workloads though, but it can take upwards of 20 minutes for this to become an issue, which is longer than a lot of typical laptop tasks. The GPU is more affected than the CPU here too, as something like repeat 3DMark runs do show larger drops between consecutive runs than we’d normally see.
Passive cooling solutions are also limited by the environment they operate in. Our airconditioned benchmark lab is set to about 23C, but if you’re in a hotter summer environment, the cooler will suffer more than an active solution, and performance will be hit harder as a result.