Database Users

[Q] Why does my S4 overheat so much?

I am experiencing problems with high temperatures on the phone, in stock TW but then I got CM. I set the max freq to 1000 and it still gets pretty hot (sometimes over 50 C using ondemand). I then installed KToonsez's kernel and using SmartassH3 getting the same. It got to the point where it was somewhat uncomfortable to touch the screen. How can I get the phone cooler?

Mine idles around 35-40c and when gaming gets to 50-60c. Try ktoonzconservitive governer
Sent from my SAMSUNG-SGH-I337 using Tapatalk 4 Beta

theo1998 said:
I am experiencing problems with high temperatures on the phone, in stock TW but then I got CM. I set the max freq to 1000 and it still gets pretty hot (sometimes over 50 C using ondemand). I then installed KToonsez's kernel and using SmartassH3 getting the same. It got to the point where it was somewhat uncomfortable to touch the screen. How can I get the phone cooler?
Click to expand...
Click to collapse
try this; its a micro USB fan.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}

XxSHaDoWxSLaYeRxX said:
Mine idles around 35-40c and when gaming gets to 50-60c. Try ktoonzconservitive governer
Sent from my SAMSUNG-SGH-I337 using Tapatalk 4 Beta
Click to expand...
Click to collapse
So 60 C is okay?

If your phone is lying on the table doing nothing, then no. But with heavy cpu activity, yes it is
Sent from my SAMSUNG-SGH-I337 using Tapatalk 4 Beta

XxSHaDoWxSLaYeRxX said:
If your phone is lying on the table doing nothing, then no. But with heavy cpu activity, yes it is
Sent from my SAMSUNG-SGH-I337 using Tapatalk 4 Beta
Click to expand...
Click to collapse
Okay, thanks.
I tried the SmartassH3 gov, and all 4 cores are working, and with ktoonservative, it's usually 1 or 2, rarely 3. Which one makes the phone hottest? If more than 2 or 3 apps are running, will they all run on the same core, or will they take advantage of all 4?

turn on power saving mode.

theo1998 said:
Okay, thanks.
I tried the SmartassH3 gov, and all 4 cores are working, and with ktoonservative, it's usually 1 or 2, rarely 3. Which one makes the phone hottest? If more than 2 or 3 apps are running, will they all run on the same core, or will they take advantage of all 4?
Click to expand...
Click to collapse
Ktoonservative does its best to not hold back on performance when its needed, but also use as little as it needs to at the same time. Its mainly running on 1 or 2 cores because your not doing much, so its putting the other cores to sleep until a heavier task comes out that needs all 4 used. I use the ktoonservative governor on my phone, and i find its the ultimate combination of performance and power saving. When i run heavy apps, all 4 cores activate and it runs hard, and when im doing very little, theres only 1 or 2 cores active.

pham818 said:
turn on power saving mode.
Click to expand...
Click to collapse
This is in CM10, not TW, there is no power saving mode.
Monkz said:
Ktoonservative does its best to not hold back on performance when its needed, but also use as little as it needs to at the same time. Its mainly running on 1 or 2 cores because your not doing much, so its putting the other cores to sleep until a heavier task comes out that needs all 4 used. I use the ktoonservative governor on my phone, and i find its the ultimate combination of performance and power saving. When i run heavy apps, all 4 cores activate and it runs hard, and when im doing very little, theres only 1 or 2 cores active.
Click to expand...
Click to collapse
Thanks!

[Official]Z5 Compact Benchmark Thread

Post your benchmarks here, and no "benchmarks are meaningless", we know that already (though some more and some less).
As far as i can tell 3D performance is on point, absolutely zero lag in console-like games like "oddworld" with full graphics, which is great and promising.
Here are some Antutu samples:
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}

It should be great with the lower resolution, everything else would be crazy ?
Skickat från min LG-H815 via Tapatalk

Just did my Z5C
Not bad from a compact with only 2gb ram!

@gadjet can you tell me which firmware (region and customization) do you use)? This is maximum from me right now :crying: And can you place ss from cpuz please; I want to see how fast run s810 at youre device; 1,56 Ghz an mine, and gsmarena write 2ghz?!
edit: strange things happens; now I enter again in cpuz and see tottaly different picture...and shure the better score! (3rd and 4th picture uploaded now)! Any ideas?!

His Proday please find attached, got a 2nd Z5C due to camera blur issue and newv antutu score

Cutting down on processes and uninstalling stuff definitely helps the score.

Just got my Z5C (Central European distribution - namely Czech Republic). Messed around a bit but was not overwhelmed by extraordinary performance. Slight lags here and there, not very impressed.
Ran the Antutu, it gave me quite a poor 53500 score, so I have no idea how you guys squeeze such a high scores from it. Seem like heavily throttled. Do I have an early revision?
Device part number - 1299-1434 Rev1
Customization - 1298-7258
UPDATE: Strange thing, inserted sim card, rebooted, ran again and got 62400, another run 59700.

@OtaruM i think that you have same problem as I. I have flashed firmware for central europe (with flashtool) on mu device and sometimes processor woark on 1.55ghz instead 1.96. Look at my photos higher, install cpu-z and check what happens.
@gadjet really impressive score. Which firmware region do you use?

It's a wolf in sheep's clothing.
Sent from my E5823 using Tapatalk

Omg 66.5k!?...gr8 score man!

Hi,
Today it is the second day I am using my z5 Compact. I noticed some lags, so I ran some Benchmarks. Also I noticed, that my Device is running max 1.5Ghz like someone mentioned before.
Devcheck, CPU-z and CPU Times are reporting a 1.5 GHz CPU.
Sent from my E5823 using Tapatalk

Antutu 5.7.1
z5 compact E5823
64bit mode benchmark: 65203 points
ROM version: 32.0.A.4.11
Devcheck also reporting max 1.55 Ghz, governor interactive, I/O cfq.
Not sure if devcheck might only show the frequency of the 4 A53 cores and not the 2ghz of the 4 A57 cores. Or Sony has done some underclocking to address the battery life and overheating issues of the 810. In any case, the benchmark result is pretty good and battery life as well as heat is satisfying (using it now for one week).

The A57 run at 2.0GHz. A53 run at 1.55GHz.
Source: http://www.anandtech.com/show/8933/snapdragon-810-performance-preview

Did anyone try the new v6 Antutu Benchmark?
In normal use and room temperature I only get around 65000 result, which according to the comparison table is far below the listed Z5. Only if I put the phone in my freezer and Bench it in there I can just match the 74k, that the normal Z5 is listed at.
Not too impressed to be honest.

I work in retail and a customer came into the store this morning with a Z3 on 5.1.1 Lollipop. It actually felt snappier than my Z5 Compact. Don't know how that's possible. Sony need to sort out the software optimisation because having an older phone that feels snappier than the new phone is illogical. I'm guessing Sony's software is not properly optimised for the big.LITTLE architecture of the Snapdragon 810. Marshmallow hopefully will fix this!
Sent from my SM-G925F using Tapatalk

69795 on antutu V6

supercakefish said:
I work in retail and a customer came into the store this morning with a Z3 on 5.1.1 Lollipop. It actually felt snappier than my Z5 Compact. Don't know how that's possible. Sony need to sort out the software optimisation because having an older phone that feels snappier than the new phone is illogical. I'm guessing Sony's software is not properly optimised for the big.LITTLE architecture of the Snapdragon 810. Marshmallow hopefully will fix this!
Sent from my SM-G925F using Tapatalk
Click to expand...
Click to collapse
I have both phones (Z5C and Z3C) and I see little difference in the overall feel regarding speed when using them. A little disappointing considering the money I spent but I'm not sure how much faster these devices can get.
Sent from my XT907 using Tapatalk

Mine is 77430

80441

81476 heuhueheu

One "BIG" A57 core always active?

Hi,
Been monitoring battery performance and CPU sleep states, and one thing I notice is that on my device the A53's spent a surprising amount of time at 1Ghz+ when the screen is on, and one of the A57 cores is always running a 384 when the device is idle, while the other 3 sleep. Is this Sony's attempt to keep the UI snappy? Having the A57 on all the time (screen on), even at that low frequency is going to use in the region of 100mW+ (which is about 4x that of each A53 core)
Can someone else sanity check my results?
I'm using perfmon, systempanel, and "floating monitors"
Perfmon can be found here: http://forum.xda-developers.com/showthread.php?t=1933284
It's also on the market for the price of a coffee, please consider buying it - great tool!

Yep sounds about right this Greek review:
https://translate.google.co.uk/tran.../2015/10/24/sony-xperia-z5-review/&edit-text=
Found that in most cases only one A57 core runs in the Z5.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}

Pardon my ignorance but...how much worse do you think it is to run only the big cores and throttle them from 2Ghz to 300Mhz without turning ON and switching to the little ones? (In other words...disable the big little technology and just use the best 4 cores).
We're now using the little cores more than the big ones....and that 1.5Ghz cap could be the reason of a lot of stutters.
Changing the topic, now that Google is selling Nexus devices (5X and 6P) with big little. I think they will improve this ARM technology in future Android versions (including the kernel).
Sent from my E5823

thesebastian said:
Pardon my ignorance but...how much worse do you think it is to run only the big cores and throttle them from 2Ghz to 300Mhz without turning ON and switching to the little ones? (In other words...disable the big little technology and just use the best 4 cores).
We're now using the little cores more than the big ones....and that 1.5Ghz cap could be the reason of a lot of stutters.
Changing the topic, now that Google is selling Nexus devices (5X and 6P) with big little. I think they will improve this ARM technology in future Android versions (including the kernel).
Sent from my E5823
Click to expand...
Click to collapse
The little cores will be perfectly fine for most daily mundane tasks, such as checking emails, messaging etc. The stutters you mention probably is because the phone launched just before Marshmallow and Sony are focusing more on that OS, same kind of thing with the camera. The current camera app is just a placeholder for the proper app due for release in a week or two.

thesebastian said:
Pardon my ignorance but...how much worse do you think it is to run only the big cores and throttle them from 2Ghz to 300Mhz without turning ON and switching to the little ones? (In other words...disable the big little technology and just use the best 4 cores).
We're now using the little cores more than the big ones....and that 1.5Ghz cap could be the reason of a lot of stutters.
Changing the topic, now that Google is selling Nexus devices (5X and 6P) with big little. I think they will improve this ARM technology in future Android versions (including the kernel).
Sent from my E5823
Click to expand...
Click to collapse
In short, no. The little cores, whilst relative to the A57's are weak, they are by no means underpowered for running the UI and such. Do consider that the A53 core is faster than the A9 cores that powered previous flagships such as the Galaxy S3. Leaving one of the A57's ticking in the background is probably not such a bad idea having thought about it, as it allows fast handover when the high performance is required; I do however hypothesise that some of the stuttering might be due to threads being handed over to the A57, whilst it is running down at 384, which will be somewhat slower than the high frequency A53 core(s) the process is being passed from. It may be that we see a touch-boost or more aggressive CPU governor brought in to rectify this.
On a related note, there is a good article on the power-aware BIG.little implementation in the Snapdragon 810 here:
http://www.anandtech.com/show/8933/snapdragon-810-performance-preview/4

Definitely for performance.
If they weren't set is this fashion the UI would feel sluggish and sluggish on wake up.
Your findings sound perfectly normal/adequate, resemble a "interactive" CPU governor.

Z5 Speed Variations

So I ordered 2 Z5's from Amazon (2 different sellers); one green and one white to see which color I'd prefer. Now that I have both, quality is equal (and great!) and I like both colors, so decided to dig deeper.
There are some differences in the 2 devices. They are clearly built for different regions as the white one came in a Spectre 007 package and uses the LTE symbol for LTE and the green one came in a standard Sony package, included a QCH10 charger and shows 4G for LTE speeds. (For reference I know that 4G = LTE). The white one is Rev1 and the green is Rev4.
Here is interesting part. I've setup and configured both phones to be completely identical from brightness to apps to theme, etc. When I run a speed test using the exact same SIM card, I get way different numbers.
Running 2 consecutive tests in each on T-Mobile band 4 with the phone on my desk in the exact same position, the white one (Rev1) gives me 82.38 Down/4.71 Up and 47.55 Down/3.52 Up. The green one (Rev4) gives me 94.83 Down/8.56 Up and 97.10 Down/4.81 Up.
On WiFi, they are consistent.
Running both Antutu and Quadrant on both devices is also interesting. Antutu scores virtually the same on each; the white (Rev1) at 56k and the green (Rev4) at 55k. However, on Quadrant, the white (Rev1) is 29k and the green (Rev4) is 22k.
So what I'm looking for are comments/thoughts from some of you senior folks on the differences to help me decide on which device to keep.
Thanks!!

Since you have gone through all the trouble to set both if them up, I would run a bigger sample size of speed tests, say at least 10. I think you would see things average out and be fairly equal. Only thing is, not sure how your data plan is configured. Running network speed test in bunches chews data.

You actually ordered unboxed and used 2 different devices just because you couldn't pick which color you like before sending back the one you don't want?
I know that you 'can' but...
Sent from my E6653

This post made me rerun Antutu 64-bit today on my black E6653 Rev3 where I got 55,092 only last night, same as OP.
Well, the latest result is 62,511
So, any of you guys notice the phone is faster during the day than at night?
Screenshots of Antutu results, phone uptime is 72 hours so no reboots between runs.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}

i merely get 50k on black rev4. it shows lte though..

ricodredd said:
This post made me rerun Antutu 64-bit today on my black E6653 Rev3 where I got 55,092 only last night, same as OP.
Well, the latest result is 62,511
So, any of you guys notice the phone is faster during the day than at night?
Screenshots of Antutu results, phone uptime is 72 hours so no reboots between runs.
View attachment 3534281 View attachment 3534282
Click to expand...
Click to collapse
Might depend on signal from carrier and how stable it is. Many apps and system modules launch during different conditions taking a toll on battery and perfomance for a moment. For example changing connectivity from 3G/4G to WiFi will launch several apps that will then be closed down as RAM is needed for other apps that you have launched (these apps might be in standby but do some work from time to time or just when initially launched). This due to them being set by respective app to start when connectivity changes happen. So bench them in airplane mode and with wifi off. Restart before benching.

Mine shows to be slower than the z4 lol

@EQ2000
We all know benches can vary from run to run. My point about 55K vs 62.5K on my Z5 is that OP shouldn't be using benchmark results to choose between two "identical" phones.

ricodredd said:
@EQ2000
We all know benches can vary from run to run. My point about 55K vs 62.5K on my Z5 is that OP shouldn't be using benchmark results to choose between two "identical" phones.
Click to expand...
Click to collapse
Agreed unless this indicates one phone is throttling due to inadequate application of thermal paste and/or contact between SoC and heatsink.

OP (me) isn't using benchmarks to decide. My questions (and title) were around speed variations between the 2 devices and the benchmarks were just a secondary observation.
I was just curious if it could be possible that Rev1 and Rev4 have different radios.
My apologies if the benchmarks became the root of this discussion.

One more observation today. The green Rev4 has a VoLTE button in the Quick Access drop down and the white Rev1 does not.

Z3+ A57 - 2 cores or 4 cores

So according android central's review, the Z3+ has two of the SD810's A57 cores disabled entirely. Only certain apps (literary a select few apps like the camera or antutu) will enable the phone to use all four A57's.
(can't post links, new user:crying
I don't own a Z3+, but I'm interested in getting one. Just curious if the Z3+ still has those A57's disabled under marshmallow, otherwise I might get a Z5.
Would someone on marshmallow test their Z3+ for me? All you need to do is download and use Trepn Profiler from the play store (Qualcomm's app, sorry can't post links).
According android central the A57's are cpu5, cpu6, cpu7 and cpu8.

According to the stats 5,6,8 were running on 0
Sent from my E6533 using XDA-Developers mobile app

Thanks for checking!
Ok so 3 of them were disabled, maybe that's why MM runs cooler?
If possible could you play a game or something, not a benchmark tool, and see how the cpu's behave?

In my case one was showing 0 (if I recall well, cpu7), rest were showing activity during ordinary tasks as browsing, moving through menus etc...
Sent from my Nexus 7 using Tapatalk

After some in depth investigation I can confirm that Z3+ won't use 3 of its A57 when it's gets hot specially during gaming. Only one core would work and that too at very low clock. After that no matter what you do it won't use the A57 cores. The cores would fire up only if you leave the device idele for some time. But as soon as you open Chrome the A57 cores would fire up but a low clock. This is happened in the lady update. Before that on MM even when the device was hot A57 cores would fire up. But after the last update Sony has done heavy change to the governor at the expense of performance. Even my 3 years old note 3 runs the game much faster after 5 mins into heavy gaming.
Sony should provide a performance mode option for people who need high performance during gaming irrespective of the heat. But the phone would never lag in the general applications. The only problem is that high level gaming and video editing impossible on this phone due to the A57 cores becoming inactive.

To be honest I think its actually making the battery last even less time, my z2 goes faster, cooler and much longer. Really really disappointed so far in this device and I dont know if they're going to bother improving on this.
I never used the device with mm because when I got it and added the 2nd user it gone into bootloop, so I had to flash it with the flashtool.
Sent from my E6533 using XDA-Developers mobile app

Just a quick one
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
Sent from my E6533 using XDA-Developers mobile app

It's how the Big.little has been configured. How I see it though is that it's a 4 core with a boost function. Um sure the 810 Uses HMp
The most powerful use model of big.LITTLE architecture is?*heterogeneous multi-processing?*(HMP), which enables the use of all physical cores at the same time. Threads with high priority or computational intensity can in this case be allocated to the "big" cores while threads with less priority or less computational intensity, such as background tasks, can be performed by the "LITTLE" cores.[10][11] wiki

I also noticed that. 3 cores are dead under pressure of gaming thats creepy.. Im on the latest mm. Hope that they will do something with it because i can feel its getting slower in direct test with Galaxy alpha g850f (with exynos).. Do someone know if this is a feature of marshmallow only or it is included in lollipop
Sent from my E6553 using XDA-Developers mobile app

I just tested some gaming and security software scan, and all cores showed activity more or less..