[dropcaps]A[/dropcaps]t long last, the generation of the problematic Snapdragon 810 powered devices seems to be drifting off on the horizon. We’ve already heard multiple rumors about the 820 and what devices it could end up powering, but today we’re going to be focusing on what’s already here, namely the 810 and 808.
As overheating is one of the characteristics so prominent in these chipsets, we decided to set up an experiment to see just how heat really affects the performance of a phone. This is obviously not a very scientific experiment, but it nevertheless paints an accurate picture of what happens when the CPU is a bit too hard at work. This is also not focused on the 810: It is tested on the G4 with the Snapdragon 808, giving a better overall picture on what happens with a device under heat.
Firstly, we began heating up the G4. With the screen brightness on 100%, we gave it the hairdryer treatment. No, this is not cheating: We want to see what happens with extreme heat, not how the chipset handles strenuous tasks, remember? (Albeit a bit exaggerated). The phone hit 133°F (56°C) in a few minutes, and was a too hot to comfortably handle.
Now, to get some idea of the performance, we ran the Antutu benchmark. For those of you that aren’t sure what that is, it is an application designed to measure your device’s speed, and gives you a total score based on the test outcome.
A few minutes passed, and we were given a score of 24,246.
Considering that the score of a G4 at normal temperature is around 49,000 (and the G2 comes in at 33,300), the heat had a massive visible impact on how well the phone performed. The device was slower and was dropping frame rates heavily. In other words, the heat totally ruined not only the synthetic benchmark, but also real life performance.
It should be mentioned that thermal throttling most likely also took a knock to the CPU. While heat can naturally stymie performance, software would’ve kicked in that told the phone to switch off some cores (the G4 has 6 cores) and limited the speed of the remaining cores. This was all the fuss about the 810 as well: OEMs were forced to deliberately slow down your phone.
Heat also degrades and damages components within your device, which is the reason throttling happens. This happens with all electronics.
During this time, the battery had 11% swiped off (in less than seven minutes) and the camera was only showing a black screen, a testament of the other side effects that can plague your phone under extreme temperatures. While it’s likely that your phone won’t reach 133°F by itself (at least, we hope), this does give an overall impression of what happens inside your beloved Android.
Then, I let it cool down itself for twenty minutes. As I was intending to send it to the freezer, I didn’t want to experience a case of spontaneous glass breakage which occurs when glass (the screen, in this case) shatters due to sudden temperature changes. Fortunately this was unlikely, and I placed it in a plastic bag before situating it in the depths of my freezer.
Another twenty-five minutes had passed before I promptly removed it and began testing again. Trying not to touch it with my warm hands, the temperature registered at 35ºF (1.6ºC). Running the application resulted in a score of 51,707, more than double the previous score, and slightly higher than the average score which was not stipulated as an exact number.
The frame rate was also significantly higher, coming in with an average of 32 fps as compared to the overheating 17 fps. This was displayed on Antutu while testing graphics.
However, there was very little gain with colder temperatures if you can recognize that the score was a bit more than that of standard room temperature. I suppose if we really want to see a boost, some liquid Nitrogen will be required, but we don’t quite want a shattered phone.
And that concludes our test, at least with the G4. We can clearly see that heat made a huge impact on the total performance, while coldness of the same extent only made marginal differences. We can only hope that this is not again the subject of topic as the new chipsets await their entrance.
Very interesting article – thank you for this test and write up. This is really interesting to me because it makes me wonder if the phone tester in Canada can produce better benchmark scores than the phone tester in Texas? Maybe the winter phone tester produces better results than the summer tester? It’s almost like they should have to report the ambient temperature along with benchmark results and that would be used as a correction factor to the normalize the benchmark results. In the car/motorcycle racing world elevation is used as a correction factor for race times at sites around the world…(i.e. the guy racing at sea level is going to have different results than the guy on top of a mountain even though they race the same distance.) http://wallaceracing.com/altitudecorrection.htm
That’s an interesting observation. Good way to cheat the scores.
My g3 was hitting 115 easy. Probably went over that many times. Can’t play COC without it heating up a bunch. And this is the problem with lg phones!
Comments are closed.