Related
Introduction
I had some spare time, and wanted to find the best ROM for my phone. So I figured I would run a series of benchmarks on the most popular ROMs i could find for the Galaxy S II. Over 350 benchmarks later, here are my results!
About the testing
- All done on the same phone
- All benchmarks were run 3 times, and then I took the average (exception is battery bench as that takes 5 hours + charge).
- Phone connected to the same network, and lied in the same spot during all battery benchmarks.
- Battery benchmark was done from 100%-50%
- Between each ROM I did a complete data wipe including dalvik.
- There were no SD-card in the phone, and the internal memory only has the ROMs and apks used for testing.
- Display brightness was cranked to max for the battery test.
Benchmarks
- Battery Benchmark
- AnTuTu
- AndroBench
- Vellamo
- CF-Bench
The ROMs
The way I chose the ROM was going through the newest couple of pages in android development and original android dev on this forum. There was a criteria that the ROM was updated during the last month.
So far I've only done 7-8 ROMs, there will be more to follow, but this is a tedious process. The next on my list is doing a couple of the latest official Samsung ROMs to see what they do. After that I'll probably do some kernel test on one of the ROMs
How to read the spreadsheet
The raw data is entered in the first 15 lines, cells show the average score, but if you click on a cell you can see the individual test scores.
The bottom part contains a score from 0-100%, this is weighted to make the highest score from all ROMs 100% and the lowest score 0%. The average score on the right is a flat average. The "Weighted" is where I've removed "Rand Write" test as its useless, and the remainding 5 storage tests count as one to not override the other tests.
EDIT: There are currently two scoring boards with different methods of calculating how the ROM did. One will be removed sooner or later, but will keep both for now.
Top one: percentage as described above
Bottom one: Average score counts 0, twice average score counts 1, and a score of 0 would count minus 1. All scores are very small, so its multiplied by 1000 in the far right column.
Statistical error
Some of these tests are extremely close, within a statistical error. On page two of the spreadsheet I've done eash benchmark 15 times on the same ROM (CM9 for anyone interested). This was to calculate a standard deviation for each test to find out just how reliable they were:
- AnTuTu and CF-bench proved to be the most reliable, with a standard deviation of under 10% of the difference found in various ROMs.
- Vellamo followed at about 20%
- Storage tests proved to be unreliable the exception being IOPS and random read tests.
I also saw some of the ROMs had much closer test results than other.
What's tested
Battery Benchmark tries to simulate a varied use where it turns on/off wifi/BT/GPS and puts the CPU through a series of cycles.
Vellamo is designed as a web benchmark simulating surfing.
AnTuTu and CF-bench is both wide benchmarks with alot of weight on java calculations.
Andro Bench is a storage benchmark testing read and write speed of the memory.
What's not tested
The big one here is stability. I did not use the ROMs as a daily ROM, I only ran the benchmarks. The best indicator here is usually the community you find in the thread of the ROM.
Also non of the ROM-specific features were tested, skins is as always a subjective case.
Top three ROMs
Based on average score:
Sensation
CyanogenMod
Oxygen
The Spreadsheet
https://docs.google.com/spreadsheet/ccc?key=0AuBi7TywV2P4dHd3ZllJUGg4TzhicUMzQm9kQWc3ZlE
--
--
Please remember to donate to the developers of your favorite ROMs
looks good, this sheet has some nice potential.
Nicely done. Lots of good data in there.
Thanks, let me know if you think of something to improve
Charging up to 100% for a battery benchmark of CM9 now.
Mixy said:
Thanks, let me know if you think of something to improve
Charging up to 100% for a battery benchmark of CM9 now.
Click to expand...
Click to collapse
Nicely done. Changing battery benchmarks from seconds to hours would be nice.
Sent from my GT-I9100 using Tapatalk 2
The 0-100% is not a very good way to rate the results. It makes the end result pretty useless to be frank, as the differences in many of the results are rather small, and it says very little. A 0% can be a split hair behind an 80% score in many of the tests.
As a suggestion, I would have given the average of each test a zero or "normal" value, and given the results for each ROM as a minus or plus percentage of the diffference from the average. At least that way your end reult (avreage across all tests) will have more meaning.
edit: But thank you very much for running the tests! Your results are certainly interesiting to compare even if I donæt think the pecentage presentation is the best. Great work!
naguz said:
The 0-100% is not a very good way to rate the results. It makes the end result pretty useless to be frank, as the differences in many of the results are rather small, and it says very little. A 0% can be a split hair behind an 80% score in many of the tests.
As a suggestion, I would have given the average of each test a zero or "normal" value, and given the results for each ROM as a minus or plus percentage of the diffference from the average. At least that way your end reult (avreage across all tests) will have more meaning.
Click to expand...
Click to collapse
Its not a perfect system, and perhaps your idea would paint another picture. I dont think it is as bad as you make it out to say though.
The system works well for most of the tests where there is an even distribution of the data. It does not work well for Random Write with .02 MB/s differantiating from best to worst. This is why I chose to introduce the "weighted" score with this one taken out.
I'll test your idea and see how it goes
EDIT: idea tested. It is quite different, and i think it is a better scoring-method. I've added it to the bottom for everyones confusion. Heres what it does:
((Test result)-(average result))/average result)
Basically giving a perfect average score a value of 0, being twice as good as the average would give 1 whole. This makes the tests count equally as just "test result-average result" would greatly favor benchmarks with high absolute numbers.
With this method best ROM is Sensation, followed by CM9 (before battery test, so could go either way here), and then Oxygen at third place.
For comparison, what do you think of benchmarking a stock ROM of some sort? A latest release of some sort, for example.
It's coming up there Guess i've done it by tomorrow/monday, will do Lightning and maybe Criskelo first.
Spreadsheet updated with benchmarks from CriskeloServan v9.
Have some problems flashing with Odin now, computer wont recognise the unit. It used to work, reinstalling drivers did not fix it, and it works with another Galaxy S II.
Mixy said:
Spreadsheet updated with benchmarks from CriskeloServan v9.
Have some problems flashing with Odin now, computer wont recognise the unit. It used to work, reinstalling drivers did not fix it, and it works with another Galaxy S II.
Click to expand...
Click to collapse
Hope you get it to work. Maybe im going to flash a rom based in info on this.
Sent from my GT-I9100 using XDA Premium HD app
Interesting that the AOSP ROMs do worst than Sammy ROMs, battery life wise :/
Not enough to make me go back to Sammy though
It seems... my battery has died
When i put the charger in all i get is a picture of a flat battery and a big fat yellow exclamation mark.
It boots with another battery, and i've been able to flash stock LP8.. But doing battery tests on a different battery is pretty useless. Will order a new one and see what i'll do, might just do the other benchmarks, not sure the battery benchmark is very reliable anyway.
I hope you get it soon. But remeber, rather battery than motherboard .
Sent from my GT-I9100 using XDA Premium HD app
I know many people have been saying the low benchmark scores on the Nexus 4 are due to "thermal throttling" but I don't really believe this. It's supposedly based on the optimus g, and if that doesn't suffer from this thermal throttling issue, then why would the Nexus 4? I personally believe it's just due to the software. People are also saying that an update won't boost the performance by that much, but I'm not sure what to believe.
So do you guys really believe that the nexus 4 is over-heating and under-clocking itself, or do you believe it is just in need of an update?
It may be a faulty unit from that reviewer, I only recall one review saying it overheated. The software is also to blame. I made a post in another thread about the performance with these pre-release software versions the reviewers have.
http://forum.xda-developers.com/showthread.php?p=33634026#post33634026
Unfinalized software on all these review devices is probably the bottleneck. Not to mention benchmarks have zero bearing on anything ever.
And Testing units and software do have a tendency to carry heavy logcats and monitoring software... I remember from the ICS days how "heavy" most leaks would run progressively getting better by the update ...
Nospin said:
Unfinalized software on all these review devices is probably the bottleneck. Not to mention benchmarks have zero bearing on anything ever.
Click to expand...
Click to collapse
Whilst benchmarks shouldn't matter too much, when this beastly specced phone is getting worse scores than the msm8960 with adreno 225, then it will obviously raise some concerns.
E3SEL said:
I know many people have been saying the low benchmark scores on the Nexus 4 are due to "thermal throttling" but I don't really believe this. It's supposedly based on the optimus g, and if that doesn't suffer from this thermal throttling issue, then why would the Nexus 4? I personally believe it's just due to the software. People are also saying that an update won't boost the performance by that much, but I know if I wrote the drivers for the phone, it wouldn't even start.
Do you guys really believe that the nexus 4 is over-heating and under-clocking itself, or do you believe it is just in need of an update? Also, another thing I wonder is this: is it called the nexus 4 because it has a 4" (4.7", I know) display, or is it called a nexus 4, because it is the fourth nexus?
Click to expand...
Click to collapse
I wouldn't say I'm good or experienced enough with Android smartphones to decide whether or not it's due to software, but I sure hope it is. I'm really only judging this particular issue by what everyone else is saying.
In regards to the sceen size and "Nexus 4" theory, I agree. Nexus 7 and Nexus 10 are a good example of device names in correlation with screen sizes here.
just wondering why did google name their nexus phone 10 wbefore the fourth one
I think it all came from a comment at Anandtech, they attempted to run all the GLBenchmark tests one after another, most web sites just chose 1 or 2 tests, usually Egypt HD. Most devices crash when trying to run all GLBenchmark test serially, it does on my Nexus 7, something to to do with running out of memory allocation.
Nexus 4 in a kind of suicidally awesome way completes the entire GLBenchmark suite in one go, but running all those test including offscreen & onscreen is a long brutal test, maxing out the SoC in a way no game is likely to do, so the fact that the device is thermal regulating itself is not that strange.
If Anandtech tested individual elements of GLBenchmark, as most other review sites do, this issue would not have occurred. In fact they admitted that the Optimus G could not run all tests consecutively, so they only tested individual elements, hence no the device didn't get downclcoked due to thermal limits. It is not good that Anandtech has this disparity in testing methodology, I like the website a lot, but some thing recently have led me to question a few things, but that is another story.
Every phone overheats nowadays so there's nothing different with the n4.
This is pretty interesting
Benchmark comparison. Once at room temp, once in freezer. Freezer scores are significantly better
http://techie-buzz.com/mobile-news/...es-show-the-real-power-of-the-s4-pro-soc.html
So yeah... Kinda does look like pretty bad thermal issues Hopefully just cause it was pre-release or something
Turbotab said:
If Anandtech tested individual elements of GLBenchmark, as most other review sites do, this issue would not have occurred. In fact they admitted that the Optimus G could not run all tests consecutively, so they only tested individual elements, hence no the device didn't get downclcoked due to thermal limits. It is not good that Anandtech has this disparity in testing methodology, I like the website a lot, but some thing recently have led me to question a few things, but that is another story.
Click to expand...
Click to collapse
But they singled out the Optimus G, because it was unable to complete more complex tests because of crashes.
mejobloggs said:
This is pretty interesting
Benchmark comparison. Once at room temp, once in freezer. Freezer scores are significantly better
http://techie-buzz.com/mobile-news/...es-show-the-real-power-of-the-s4-pro-soc.html
So yeah... Kinda does look like pretty bad thermal issues Hopefully just cause it was pre-release or something
Click to expand...
Click to collapse
Wow, that is a very interesting find. I was actually thinking that someone should do a freezer test just see if it is overheating. This article would seem to prove that it is. Those retests show dramatically higher scores, more on par with what the S4 processor should be capable of.
They said that perhaps the retail versions will have a higher tolerance for heat because they did not think they felt that hot. More testing and info is needed though.
Ryukeima said:
They said that perhaps the retail versions will have a higher tolerance for heat because they did not think they felt that hot. More testing and info is needed though.
Click to expand...
Click to collapse
To elaborate, straight from the source, in the Anandtech podcast this morning Brian commented that the phone seemed to be set to throttle at 60 degrees (for the dual-krait S4 at least it's usually around 80 afaik) and at that point the exterior of the phone was much cooler than a lot of other phones. (He talks about it at about 00:51:00.)
Something to think about.
Sjael said:
To elaborate, straight from the source, in the Anandtech podcast this morning Brian commented that the phone seemed to be set to throttle at 60 degrees (for the dual-krait S4 at least it's usually around 80 afaik) and at that point the exterior of the phone was much cooler than a lot of other phones. (He talks about it at about 00:51:00.)
Something to think about.
Click to expand...
Click to collapse
That means it's a simple tweak, and so as long as people aren't noticing excessive heat (which we would have heard from reviews) on the phone itself then is sounds like things will be fine for the release.
Yes IT FREAKING OVERHEATS
E3SEL said:
I know many people have been saying the low benchmark scores on the Nexus 4 are due to "thermal throttling" but I don't really believe this. It's supposedly based on the optimus g, and if that doesn't suffer from this thermal throttling issue, then why would the Nexus 4? I personally believe it's just due to the software. People are also saying that an update won't boost the performance by that much, but I know if I wrote the drivers for the phone, it wouldn't even start.
Do you guys really believe that the nexus 4 is over-heating and under-clocking itself, or do you believe it is just in need of an update? Also, another thing I wonder is this: is it called the nexus 4 because it has a 4" (4.7", I know) display, or is it called a nexus 4, because it is the fourth nexus?
Click to expand...
Click to collapse
I've had the phone for five days now and it definitely overheats. The last time it overheated was this morning. I got up, looked at the phone and unplugged it from the charger. Went to attend to my toddler for 20 mins and came back and saw that the phone was very warm and battery life was hit 20 percent. I did a check on battery usage under 'battery' in settings and saw that the playstore app had sucked 50% of the power of late. This is HIGHLY unusual becuase usually the screen is what sucks the most juice. I turned it off and turned it back on and it went back to being normal. During this overheating the phone stutters EXTREMELY visibly. Probably due to the massive thermal throttling of the cores. Once it cools off it goes back to being normal. This happens randomly and occurs once or twice a day. I've reported the bug to google. Otherwise the phone is superb.
That sounds like the Play Store app is misbehaving. Have you tried clearing the data?
Sent from my Nexus 4 using Tapatalk 2
Remember it's winter at northern hemisphere. So overheating might not be obvious.
Any friends in southern hemisphere (Australia) wanna chime in with their temperatures?
Paradisle said:
I've had the phone for five days now and it definitely overheats. The last time it overheated was this morning. I got up, looked at the phone and unplugged it from the charger. Went to attend to my toddler for 20 mins and came back and saw that the phone was very warm and battery life was hit 20 percent. I did a check on battery usage under 'battery' in settings and saw that the playstore app had sucked 50% of the power of late. This is HIGHLY unusual becuase usually the screen is what sucks the most juice. I turned it off and turned it back on and it went back to being normal. During this overheating the phone stutters EXTREMELY visibly. Probably due to the massive thermal throttling of the cores. Once it cools off it goes back to being normal. This happens randomly and occurs once or twice a day. I've reported the bug to google. Otherwise the phone is superb.
Click to expand...
Click to collapse
Same thing happened to me earlier today, I unplugged the phone from the charger and was extremely hot. I panic and shut the phone off. I rooted the phone last night so that was the only thing I did differently. I've had the phone for almost a week.
I also noticed the battery was draining way too quickly.
More like rogue apps... had the phone for over a month now, heavy usage every day and never once has it "overheated"as people say.
If im multitasking, like listening to music, downloading a torrent in the backgroud while playing angry birds or something the top back of the phone will get warm, but nothing unbearable if i deliberatly grab and hold the phone at that spot.
It's a glass phone... it will get warmer than most people are used to... its bascially the same glass that the new kitchen ovens use as a top surface... if you can cook on a glasstop stove... a phone heating up will be the same principle, albit on a smaller scale..
So yes, for the people complaining about heat.. then say in the next breath that they lost battery in some % form... funny how nobody is telling what they have installed as extra..or post screenshots of the battery page to back up the claim with info so we can help...large loss of battery % right away points to a rogue app somewhere.. you dont magically lose 10-20-50% whatever battery when unplugging the phone... thats something stuck running that is forcing your cpu to run at max for an extended period of time.
so as the internet expression says "fraps or it didnt happen" (screenshot or it didnt happen) lol
I am seeing battery temp reach 40C during antutu benchmark test running 4.2.2. Looks like it is only affecting a few devices. It gets warm during the benchmark test but nothing like unbearable heat.
Hi guys,
I've been an iPhone user for four years in a row. This year though, I bought HTC One developer edition for me and my wife.
Anyway, I've seen lots of threads on battery usage and stats but they reflect results that are dependent on each particular user. I felt like my battery is not lasting long so I ran AnTuTu battery benchmark test and got 457. I still have a week for the unquestioned exchange term, so if anyone out there had the same test could tell me his score, I would decide whether to return it or keep it. Cheers,
P.S. Skype was running on background during the test. I don't know if this benchmark software is clever enough to calculate that difference, but antutu was measuring cpu-usage/battery-usage so Skype shouldn't effect the test result. Advance users might know better of course. :fingers-crossed:
djan84 said:
Hi guys,
I've been an iPhone user for four years in a row. This year though, I bought HTC One developer edition for me and my wife.
Anyway, I've seen lots of threads on battery usage and stats but they reflect results that are dependent on each particular user. I felt like my battery is not lasting long so I ran AnTuTu battery benchmark test and got 457. I still have a week for the unquestioned exchange term, so if anyone out there had the same test could tell me his score, I would decide whether to return it or keep it. Cheers,
P.S. Skype was running on background during the test. I don't know if this benchmark software is clever enough to calculate that difference, but antutu was measuring cpu-usage/battery-usage so Skype shouldn't effect the test result. Advance users might know better of course. :fingers-crossed:
Click to expand...
Click to collapse
Oddly enough I have the exact same grace period so to speak to return my one also , I have also not had very good results with battery and I am considering returning it also , I even downloaded wake lock detector , gem battery monitor, to try and see what is keeping this phone awake, so far I have not been able to nail down anything, I was hoping not to have to worry about every setting , ie, WiFi,brightness, etc etc , my s3 outperforms my one by at least 30 percent , I'm finding the phone is not really going Into deep sleep the way I think it should , anyway sorry for rambling , when. Saw your post I was like uhhhh yeah me too , anyway before I posted I went looking for antutu battery bench mark and I can't find that app , do you mean antutu benchmark, and that app has a battery score I it somewhere....? I never used that app , although I am quite the flasher , sorry for rambling , but I saw on another thread that someone else was not getting Decent battery results and had to exchange there phone and when he did he got much better battery results , I thin its a software glitch , I was hoping not to have too root, to achieve better performance ..........nope other people chime in , I'm a very light user lately and I'm barely getting 3 hours of screen time and 13 hours total , phone is draining around 4 % an hour while off....
I have the one since friday, after the onex has driven me nuts for a year because of the sh**y battery.
I am absolutly amazed by the battery of the one, getting 4-5 hours screentime (watching hd movies and playing games) and still having 20% left
Philly Jim said:
Oddly enough I have the exact same grace period so to speak to return my one also , I have also not had very good results with battery and I am considering returning it also , I even downloaded wake lock detector , gem battery monitor, to try and see what is keeping this phone awake, so far I have not been able to nail down anything, I was hoping not to have to worry about every setting , ie, WiFi,brightness, etc etc , my s3 outperforms my one by at least 30 percent , I'm finding the phone is not really going Into deep sleep the way I think it should , anyway sorry for rambling , when. Saw your post I was like uhhhh yeah me too , anyway before I posted I went looking for antutu battery bench mark and I can't find that app , do you mean antutu benchmark, and that app has a battery score I it somewhere....? I never used that app , although I am quite the flasher , sorry for rambling , but I saw on another thread that someone else was not getting Decent battery results and had to exchange there phone and when he did he got much better battery results , I thin its a software glitch , I was hoping not to have too root, to achieve better performance ..........nope other people chime in , I'm a very light user lately and I'm barely getting 3 hours of screen time and 13 hours total , phone is draining around 4 % an hour while off....
Click to expand...
Click to collapse
here is the link for the app:
https://play.google.com/store/apps/details?id=com.antutu.tester
I found one review in which they've done this test and got 574 but that was a prerelease unit so things night have changed since.
djan84 said:
here is the link for the app:
https://play.google.com/store/apps/details?id=com.antutu.tester
I found one review in which they've done this test and got 574 but that was a prerelease unit so things night have changed since.
Click to expand...
Click to collapse
I will report back with my score , thanks for the app I'm interested to see where I stand and if you get a chance you might try one of the two apps I use at least they tell you what's using your juice.........
Edit Okay maybe i'm mistaken but in order to run this test i think you need to be at at least 95% and it must take a long time , i've tried several times but no good phone is on charger will try again ....................jeeze lol
Dear Community,
In this thread I want to gather a sample to investigate to what extent the S4 suffers form overheating and how this affects the battery drainage/life. I urge you to participate as I am collecting a sample to submit a formal complaint to Samsung to raise awareness to an issue that is widespread among S4 users and to push Samsung to address this issue accordingly. In order to provide your sample please follow the instructions below and report back in this thread with the relevant information. Please first indicate which variant you have: Quad/Octa (i9505 or i9500)
Please download BatteryGraph (it's a great app to measure Battery drainage accurately) https://play.google.com/store/apps/details?id=com.modroid.battery
Run it once after download to initiate logging. Let it run for 10 minutes to allow the graph to built up a bit.
Then please Download Antutu benchmark run the benchmark fully exactly 5 times back to back (continuously). After that please enter the stock dialer (as quickly as possible) and enter *#0228# and post the temperature (external thermistor) after the tests in order to determine the maximum temp that the device reaches under continuous stress.
After that, please enter the app Battery Graph and zoom into the graph to the maximum level. Scroll slowly along the curve at the time when you performed the Benchmarks and tell me if there are interruptions or 'skips' where the battery drops 1 or 2 percent at a time and where the App does not Register a graph for those drops. Sequence/scroll slowly along the graph 1% at a time and make sure that the battery dropped 1% at a time. Please report of the battery has dropped more than 1% at a time. Please make a screenshot of your battery graph in the App and post it here for collection. I will analyze the graphs subsequently and compile it in SPSS to submit our findings.
If there is a drop of 2 or 3% at a time it means that you are experiencing Battery Percentage skips/cliffs which could indicate either a defective battery or defective device which might be the cause for the extensive heat development.
Thank you for your participation. Hopefully Samsung will listen to us and address the issue to give us the perfect S4 that we deserve.
Best,
Thomas from Germany
exxi said:
Dear Community,
...
Thank you for your participation. Hopefully Samsung will listen to us and address the issue to give us the perfect S4 that we deserve.
Best,
Thomas from Germany
Click to expand...
Click to collapse
That battery app polls for battery level changes every 5 minutes by default. I think almost every device will experience a drop of more than 1% every 5 minutes under constant stress testing. You can change it to poll every 1 minute in the settings, which you might want to mention, but even so isn't it entirely possible to lose more than 1% a minute with everything turned on (GPS, Gestures, etc.) and ~17 minutes of constant cpu/gpu stress testing? It's a nice idea but it doesn't sound like it would yield very accurate results, unless I've misunderstood your post.
EDIT: Nevermind, I see how you can step through it 1% at a time now.
Meltus said:
That battery app polls for battery level changes every 5 minutes by default. I think almost every device will experience a drop of more than 1% every 5 minutes under constant stress testing. You can change it to poll every 1 minute in the settings, which you might want to mention, but even so isn't it entirely possible to lose more than 1% a minute with everything turned on (GPS, Gestures, etc.) and ~17 minutes of constant cpu/gpu stress testing? It's a nice idea but it doesn't sound like it would yield very accurate results, unless I've misunderstood your post.
I've also noticed that you cannot really zoom in very far and at the highest zoom setting the % is still displayed in multiples of 10. Kind of tricky to differentiate between a 1% and a 2% drop.
Click to expand...
Click to collapse
You might be right, what would you suggest in order to yield more accurate results? I suspect that the heat is directly linked to a very quick drainage of the battery. How can we measure how much it affects the actual battery drainage?
exxi said:
You might be right, what would you suggest in order to yield more accurate results? I suspect that the heat is directly linked to a very quick drainage of the battery. How can we measure how much it affects the actual battery drainage?
Click to expand...
Click to collapse
I'm not criticizing you or anything and I'm not saying it's a bad idea, It's just that obtaining accurate battery results are tricky as everyone's device is different, has different settings enabled and are running different kernels/ROMs/etc. Batteries also seem to get more efficient as time goes on so newer devices might suffer from higher battery drains.
I'll run the battery temp test now though and see how hot mine gets!
I like your critical thinking, obviously you are right it is very trick but I believe the heat test itself could be very indicative. In the meantime I will try to find a better way to measure battery drainage.
So everyone please try to run the benchmark multiple times and submit the temperature.
The most I was able to get was 52C
Thanks
Pre-test: 29.9°C
After 3 tests: 35.0°C
After 5 tests: 37.7°C
After 7 tests: 38.4°C
After 10 tests: 38.8°C
The temperature increased less and less so I don't know how much higher I could get it.
Also, an easier way to poll the temperature is to do "adb shell dumpsys battery" over ADB. The temp will be displayed as something like 388 which means 38.8°C
Edit: It might be worth noting that the top of the phone, around the camera but more so on the screen side (so the cpu?) got incredibly hot. Almost too hot to touch. Kinda worrying, but I have spent the last half an hour stress testing, I guess
Meltus said:
Pre-test: 29.9°C
After 3 tests: 35.0°C
After 5 tests: 37.7°C
After 7 tests: 38.4°C
After 10 tests: 38.8°C
The temperature increased less and less so I don't know how much higher I could get it.
Also, an easier way to poll the temperature is to do "adb shell dumpsys battery" over ADB. The temp will be displayed as something like 388 which means 38.8°C
Edit: It might be worth noting that the top of the phone, around the camera but more so on the screen side (so the cpu?) got incredibly hot. Almost too hot to touch. Kinda worrying, but I have spent the last half an hour stress testing, I guess
Click to expand...
Click to collapse
Thanks for your contribution Meltus! Really appreciated my friend. Also thanks for the ADB command it does make it easier but I doubt that you've read the correct temp as 38 appears to be quite low compared to what i got. Is your S4 a i9505 ?
exxi said:
Thanks for your contribution Meltus! Really appreciated my friend. Also thanks for the ADB command it does make it easier but I doubt that you've read the correct temp as 38 appears to be quite low compared to what i got. Is your S4 a i9505 ?
Click to expand...
Click to collapse
It's also the temperature that dialling *#0228# gave me (I double checked they were both the same each time). And no, I have an i9500, running Omega ROM, if that helps.
Might want to mention the elementary precaution of taking off all case covers except for the stock case before testing to prevent insulation effect of case covers increasing battery temperature...
While you're at it you might want to standardize room temperature to 25 degrees C... I know for a fact that my device would never get hot no matter what it does if running in the freezing cold air-con'd metro...
What about brightness settings and whatnot. Should all be the same, so you have to give a standard.
Pre-test: 36.2°C
After 3 tests: 51.4°C
After 5 tests: 54.6°C
After 7 tests: 56.9°C
dafaq right??? im really annoyed at sammy for this overheating S4... after 7 times testing the Antutu, i stopped and LITERALLY put my S4 in the lower compartment of the refrigerator to get it cooled down quickly cause im pretty sure the heat sensors inside would definitely be having some testing errors and the temperatures would be really really higher than that. even when i swipe between different homescreens the phone gets to 47°C which is really annoying.
can someone pls for the LOVE OF GOD provide any solution to this freaking problem? i have been worrying a lot for spending my $730 on an overheating phone...
i have S4 i9500 Exynos version
---------- Post added at 10:02 AM ---------- Previous post was at 10:01 AM ----------
oh and i have half the brightness all the time for everything and auto brightness turned off
Are you talking about the battery temperature? If so something is extremely wrong with your unit, my battery temperature never went past 45 even with high stress.
If in case you are pointing out the cpu temperature then there is no reason to freak out because the cpu can handle upto 90c + and can cool off pretty fast
Edit:
And why on earth you ran Antutu for 7 times continously lol? None of the current generation phones can handle it for 7 times on a row afaik
Sent from my GT-I9500 using Tapatalk 4 Beta
bala_gamer said:
Are you talking about the battery temperature? If so something is extremely wrong with your unit, my battery temperature never went past 45 even with high stress.
If in case you are pointing out the cpu temperature then there is no reason to freak out because the cpu can handle upto 90c + and can cool off pretty fast
Edit:
And why on earth you ran Antutu for 7 times continously lol? None of the current generation phones can handle it for 7 times on a row afaik
Sent from my GT-I9500 using Tapatalk 4 Beta
Click to expand...
Click to collapse
You can read the CPU temp instead of the battery temp?
bala_gamer said:
Are you talking about the battery temperature? If so something is extremely wrong with your unit, my battery temperature never went past 45 even with high stress.
If in case you are pointing out the cpu temperature then there is no reason to freak out because the cpu can handle upto 90c + and can cool off pretty fast
Edit:
And why on earth you ran Antutu for 7 times continously lol? None of the current generation phones can handle it for 7 times on a row afaik
Sent from my GT-I9500 using Tapatalk 4 Beta
Click to expand...
Click to collapse
im talking about the cpu temps. i ran antutu cause it was asked to be done for this thread.
and why would it be silly to run any app 7 times on the flagship device? its made to run the apps as much as people want thats why we pay a huge amount of money for these devices dont we? or else we should buy the low end devices
just saying cause i had heating up issues with devices before infact my last device Note 2 used to get a lot hot but not as much as this S4...
i desperately need a solution for this...
Joe0Bloggs said:
You can read the CPU temp instead of the battery temp?
Click to expand...
Click to collapse
Yes we can read cpu temp and battery temp separately, use system tuner app to read cpu temperature.
Sent from my GT-I9500 using Tapatalk 4 Beta
aami.aami said:
im talking about the cpu temps. i ran antutu cause it was asked to be done for this thread.
and why would it be silly to run any app 7 times on the flagship device? its made to run the apps as much as people want thats why we pay a huge amount of money for these devices dont we? or else we should buy the low end devices
just saying cause i had heating up issues with devices before infact my last device Note 2 used to get a lot hot but not as much as this S4...
i desperately need a solution for this...
Click to expand...
Click to collapse
You are not buying a flagship device just to run Antutu n number of times, do you? yes I accept we do pay huge money but that doesn't mean we can push the limit of a mobile and complain about it from a general user perspective . , it can be only done for experimental purposes and to understand the power /thermal envelope provided that it's done in the right way.
But how does running Antutu 7 times matches a real life scenario? Highly unlikely isn't it?
We are already pushing the limits of raw cpu power for a mobile, raw a15 cores are power hungry and often tend to heat faster than previous generation. Hence arm introduced the big little. The mobile heats up during heavy stress, but but that's expected with these powerful cores right? I'd better utilize the power wisely when needed rather than using it all the time. The cpu indeed gets hotter when stressed, it's the same With my s3 too the cpu temps reached upto 80,there is built in throttling mechanisms which will take of the cpu once the cut off temperature is reached. Have seen many benchmarkers these days using freezer test due to the thermal throttling in new gen devices. Yes the s4 heats up pretty quick than other mobiles but also cools down pretty fast and I think that's the way it is destined to work.
Hei one more thing, did you note how long it took to fallback to normal temps?
Sent from my GT-I9500 using Tapatalk 4 Beta
bala_gamer said:
Are you talking about the battery temperature? If so something is extremely wrong with your unit, my battery temperature never went past 45 even with high stress.
If in case you are pointing out the cpu temperature then there is no reason to freak out because the cpu can handle upto 90c + and can cool off pretty fast
Edit:
And why on earth you ran Antutu for 7 times continously lol? None of the current generation phones can handle it for 7 times on a row afaik
Sent from my GT-I9500 using Tapatalk 4 Beta
Click to expand...
Click to collapse
We were doing it to see if there was a correlation between excessive battery drain and overheating batteries. Mine never got any hotter than 40 and I have pretty great battery life. The OP was seeing temps of 50+ and I'm guessing he has poor battery life.
Could there have been a few batches of bad batteries sent out with devices?
Sent from my GT-I9500 using xda app-developers app
Meltus said:
We were doing it to see if there was a correlation between excessive battery drain and overheating batteries. Mine never got any hotter than 40 and I have pretty great battery life. The OP was seeing temps of 50+ and I'm guessing he has poor battery life.
Could there have been a few batches of bad batteries sent out with devices?
Sent from my GT-I9500 using xda app-developers app
Click to expand...
Click to collapse
Are you talking about battery temperature or cpu temperature? Haven't heard any stories so far regarding damaged battery, but may be possible.
Sent from my GT-I9500 using Tapatalk 4 Beta
Thanks for the Updates > and clear step to analize the issue to fix in next Sw update
Thanks for your efforts ]
Dear Community,
In this thread I want to gather a sample to investigate to what extent the S4 suffers form overheating and how this affects the battery drainage/life. I urge you to participate as I am collecting a sample to submit a formal complaint to Samsung to raise awareness to an issue that is widespread among S4 users and to push Samsung to address this issue accordingly. In order to provide your sample please follow the instructions below and report back in this thread with the relevant information. Please first indicate which variant you have: Quad/Octa (i9505 or i9500)
Please download BatteryGraph (it's a great app to measure Battery drainage accurately) https://play.google.com/store/apps/details?id=com.modroid.battery
Run it once after download to initiate logging. Let it run for 10 minutes to allow the graph to built up a bit.
Then please Download Antutu benchmark run the benchmark fully exactly 5 times back to back (continuously). After that please enter the stock dialer (as quickly as possible) and enter *#0228# and post the temperature (external thermistor) after the tests in order to determine the maximum temp that the device reaches under continuous stress.
After that, please enter the app Battery Graph and zoom into the graph to the maximum level. Scroll slowly along the curve at the time when you performed the Benchmarks and tell me if there are interruptions or 'skips' where the battery drops 1 or 2 percent at a time and where the App does not Register a graph for those drops. Sequence/scroll slowly along the graph 1% at a time and make sure that the battery dropped 1% at a time. Please report of the battery has dropped more than 1% at a time. Please make a screenshot of your battery graph in the App and post it here for collection. I will analyze the graphs subsequently and compile it in SPSS to submit our findings.
If there is a drop of 2 or 3% at a time it means that you are experiencing Battery Percentage skips/cliffs which could indicate either a defective battery or defective device which might be the cause for the extensive heat development.
Thank you for your participation. Hopefully Samsung will listen to us and address the issue to give us the perfect S4 that we deserve.
Best,
Thomas from Germany[/QUOTE]
bala_gamer said:
Are you talking about battery temperature or cpu temperature? Haven't heard any stories so far regarding damaged battery, but may be possible.
Sent from my GT-I9500 using Tapatalk 4 Beta
Click to expand...
Click to collapse
Battery temperature. If someone's experiencing temperatures in excess of 50C and bad battery drain it would indicate a problem.
I've been running benchmarks on this phone.
I've noticed a couple of things.
Thermal throttling on this phone starts at a very low threshold and is very aggressive.
Supposedly the maximum frequency of the large cluster is 2.45 GHz
At room temperature my phone never reaches that performance.
Here is a copy of the thermal-engine.conf, pulled from system/etc.
As soon as you start to benchmark the speed starts dropping:
At room temp I get Antutu ssores in the 130-150K range.
If the phone has been in a very cool room I get 160K numbers.
If I put it in a very cold environment for a few minutes I can get numbers in the 178K range.
Back to back runs plunge performance.
So it does not surprise me that the phone freezes at times.
Not happy that the phone throttles like this.
I have also found that if the phone gets to 44C then the throttling is so bad that a LG G4 or Samsung Galaxy S6 will outperform this phone.
So the thermal-engine.conf has this phone crippled.
This is speculative, but I wonder if overheating was the cause of the release delay. If so, perhaps this was the workaround they implemented to get the device out. That said, if throttling is the culprit then it could explain why my phone runs like an Evo3D much of the time. Generally before use, it's in my pocket, which is body temperature and the phone is always warm. Maybe even pocket-temp causes throttling?
Oh well, return initiated yesterday, awaiting response from CS - I just can't stand to take a chance a potentially crippled device. While the updated did increase the responsiveness of the screen a slight bit, I didn't see an overwhelming change that put this in the category of other devices running similar hardware. Back to the OP3...
cadbomb said:
This is speculative, but I wonder if overheating was the cause of the release delay. If so, perhaps this was the workaround they implemented to get the device out. That said, if throttling is the culprit then it could explain why my phone runs like an Evo3D much of the time. Generally before use, it's in my pocket, which is body temperature and the phone is always warm. Maybe even pocket-temp causes throttling?
Oh well, return initiated yesterday, awaiting response from CS - I just can't stand to take a chance a potentially crippled device. While the updated did increase the responsiveness of the screen a slight bit, I didn't see an overwhelming change that put this in the category of other devices running similar hardware. Back to the OP3...
Click to expand...
Click to collapse
So the device won't get really hot.
It won't even get that warm unless it's charging.
If you put it on a charger, it will be throttled after about a minute.
If I run Antutu and start out with the device sitting on a table in a cool room (too many conditions), I can get a score of about 167K.
If I sit it on a baggy of ice, I can get a score of about 177K but never unless the device is artificially chilled.
But if you run Antutu back to back or GeekBench and look at the temp, as soon as it hits 43C, you hit a wall.
The device won't heat up more than that but it's now running only slightly faster than a 6P and slower than a Pixel XL or Moto Z.
I don't want to pay $850 (yes that much if you include the extended warranty and tax) for a marginal device.
If I had root, I would have played with the thermal-engine.conf. It can't be right.
You go from full performance to half in a matter of 4 degrees C.
I have looked at other files for other devices and I've never seen anything that aggressive.
Even the 6P was not that bad.
I have an iPhone 6s (work) and no matter ho many times I run the test I always get about the same numbers.
I have my phone charging and I ran the test and I just got 115K. So on a charger the performance is 65% of max.
This is why I'm sending it back. Might as well be a LG G6.
tech_head said:
So the device won't get really hot.
It won't even get that warm unless it's charging.
If you put it on a charger, it will be throttled after about a minute.
If I run Antutu and start out with the device sitting on a table in a cool room (too many conditions), I can get a score of about 167K.
If I sit it on a baggy of ice, I can get a score of about 177K but never unless the device is artificially chilled.
But if you run Antutu back to back or GeekBench and look at the temp, as soon as it hits 43C, you hit a wall.
The device won't heat up more than that but it's now running only slightly faster than a 6P and slower than a Pixel XL or Moto Z.
I don't want to pay $850 (yes that much if you include the extended warranty and tax) for a marginal device.
If I had root, I would have played with the thermal-engine.conf. It can't be right.
You go from full performance to half in a matter of 4 degrees C.
I have looked at other files for other devices and I've never seen anything that aggressive.
Even the 6P was not that bad.
I have an iPhone 6s (work) and no matter ho many times I run the test I always get about the same numbers.
I have my phone charging and I ran the test and I just got 115K. So on a charger the performance is 65% of max.
This is why I'm sending it back. Might as well be a LG G6.
Click to expand...
Click to collapse
Good to know, and thanks for the thorough diagnostic! Even with rooting and changing the throttle configs, I fear a voided warranty and bricked phone if it's a design flaw and not a software bug. It's just not worth it for me to wait and see when there are other new and shiny things just around the corner. *squirrel*
cadbomb said:
Good to know, and thanks for the thorough diagnostic! Even with rooting and changing the throttle configs, I fear a voided warranty and bricked phone if it's a design flaw and not a software bug. It's just not worth it for me to wait and see when there are other new and shiny things just around the corner. *squirrel*
Click to expand...
Click to collapse
I'm pretty sure it's a mistake in the thermal-engine.conf.
I actually contacted them and submitted my information.
You don't throttle from 39-40C down to 50%
I've worked designing chips for better than 25 years.
If the 835 is more efficient than the 810, then there is no need for anything that aggressive.
Anyway, I asked for a RMA and they are sending a the boxes/envelopes.
No cases is a big deal for me also.
Since I'm already at $900 for the phone and camera (taxes and insurance). I might as well give the X a look.
Yeah, it costs more but if I'm already at $900 another couple of hundred isn't going to matter.
Unfortunately, I have to go back to my LG G3 since my 6P was sold and shipped last week.