Hey there,
anyone interested in running some benchmarks to compare heat development?
I've recently replaced the thermal paste on mine and would like to know if it made a difference.
While I have yet to find something that'd push the switch in the same way some of it's games would in terms of temperature, comparing even low temp development should at least make some things more objective than just 'well it feels colder in korok forest I think'.
I've been using Antutu/Antutu 3d, 3D Mark and CPU Prime almost back to back.
Performance Settings were at original standard (middle) for these tests.
I was charging with unofficial charger cause battery was pretty low after installation, undocked, lying on a table (joycons connected), so there was airflow to the vents.
My switch has the copper shim removed, MX-4 thermal paste between die and heatsink, as well as additional thermal pads connecting the ramshield and the copper heat pipe to the metal shielding.
I didn't remove the ramshield so there's still the original compound below it.
Ambient temp is about 20-21°C
Antutu 3D gave me somewhat in the 200000 point range but didn't really raise temps a lot.
Antutu stress test (15 minutes) slowly got it towards ~40°C
____
3D Mark
Here I ran the suggested
sling shot extreme opengl
and
sling shot extreme vulcan
temp again went up to 40°C
I was using antutus temp display to check though, as 3D mark seems to measure temps inaccurately (Antutus temp display is in the notification area.)
____
Finally, I ran
CPU Prime stress-test
Starting at about 38°C leftover from 3D Mark.
It slowly reached 43° after 10 minutes, at which point I had to stop to charge the battery.
Again, using Antutus Temp display, as CPU Prime only measures battery temp.
Overall, it seems CPU Prime is the only one that can, given time, tax the system enough to reach the ~50°C a BOTW or DQB2 could bring the switch to after a while, so it's probably the best to compare temp development.
Once the battery is charged (jesus android charges slowly...), I'll do another CPU Prime Stress test at the highest performance setting, probably checking temps every 5 minutes / seeing how fast it reaches 50°C.
It seems that docking the system aborts this stress test and I'm not far enough into android to know how to tap stuff when docked...
I'd really be interested to see how different thermal mods stack up against each other, but for the time being, it'd be enough if one or two people with an unmodded (well in the temperature sense) system are interested in comparing data?
If so, just download Antutu Benchmark (for the temp display) and CPU Prime. Both are free. Start the stress test and pull down notifications to keep an eye on temps.
Charging is definitely advised as this drains the battery.
Here's a format to make it easier:
1. Switch setup
Room Temp (estimate):
Charger: Official/Unofficial
Docked: Y/N
Android Performance Setting: Max Performance please
2. Any thermal modification?
If Y:
Thermal paste:
Copper Shim removed?
Cleaned under RAM shield?
Thermal pad over RAM area?
Thermal pads over heat pipe?
Anything else of note?
3. CPU Prime Stress test
Starting Temp:
After 2 Minutes:
After 5 Minutes:
After 10 Minutes:
After 15 Minutes:
After 20 Minutes:
Reaching 50°C:
Reaching 55°C:
Just in case max performance actually lets it jump above 55°C or something, I'd say abort then. The switch can take a lot more, but I don't wanna be responsible for any damages either.