Related
Hello all,
Following XDA for years, but wasn't posting that much. My apologize for that, trying to give my contributions now
So, to start, hereunder, a sum up of my recent tests over last German ICS rom, Stock XXLRG.
======================
This is the second version of this post, taking in account various comments so far.
Firtsly, this is aimed to share my feelings about my own experience. Hope this could help anyone in anything. Nothing more
Accordingly to Roy_W, I provide which I will call my "test phase" description at the end of post, as simple infos for you to better make your opinion according to your own use. It is not about precise calculations & results, but about comparing the same things.
Hope this helps.
======================
Tests
1. Stock LRG + Stock Kernel, No root
2. Stock LRG + Stock Kernel + Root + remove some samsung apps.
3. Stock LRG + Speedmod 3-7 + remove some samsung apps
4. Stock LRG + CF-Root LRG + remove some samsung apps.
Impressions
1. In my opinion, do not use this setup. Rom feels very laggy, many little freeze here and there but the must annoying is the battery drain and heat while wifi download.
I think this is due to screen ON + heavy download (1GB+, games like Max Payne or Nova)
Pen is not well reponsive also.
Note : Except in test 3, there should be no difference with pen responsiveness between 1,2 and 3. but that was the case to me. Dunno why ^^
2. Starting to feel better, due to the removal of some Samsung apps (Rooting should not have this kind of effect .
Complete list of removed applications at the end of this post.
The phone is more responsive, lesser lags, but other things from test 1 (battery drain, pen responsiveness) remains the same.
3. Very good setup.
Lags are mostly gone, and Pen responsiveness is far better.
Don't know if Speedmod have done some special work for that, or is just due to his overall tweaks. I think it is also related to Samsung S-Suite updates. (look at test 4)
Good battery, feels like it could be better but mostly due to Samsung Roms than kernel. With normal use (sms, few calls, internet, XDA app , new review, some always active sync like Weather, Gmail, Talk, ...), you can easily achieve 2 days (48h) before running out of juice.
Also lesser memory consumption in idle (maybe related to samsung apps that were removed)
But, bluetooth was not working and since it is a must have for me, I needed to go to test 4.
4. Currently in test. Don't have my bluetooth headset with me, so this test will be updated tonight or tomorrow for feedback about battery, bluetooth and other.
So far, overall responsiveness of the phone is ok, but not as good as with Speedmod.
Maybe a placebo ( ), but I always felt that Chainfire's kernel is behaving a little better than stock one even if the only difference is root rights.
Also lesser memory consumption in idle (maybe related to samsung apps that were removed)
Samsung Removed Apps
Always doing this since my GS. It always make the system smoother to me. Can't tell about battery save.
- All Hubs (Reader, Music, ...)
- All apps related to hubs. (Zinio, Press reader, kobo...are examples of ones eating too much storage space).
- the big one : Samsung social Hub (2 apks)
- All stock live wallpaper (storage space eater)
I did not removed the Kies apps, since there were never launched, do not bother me and was having some weird warnings about missing apks while trying.
I advised to keep All Share (which is functionnal and useful for cross hardware connection) and Samsung Apps (but deactivate sync from app settings)
Go also for setting, and deactivate Wifi sync for system updates (About Phone -> System updates).
I did not test NOTECore kernel which has good feedback (3.1 version) because of the bluetooth not working.
Test Phase
Mostly due to my work, I have quitely the same daily use.
- Kernel were flashed the morning, at full charge, with cache/dalvik cache wipe.
- Only 3G, no Wifi/GPS/Blutooth. Auto Brightness, no added/removed apps (except for tests cases above).
- Betterbattery stats, stock battery monitor and Go power master for battery monitoring
- Facebook sync off (for battery save)
- Gmail sync on
- Google+ sync & Instant Upload off (for battery save)
- Weather sync every 4 hours
- Samsung apps sync off (for battery save)
- 1h call
- Gtalk & Whatsapp always connected
- No Flash installed (feels like I should mention this)
- No kills or memory clean
- Fixed wallpaper (for battery save) & lesser lags for screens transitions and opened drawer using NOVA.
What I take into account while making my opinion about battery :
- screen on time versus battery lost for the same period of time. Having used different 720p androphones, I have my own appreciation of that.
For 2000mah full charged batteries in my use with full stock, 2-3h screen on time is really bad, 3-5h should be considered as a good average, 6h+ is a dream.
GNote has a 2500 maH battery, so I have my expectation and maybe that's why I feel bad battery bahaviour when it won't hurt you at all.
- if apps were differently used by me according to their respective impact on battery
- if the basic services of Android are stable towards battery (custom kernels often reduce that )
- any wake locks
- if battery seems not affected by side-effects like mentionned above, then wait for one week of use before stating.
What I take into account while making my opinion about smoothness:
It is about shutters between screens transition, apps transition, feeling about time for the usual apps to be launched, multitasking...nothing more that a simple user feelings
Bests
thanks for the tests... but we prefer the wiping tests... just kidding...
thanks for the info,
which method do you use to remove these apps please ?
Haha..only wipe caches & dalvik from speedmod and CF kernel. I won't take any rosk from stock kernel
For system apps removal, jusy need root and titanium backup
Sent from my GT-N7000 using xda premium
CF-root and stock kernels are identical. No tweaks. Nada. Just root and Clockworkmod recovery.
So, how can the phone lag with stock kernel and be OK with CF-root ???
Placebo ?
Maybe related to the init scripts. For sure the main thing is the root with this kernel but wihin these tests and also for all samsung devices I had with Cf, it were always behaving better.
As I also said, should be related to samsung apps removal, especially the hubs.
Sent from my GT-N7000 using xda premium
Phenryth said:
Hello all,
Following XDA for years, but wasn't posting that much. My apologize for that, trying to give my contributions now
So, to start, hereunder, a sum up of my recent tests over last German ICS rom, Stock XXLRG.
Hope this helps some of you.
Tests
1. Stock LRG + Stock Kernel, No root
2. Stock LRG + Stock Kernel + Root + remove some samsung apps.
3. Stock LRG + Speedmod 3-7 + remove some samsung apps
4. Stock LRG + CF-Root LRG + remove some samsung apps.
Impressions
1. In my opinion, do not use this setup. Rom feels very laggy, many little freeze here and there but the must annoying is the battery drain and heat while wifi download.
I think this is due to screen ON + heavy download (1GB+, gamers like Max Payne or Nova)
Pen is not well reponsive also.
2. Starting to feel better, due to the removal of some Samsung apps.
Complete list of removed applications at the end of this post.
The phone is more responsive, lesser lags, but other things from test 1 (battery drain, pen responsiveness) remains the same.
3. Very good setup.
Lags are mostly gone, and Pen responsiveness is far better.
Don't know if Speedmod have done some special work for that, or is just due to his overall tweaks. I think it is also related to Samsung S-Suite updates. (look at test 4)
Good battery, feels like it could be better but mostly due to Samsung Roms than kernel. With normal use (sms, few calls, internet, XDA app , new review, some always active sync like FB, Weather, Gmail, Talk, ...), you can easily achieve 2 days (48h) before running out of juice.
Also lesser memory consumption in idle (also related to samsung apps that were removed)
But, bluetooth was not working and since it is a must have for me, I needed to go to test 4.
4. Currently in test. Don't have my bluetooth headset with me, so this test will be updated tonight or tomorrow for feedback about battery, bluetooth and other.
So far, overall responsiveness of the phone is ok, but not as good as with Speedmod.
Should be related to various tweaks, and ChainFire always aimed to provide his kernels, with Root and few tweaks.
I had GS, GS2, and CF kernels were always really nice for those looking to a Rom not too far from Stock, but with some performance tweaks and root.
Also lesser memory consumption in idle (also related to samsung apps that were removed)
BUT, having some troubles with NOVA with this setup. Several freezes, and needs to restart the launcher.
I think it is related to root access which needs to be set again coming from Speedmod to CF. Will keep you informed.
Samsung Removed Apps
Always doing this since my GS. It always make the system fly
- All Hubs (Reader, Music, ...)
- All apps related to hubs. (Zinio, Press reader, kobo...are examples of ones eating too much storage space).
- the big one : Samsung social Hub (2 apks)
- All stock live wallpaper (storage space eater)
I did not removed the Kies apps, since there were never launched, do not bother me and was having some weird warnings about missing apks while trying.
I advised to keep All Share (which is functionnal and useful for cross hardware connection) and Samsung Apps (but deactivate sync from app settings)
Go also for setting, and deactivate Wifi sync for system updates (About Phone -> System updates).
I did not test NOTECore kernel which has good feedback (3.1 version) because of the bluetooth not working.
Hope this helps.
Bests
Click to expand...
Click to collapse
Got to be honest I disagree with a lot thats written here.
Point 1 :- Why shouldn´t we use this set up??You might have or had problems but a lot of other users don`t,me being one of them.I have had NO laggs,my battery doesn´t heat up that much with WiFi,nothing freezes and the battery life is 100% top,all what you did in Point 3 with usage I get in Point 1 plus a few hours more.
Point 2 :- Once again I disagree,I´ve had no problems what so ever,only rooted for a few extra apps and removed NONE of the Samsung apps,still fast and smooth.Battery life still top,even upto 3 day usage depending on what I´m doing.
Point 3 and 4.....can´t say anything about these points because I won´t go this far,why should I ,everything is super fast,no laggs,a lot less battery drain:good: no need for any mods.(bluetooth also works )
Maybe I´m one of the lucky ones with a good Note :laugh:but this report won`t help me,maybe a few with problems.
Hi I think you don't agree wih my report cause of different perceptions between you and I.
Battery perf is linked to what is acceptable or not for both of us and also from various comparison through many flashes since GB.
And for sure, Linked to the type of use we both have.
It is the same for smoothness. I must say I am very sensible to lags and freezes even milliseconds one For instance, I am so maniac with it that I can detect a lag in a movie while no one noticed around me.
Plus, you can tell about many things to explain mobile OS perf and I will not accept this since I can not accept that a phone running a spec near to a PC ones, suffers such perf issues.
The best example is the smoothness experienced with my nexus with Jellybean against ICS. It is day and night.
Sent from my GT-N7000 using xda premium
Thanks OP for the comparison :victory:
Good initiative from OP for the comparison. Each person use the Note in a different way and will have different results surely from OP. But the main point of this thread was to give us a basic idea on the different setups OP has used. Thanks!
Thanx for making thread with detail study, thogh i can not be agree with many of statements.
Here something i like to say
- There is nothing much difference between test 1 & 2 except you have Root and removed some applications.
So does it mean
*Rooting can improve performance - I dont think so
*Removing application just making space on system ROM, so does making room on system ROM, can increase performance? I dont think so
many application you have removed may not be activated on start, naturally that doesn't have any relation with performance, b'coz it just occupy space not RAM
Now some application which may have enabled on boot, even that gives you temporary extra RAM, As android itself have LMK(low memory killer) which auto kills background applications which is not in used - LRU(least recently used), and when there is enough RAM available, even you exit from applications it remains in background for quike reopening, that is basic android mechanism. So if your aim to remove some application to make available more RAM, it won't be fulfilled, b'coz that application will be replaced by recently used application by users till android system dont sense as RAM is full. (you ca test it - when you have many applications removed, sure you will have much free RAM on boot, but if you look again after 15 min of play with device, it will be again full, same as who dont have remove blotwares, so wat benefits we got by removing those?)
Yes atleast i can say applications which was running in background for some time can consume little juice, but if you believe me, my experience is it can save maximum 3-4%(of full cycle) nothing more than that. But performance? I dont have any logic to believe it.
-I can't comment on test 3, as i have never used any custom kernel, and leaving your opinion untouched.
-Again between 2 and 4
yes you have also not tested much, but somewat if your opinion goes in favor of 4 (as you said in 1&2 responce is not good but in test 4 you saying it is OK) then if i am not wrong chainfire himself said, his kernel is complete stock (minus mmc_erase_cap command +root+su binary+CWM)
so in that case your opinion may not be placebo?
Anyway dear don't feel bad. appreciate your workout, but some points which i said dont favor me to agree completely.
Regards
I think my note must have been XXLRG when I got it but after trying to do some rooting things and bricking it, then finally loading XXLR1 (ICS) It is working great with no lag or freezing what so ever. Also have scored up to 4050 on Quadrant!
Da mOnKeY said:
Got to be honest I disagree with a lot thats written here.
Point 1 :- Why shouldn´t we use this set up??You might have or had problems but a lot of other users don`t,me being one of them.I have had NO laggs,my battery doesn´t heat up that much with WiFi,nothing freezes and the battery life is 100% top,all what you did in Point 3 with usage I get in Point 1 plus a few hours more.
Point 2 :- Once again I disagree,I´ve had no problems what so ever,only rooted for a few extra apps and removed NONE of the Samsung apps,still fast and smooth.Battery life still top,even upto 3 day usage depending on what I´m doing.
Point 3 and 4.....can´t say anything about these points because I won´t go this far,why should I ,everything is super fast,no laggs,a lot less battery drain:good: no need for any mods.(bluetooth also works )
Maybe I´m one of the lucky ones with a good Note :laugh:but this report won`t help me,maybe a few with problems.
Click to expand...
Click to collapse
I have to agree with this post. I didn't recognize any of the problems stated in the 1st option. I am currently running stock everything after having tried several different combinations of ROMS/Kernels.
There is also a lot of missing information in relation to what exact tests were performed and how they were measured : Whats apps were running at the time, were Wifi / GPS / Power Saving active, what was the state of the battery, how was "lag" measured, what brightness setting was used, Was a good 3g station nearby or was the telephone continually searching for a stronger signal, etc etc etc , The list could go on eternally.
As for anything batttery related, I do not understand how tests can be performed that have not been run over a relatively long period.
There are far too many parameters and options possible, that it would appear that the tests are merely personal interpretation.
I appreciate that you have taken the time and effort to do the testing but I am not convinced that the results can be treated as anything other than one persons thoughts at a given time.
I think that you can only perform these kind of tests when they are performed "identically" on a device that is reset from scratch and that the tests performed are measurable with precise calculation rather than the human eye/sentiment.
It is an almost impossible achievement outside of a qualified environment with a very strict set of controls.
Can the results be reproduced, is there any possibility that the results could change, would a third party be able to reproduce the same results ?
Unless of course detailed results and calculations are actually available, in which case it would be very interesting to be able read them.
Hi Dr. Ketan. Again, thanks for your guide, test 1 & 2 was done following it
Do not worry, I won't feel bad about your feedback, as for the previous one.
For sure, I think you and Chainfire knows that type of things better than me (no needs to tell why ^^), just reporting here my feelings over my tests.
*Rooting can improve performance - I dont think so
Click to expand...
Click to collapse
Think the same.
*Removing application just making space on system ROM, so does making room on system ROM, can increase performance? I dont think so
many application you have removed may not be activated on start, naturally that doesn't have any relation with performance, b'coz it just occupy space not RAM
Now some application which may have enabled on boot, even that gives you temporary extra RAM, As android itself have LMK(low memory killer) which auto kills background applications which is not in used - LRU(least recently used), and when there is enough RAM available, even you exit from applications it remains in background for quike reopening, that is basic android mechanism. So if your aim to remove some application to make available more RAM, it won't be fulfilled, b'coz that application will be replaced by recently used application by users till android system dont sense as RAM is full.
Click to expand...
Click to collapse
Removing samsung apps does not affect that much battery, but smoothness. I certainly wrongly express myself.
Maybe CF kernel has nothing to do with it since I was always flashing this and removing the hubs (especially the social one) together.
But fore sure, since my galaxy S, even on the S2, and now on the NOTE, doing this does increase system smoothness.
I do not know that much about Android memory system managment, but the first time I've done this was because I was sick about seeing these services running in background, even without ever launching them once and it does consume memory.
Yes atleast i can say applications which was running in background for some time can consume little juice, but if you believe me, my experience is it can save maximum 3-4%(of full cycle) nothing more than that.
Click to expand...
Click to collapse
I believe you without any problems
You know the deal, 1% here and there...at the end, it saves some battery.
I will update my post, but I can now say that Speedmod provides better battery than Stock/CF, looking at my battery stats for today.
Roy_W said:
I have to agree with this post. I didn't recognize any of the problems stated in the 1st option. I am currently running stock everything after having tried several different combinations of ROMS/Kernels.
There is also a lot of missing information in relation to what exact tests were performed and how they were measured : Whats apps were running at the time, were Wifi / GPS / Power Saving active, what was the state of the battery, how was "lag" measured, what brightness setting was used, Was a good 3g station nearby or was the telephone continually searching for a stronger signal, etc etc etc , The list could go on eternally.
As for anything batttery related, I do not understand how tests can be performed that have not been run over a relatively long period.
There are far too many parameters and options possible, that it would appear that the tests are merely personal interpretation.
I appreciate that you have taken the time and effort to do the testing but I am not convinced that the results can be treated as anything other than one persons thoughts at a given time.
I think that you can only perform these kind of tests when they are performed "identically" on a device that is reset from scratch and that the tests performed are measurable with precise calculation rather than the human eye/sentiment.
It is an almost impossible achievement outside of a qualified environment with a very strict set of controls.
Can the results be reproduced, is there any possibility that the results could change, would a third party be able to reproduce the same results ?
Unless of course detailed results and calculations are actually available, in which case it would be very interesting to be able read them.
Click to expand...
Click to collapse
Many good hints in your posts.
For sure, I've made my tests with some knwoledge about that, and trying to be in the same scenario each time.
I will update my post towards the parameters activated and other you stated.
Thanx
yes custom kernel may have improvement, b'coz that is made for that purpose, and we know developer adds some tweaks to improve performance. though i have not tested it, i can be agree on that point. but i have not make review just b'coz i have not used.
Regards.
Phenryth said:
Many good hints in your posts.
For sure, I've made my tests with some knwoledge about that, and trying to be in the same scenario each time.
I will update my post towards the parameters activated and other you stated.
Click to expand...
Click to collapse
Phenryth,
I look forward to your results, the basic idea of what you are doing has the potential of becoming a very interesting article.
These are tests without any statistical significance.
So in the end your heart or gut decided.
Everyone has there own view on how their phone should perform. If you are happy with it then there is no need to change it. If not then this guide provides some tips on how to improve it.
Sent from my GT-N7000 using xda premium
To me for the first time ever since the ICS was introduced the 1- setup is absolutely the best I have tried, and I did try all four of them including removing Samsung bloatware.
The touch response is super sensitive, the pen works not just perfect but it also marks when in a diagonal position something never happened before, my one and only complain is the same one since seven months: the lag which comes when opening the Contacts application for the first time ,which is absolute outrageous other than that I am satisfied.
Hi, so I've had my G2 D800 model for a while (about a year or two), and for the most part, it's been an excellent phone despite being 3 years old now. Never had any slowdown problems, and it's always performed well in games and in general.
But just recently, about a week ago, it just started lagging out and being very, very slow.
For example, if I'm watching a YT video through the app, if I rotate the screen, it takes maybe 5-8 seconds to actually rotate. Also, in a lot of circumstances where the keyboard has to pop up, it just freezes the app until the keyboard pops up maybe 10 seconds later.
It's also just laggier in general in most apps, scrolling through FB or Twitter freezes nearly every second when I'm scrolling. It's generally unresponsive and sometimes just freezes outright. It also acts this way on the homescreen, and in most everything else.
The most annoying part, however, is that sometimes the devices just doesn't turn on for sometimes up to 15 seconds after pushing the power button (and it usually just refuses to turn on when using my knock code).
The odd part, however, is that once it's actually in something like a game, it performs perfectly fine.
I've played a number of Cardboard games, as well as games like Nova 3, and it performs perfectly.
I've also ran multiple benchmarks (Geekbench, 3DMark, etc), and it scores around where it should.
Info about the device:
LG G2 D800
Stock Lollipop 5.0.2
Software Ver. D80030f
Kernel Ver. 3.4.0
Rooted
TWRP Recovery installed
XPosed Lollipop installed
G3 Tweaks installed, a number of tweaks applied
Everything else is stock (homescreen, etc)
What I've tried/checked:
Checked memory usage during "lag sessions", rarely dips below 300mb free
Changing minfree values through Rom Toolbox (tried every preset, as well as setting everything to the max to free as much as possible)
Rebooting (affects nothing)
Checked CPU speed with Rom Toolbox (is at max, and does achieve the max speed of 2265 MHz. Setting it to performance mode/setting the min speed to be 2265 MHz has no effect, but reduces battery time)
Changed default cache sizr (tried every preset from 128kb to 4096kb. With multiple benchmarks, the optimal value for r/w bounced between 512kb and 4096kb, currently on 4096)
Things I have not tried (that I am aware would possibly have an affect):
Changing anything in the kernel tweaks/build.prop
Completely wiping, reformatting, and reinstalling (I really would rather avoid this if possible, as it would take a while to get everything back to what it is now, and it was a pain to upgrade to Lollipop in the first place)
I am unaware of anything else I could try to fix it, which is why I'm here
So, does anyone know of any other solutions or things to try? This has made my phone very unreliable, even for phone calls and messages, which is a problem.
Thank you for your time and help!
Jtpetch said:
Hi, so I've had my G2 D800 model for a while (about a year or two), and for the most part, it's been an excellent phone despite being 3 years old now. Never had any slowdown problems, and it's always performed well in games and in general.
Click to expand...
Click to collapse
Hi,
First thing i need you to try out is this custom kernel : http://forum.xda-developers.com/lg-g2/development/lp-5-0-2-kernel-3-4-107-dorimanx-1-0-lg-t3102512
Then install greenify, and hibernate all the unnecessary apps.
This should solve the issue.
iubjaved said:
Hi,
First thing i need you to try out is this custom kernel : http://forum.xda-developers.com/lg-g2/development/lp-5-0-2-kernel-3-4-107-dorimanx-1-0-lg-t3102512
Then install greenify, and hibernate all the unnecessary apps.
This should solve the issue.
Click to expand...
Click to collapse
Thanks for the quick reply!
I'll try out greenify, though there aren't many unnecessary apps running (i uninstalled all the bloatware first thing, when I rooted).
And about the custom kernel, it's something I've never tried, so I have a couple questions.
Will it wipe my device, or break TWRP, root, or Xposed? (Sorry, didn't see the answer on that forum)
Will a TWRP backup I make be able to fully restore my device to it's current condition?
Is everything that is supported on my device be supported with that kernel? (Apps, etc. Don't know a whole lot about how kernels work, so I don't know)
Thank you!
Jtpetch said:
Thanks for the quick reply!
I'll try out greenify, though there aren't many unnecessary apps running (i uninstalled all the bloatware first thing, when I rooted).
And about the custom kernel, it's something I've never tried, so I have a couple questions.
Will it wipe my device, or break TWRP, root, or Xposed? (Sorry, didn't see the answer on that forum)
Will a TWRP backup I make be able to fully restore my device to it's current condition?
Is everything that is supported on my device be supported with that kernel? (Apps, etc. Don't know a whole lot about how kernels work, so I don't know)
Thank you!
Click to expand...
Click to collapse
Hi,
Even you uninstalled bloatware, the apps you use daily could be running in background and taking all those rams. So whenever you play a game or dont use a certain app, just go to greenify and hibernate. Since you have xposed installed, you can do more than that such as u can stop sync , etc.
About the kernel, it has a ram management settings and cron task.
Ram management dont need to be altered. It has some preset profile so if you seek performance, just choose performance or if u need battery , select battery profile etc.
Cron task is useful because it will do alot of things automatically. For example, it will release ram at certain amount of time from the apps , it will restart google play service since that could be an issue , etc.
All you need to do is take a nandroid backup just to be safe. Then download Dorimanx Ota from playstore and it will detect ur device and show u which version of kernel is available. Then download that, go to recovery, flash it, and you will see three app installed : Synapse -- will give you all the details of ur setup including cpu speed, temp etc for monitoring purposes
Dorimanx settings - this is the main settings of the kernel. You will see a P icon on top, tap it and choose your desired profile.
Color management : It is as the title says, just leave it be.
Then use your device, check your ram and report back. Hopefully it will resolve your issue.
Wow. Dorimanx did a lot more than I was expecting.
Downloaded the app, downloaded the kernel, flashed it no problem, went through and checked out synapse.
Went into the Dorimanx settings, and set it to the "Performance" profile.
Immediate improvement.
The whole phone seems a lot more snappy, and I switched back and forth between a 1080p60 YT video, facebook, and twitter with absolutely no problem.
Free memory when idle (no apps open) stays around 1gb; it used to be about 500mb before.
Tested a few games for the heck of it, and it actually seems to have improved game performance somehow. (It seems to have overclocked my CPU to 2.5ghz, though, so that makes sense.)
Temps stayed at around 58-60c while running a few Cardboard apps, so it seems fine.
Thanks for the help iubjaved! This has fixed my issue and then some! Ah, the wonders of the Android dev community.
Jtpetch said:
Wow. Dorimanx did a lot more than I was expecting.
Downloaded the app, downloaded the kernel, flashed it no problem, went through and checked out synapse.
Went into the Dorimanx settings, and set it to the "Performance" profile.
Immediate improvement.
The whole phone seems a lot more snappy, and I switched back and forth between a 1080p60 YT video, facebook, and twitter with absolutely no problem.
Free memory when idle (no apps open) stays around 1gb; it used to be about 500mb before.
Tested a few games for the heck of it, and it actually seems to have improved game performance somehow. (It seems to have overclocked my CPU to 2.5ghz, though, so that makes sense.)
Temps stayed at around 58-60c while running a few Cardboard apps, so it seems fine.
Thanks for the help iubjaved! This has fixed my issue and then some! Ah, the wonders of the Android dev community.
Click to expand...
Click to collapse
No worries, happy to help. I gave you advice based upon my experience using all those .
Feel free to post in XDA if you need any sort of help, you will find many that will give you insights and stuff to solve your issue.
GL :good:
Here are some screen shots of perfect CPU utilization with stock system and a ton of exposed modules.
What you want is OSmonitor to be at 1 and 2 about 50% to 80% of the time on the usage list, but that is more common with custom kernels. With stock, 1 and 3 becomes more common.
What causes hits to the system to keep OSMONITOR from stabilizing in the top ranges of utilization... quite simple; Modules that directly require the system to update screen or pol processes in background. In simple terms, the need for the OS to do the dirty work.
That is why I don't use Tasker, battery performance enhancements, and themed system type mods. Without them I can get 12 hours on a single charge with minimal sleep time, so by identifying how apps effect the system, one does not need to add overhead to correct, just need to do some detective work.
I noticed that Wanam modules use some CPU by poling the system. Even though there are plenty of good xposed modules, some are dirty in how they operate.
Not that they cause some overhead, but how they interact with other modules if used together.
Even if you find a nice app to do some custom work to your device, some use network connections to send data, either to play store, Google services, or to the developer site for debug or usage tracking. I suspect a bit of overhead is why network polling by such apps should be stopped if possible before dumping one or more because they hot on utilization.
One in particular is Solid Explorer. Not only I blocked ads, I had to block the app as well from background network usage. It is easy to find by looking at the usage data in network setting. If you are not using IPC$ or bothering with cloud storage, then by all means kill network usage for the app in question.
Right now, I am looking into Tunnelbear, Samsung's video player, and the stock video player on TouchWiz devices. Both do some odd things but are manageable. Mostly because I know what to expect.
Tunnelbear is now hooked to auto run when on an unsecured network. So it leaves its network service on when not used. Plus registration may stay on after closing. Very messy for programming, even through one may see the app as well made. IMO registration should turn off with the app closing. As for its network service, it should only enable when the option to auto start is enabled. A simple foresight many programmers don't see until pointed out.
Here I don't program and know better... a true programmer should know the processes inside and out before writing a line of code. Maybe resorting back to assembly will brake such programming habbits, that are easy to make with tools to plug away with in programming technique. Yep I know how to program, but I am not a programmer! It's a job that does not peak my interests.
Notice the attached images... the CPU usage will fluctuate but the CPU speeds will hold at one core enabled and at 200Mhz on stock kernel. On custom, it could be down to 100Mhz and any number of active cores depending on the kernel tweaks.
Partial fix for tunnelbear
Since I touched on tunnelbear keeping active connections for registration and network service. I was able to keep network service from being an issue, by preventing the service from running. I also prevented the quicksettings tile service from running as well. Both are not required and for my usage tunnelbear does not work well as a full vpn outside of streaming and messaging, due to the lack of multiple connection limit. So even if I wanted to have tunnelbear auto connect when connecting to a public Hotspot, I would be unable to use the VPN the way i want to.
So far the only issue is the registration service sticking in background once tunnelbear is closed.
Latest vs Older apps
Again I was forced into doing a trial and error run of older app usage vs that of newer and latest app functionality.
Tunnelbear and BBCiplayer were knocked down, for two different reasons.
BBCiplayer had difficulty with downloads completing and clear playback, as it seemed below 720i standards. Had a older good copy in my tibackup backups. That fixed one issue. As for tunnelbear... there was way too many so called improvements that I did not use. Plus having the old and dated version cleared up network service loading, as it did not hook with detecting unsecured networks for VPN auto connection. As for registration being loaded and running, it seems to be a problem since v 100 or so, and all the older ones fail to connect now. At least the registration service will close it's connections when not in ise, unlike the newer versions.
There are many other apps I will not update because of such trouble. I rather have 100% functionality with an app, than to have a bug or two plague what I expect out of it. Plus I don't need all that Facebook, Twitter, and other social stuff tied into what I use. Actually a browser is all I need for social activity if I choose. Disconnecting your devoce from social functionality and syncing. Truely releases your device from the power being consumed by these unnecessary apps, even when they are not in use.
You could freeze them, but it is more of an Inconvenience than what it should be.
At least with a browser, once your done, it's is done with its connection and you did not need the use of a background service.
Idle utilization
I performed a cpu idle during normalization usage stat after tweaking my disabled services and frozen system apps.
The screen shot was done 12 minutes after wake up, screen always on and had just hit 99% on a full charge.
You're busy and don't have time to wait, which is why you need to stop reading this thread and get back to organizing your Pogs. Rate this thread to express how the Xiaomi Mi A2 Lite performs when multitasking. A higher rating indicates that the Xiaomi Mi A2 Lite keeps many apps in memory so that they don't need to reload, and that when moving between apps, transitions are smooth and performance is excellent.
Then, drop a comment if you have anything to add!
I noticed recently that I have very few apps open in recents. It seems to me like the phone is killing apps prematurely. Thankfully, app launch times are not that bad, so even with reloading it is not terrible.
Hello, I'm creating an account in order to give my real-life review of this phone. The previous phones I had were HTC One M8, LG G6.
Ram management in this phone is not the best in the world. I am not the kind of guy to tap of “clear all apps” all of the time. But if you keep an app open in the background and let it sit of half a day, it will have to reload.
While I pretty much never exceeds 3.5GB of ram usage, I wish apps could be kept open without reloading.
i have 4gb of ram its very smooth phone
I changed LMK profil in Kernel Auditor to Light and turned off adaptive LMK and changed to 0mb for foreground and visible applications and I can use full ram whitout app reloading, but the phone is still smooth and fast and stable.
I have a problem with multitasking.... Processes of all apps turns off right after leaving them. Tried everything, went through all battery settings, developer's settings and such. Tried every option. Multitasking is simply impossible on Mi A2 Lite (Pie 9.0). Apps turns off instantly after I leave them. Optimization for each app is turned off, Adaptable battery turned off, Battery saver turned off, also. Nothing helps; So sad, and it seems there is no solution for this.
EDIT:
It works, phone needs restart in order for settings to be applied.
So far no problems encountered in multi-tasking, on Android 9. Not rooted.
Has anybody found any fixes for this?
VS88 said:
I noticed recently that I have very few apps open in recents. It seems to me like the phone is killing apps prematurely. Thankfully, app launch times are not that bad, so even with reloading it is not terrible.
Click to expand...
Click to collapse
It is terrible.
I use the original ram and if I copy text paste it to other app then go back to copy another when I go to paste it's already lost the first before the reload.
For example I copy a bank account number, paste it go back and copy the name and after gone. Because of reload the account number disappears.
And this is just one example how annoying it is
It is pretty bad. Although, I feel that it had been better before certain security updates, but I couldn't remember specifically which ones. Right now, at August update, I've got all optimisations turned off, battery saver turned off, adaptive battery turned off, individual app optimisations are also turned off, in developer settings I've got standard limit of background processes, killing apps is turned off.. I really don't know what else one could turn off, in order to render multitasking operable. I'm on stock with unlocked bootloader. Any help would be massively appreciated, because I am really frustrated by having two out of four gigabytes of RAM which is always free and unused, but my apps are getting killed almost instantly after I've left them.
My multitasking is better since I have disabled Adaptive Battery.
godoy.rafa said:
My multitasking is better since I have disabled Adaptive Battery.
Click to expand...
Click to collapse
Well, it should be, because, what this option does - it basically either kills all apps (if on) or lets them all work in background literally all the time (if off). That's why it's not really a good idea to keep it off because it will absolutely destroy your battery life.
I haven't updated since the October update and am not planning to - it is the last update that actually works properly regarding multitasking. All my optimisations are turned on and I reboot my phone every time after I've charged its battery - for some odd reason the multitasking works flawlessly after this.
novak.vujacic97 said:
and I reboot my phone every time after I've charged its battery - for some odd reason the multitasking works flawlessly after this.
Click to expand...
Click to collapse
Yes, if I do this even in November update, my multitasking is good. When I keep the phone on for a couple of days, it deteriorates. For a 3Gb phone, it makes no sense to have bad multitasking... Poor Android development I guess.
godoy.rafa said:
Yes, if I do this even in November update, my multitasking is good. When I keep the phone on for a couple of days, it deteriorates. For a 3Gb phone, it makes no sense to have bad multitasking... Poor Android development I guess.
Click to expand...
Click to collapse
Thats what i am talking about on my Ram Leak thread. Thank you for your comment.
Multi-tasking used to be waaaaay better in Android 8.1, but with Pie update it's worsened. I remember being amazed by how many apps could be held open in the background back then. I have the 4/64 version.
I have the 4GB version and I do not really do multitasking (multiple open windows), just share to Telegram and copy-paste between applications. It is anything but snappy; sometimes it takes looong. It does not feel like 4GB memory.
I may have discovered a way to make Shield Tablet very smooth and stable very consistently. I don't have the technical backing for it - im just a regular user doing trial and error. I hope it helps the others who want to keep this awesome Device live up to its potential...longer.
So here's my Device setup:
- Shield Tablet US-LTE
- using stock 5.4 debloated for K1 by Sh0X31 for K1 (yes, it works well with my us-lte as long as you don't care about system updates and stylus app)
- rooted via magisk
- Xposed framework
- Nova launcher
Using the ROM as it is significantly improved the performance of the device compared to stock 5.4 for US or RoW LTE. But the sluggish experience Everytime the device wakes up from deep slumber, or connects to WIFI for the first time, is still there. Using Antutu system monitor on the notification panel, I could correlate the sluggish performance with the CPU hitting above 94%. That happens 99% of the time so I suspected that most of the persistent sluggish performance is a CPU issue rather than a RAM issue. So I looked for a way to pinpoint which apps are hogging CPU operations, and hopefully discover what triggers it.
I tried enabling CPU usage in Developer's Options but alas, it could only show com.android.systemui so nothing much can be done from there. I understand this is a bug in Android 7 and it was supposed to be fixed on 7.1 but we all know Shield Tablet is stuck at 7.0 for now.
Next I tried checking out Running Services, still in Developer's Option. So I saw a lot of apps there that seem to be hogging CPU - Malwarebytes, ShareIt, Antutu, Google Services, etc. I uninstalled Malwarebytes and ShareIt but the CPU was still operating at 94% and it takes the device a few seconds to respond to my inputs.
Finally, I noticed these Nvidia-specific processes - these few KBs of operations particularly the ones for OTA upgrade and Media as well as Nvidia -customize (something like that). I forced them to stop and voila! My device CPU suddenly calmed somewhere between 30-50% and as expected, it's operating smoothly and snappily.
When I rebooted, those Nvidia processes naturally became active. I suspect that it's probably part of the device startup process. And as expected as well, whenever they're present the device is expected to go sluggish at some point before its CPU can calm down again.
So now, it's been 2 days and my CPU operates between 18-54% on normal operation, sometimes hitting 80% when I transfer files or play games. But it has NEVER gone sluggish, not even once (waking from deep slumber, connecting to WIFI, etc). I'm very satisfied and back to enjoying my Shield Tablet.
DISCLAIMER: I don't have proof that this is indeed the solution. I can't explain it technically as well. Hoping someone with more knowledge on system process can enlighten us. One thing I'm convinced though is that the sluggishness we all experience comes from NVIDIA-specific processes. I make this inference from my observation that having AOSP ROMs make the device smooth consistently. But I don't like any other custom ROMs coz I need the controller support for the tablet. I hope I make sense and hopefully this could help some Shield Tablet Fans. Sorry for the long write up.
Thanks for the post! If you want to try and dig deeper into this, consider using Tasker, I believe it can be used to make those force stops automatic on start up. Good luck!
fpozar said:
Thanks for the post! If you want to try and dig deeper into this, consider using Tasker, I believe it can be used to make those force stops automatic on start up. Good luck!
Click to expand...
Click to collapse
Thanks for the TIP. I've checked it out in Google Play, the paid version listed what it can do. Need to study if it can disable those mentioned process at startup.
I really wished Nvidia would finally make it's firmware documentation open to the public.