Does using an M1 in Clamshell reduce Performance?
February 1, 2023
I ran some benchmarks on the M1 to see if it suffered performance throttling by running in clamshell mode. I’ve always heard that laptops don’t run as well with their screens closed, and it seems like it makes sense from an airflow perspective.
Tools Used
Benchmark | Version |
---|---|
Geekbench | 5 |
Blender | 3.4.0 |
System specifications: M1 Pro (10-core), 16gb RAM, 512GB SSD.
I expected the results to be more exciting than they were. There are essentially no differences between lid open and closed on my system.
For my test methodology, I ran the tests with a couple of minutes between each one, and rebooted after I ran the open lid tests. I’m not going for scientific rigor, I ran each test once. If the results had shown different results I’d have dug deeper.
Benchmark Results
Benchmark | Lid Closed | Lid Open |
---|---|---|
Geekbench Single-Core | 1765 | 1766 |
Geekbench Multi-Core | 12361 | 12336 |
Geekbench Compute - Metal | 40949 | 41707 |
Blender Monster | 109.51 | 109.18 |
Blender Junkshop | 60.7 | 59.35 |
Blender Classroom | 47.96 | 47.72 |
Okay, so on the M1 it doesn’t matter. What about the notoriously throttled 2019 16” Macbook Pro with the i9? I ran the same tests on that system and was also surprised to see no difference. Either it truly doesn’t matter, or these benchmarks don’t tax the systems enough to get them to thermally throttle due to real or perceived airflow differences.
Benchmark Results - 2019 Macbook Pro (i9 CPU, upgraded GPU)
Benchmark | Lid Closed | Lid Open |
---|---|---|
Geekbench Single-Core | 1149 | 1145 |
Geekbench Multi-Core | 7292 | 7349 |
Geekbench Compute - Metal | 31597 | 34183 |
Blender Monster | 78.3 | 77.65 |
Blender Junkshop | 43.74 | 44.4 |
Blender Classroom | 31.36 | 31.3 |
The benchmarks did make the Intel Mac’s fans spin up loudly. The M1 was silent as ever.