Armed with a new HTC One M9, the AnTuTu benchmark app (available on iOS and Android) and an infrared thermometer – I set about trying to find out whether the Qualcomm Snapdragon 810 chipset throttles its performance and how significant, if any, the overheating issues are.
- Record M9 temperature in an idle, non-stressed state for the control.
- Run AnTuTu benchmark 7 times (note: there was only a 1.8% difference between the last 3 tests – so I believed the result had plateaued and therefore stopped after 7 tests).
- Record the temperature of the M9 at the end of the final benchmark.
- Analyse temperature and performance change.
It’s worth noting that the M9’s body is aluminium and since metal is a good conductor of heat, the temperature of the body will likely be higher than other non-metal or plastic smartphones with similar hardware. What this means is that although the heat dissipation through the metal body may help the Snapdragon 810 to keep cool, it also means the body becomes hotter and therefore far less comfortable to hold.
So I start by taking the temperature of the M9 in a idle condition for the control:
I record 23°C (73.4F) as an idle temperature, so I proceed with running the first benchmark which will provide the optimal performance result.
So a pretty impressive 52343 from the HTC One M9 running the first benchmark.
Now I run the benchmark 6 more times and screenshot the results:
As you can see over the process of running the benchmark 7 times, the score has dropped from 52343 to 43770 – that’s a 17.8% drop.
So there is a clear drop in performance, but what about temperature? I used the infrared thermometer to measure the temp after the final benchmark:
So the temperature has climbed from 23°C (73.4F) for the control, to 38.2°C (100.8F) after the tests, a 66% increase.
Conclusion & Evaluation:
With a performance drop of 18.2% and a temperature rise of 66%, it’s clear that running an intensive app (a benchmark in this case) does have an impact on the HTC One M9’s Snapdragon 810 chip.
It’s difficult without further testing to establish if the performance drop is a direct result of the temperature gain, but that’s not really the point here. Yes, there is a noticeable performance drop and yes it does get a lot hotter, but the higher recordable temperature is likely a result of the metal body which conducts heat away from the chip and onto the body which causes it to feel a lot hotter.
In the real world, this test supports my own personal experience of the One M9 heating up, especially when playing power-hungry or intensive apps, to a reasonably uncomfortable level. I would describe it has becoming ‘very warm‘ as opposed to hot, so not a significant issue but something you will notice.
Less important is the performance drop. With the score dropping by around ~18% following the tests, this is unlikely to be noticeable in real-world use. You may lose a few frames while playing games but it is doubtful you would notice it or it would become an issue. This decrease in performance, in my opinion, is really only observable and measurable in synthetic benchmarks.
Of course these tests are of an abstract nature compared with your real-world experience. By concentrating on one synthetic benchmark and recording performance drops and temperature gains following an unlikely scenario of running an intensive benchmark 7 times – I am only trying to establish if, as it has been rumored, the 810 chip does lose performance after extensive use.
TL:DR – Whether or not the 810 throttles as a direct result of the temperature will take further testing, but for now the most important thing to take away is that it doesn’t really matter that much. Yes there’s a slight drop in performance, but aside from the body becoming warmer it’s unlikely you’ll notice.
I hope you found this test useful, let me know if you have any thoughts or feedback by leaving a comment! Cheers.