Database Users

i only got 6.9 mflops on my x10 benchmark!

Hello folks,
I just tried to benchmark my x10 with linpack and got only 6.8 or 6.9 mflops. is this normal, is there anything i can do to make it get a higher score?
I'm using the latest .435 firmware and rooted and have all the junk pps from the 'x10 junk apps' thread removed from my phone.
Thanks for any help.

How does the phone perform? Are you satisfied with it or does it lag considerably? A benchmark really means nothing other than bragging rights. If it feels fast and feels good then why worry about mflops?
Sent from my Delorean using a flux capacitor!

If you look in the developer forum just pick one of the ROMs and you will find info on enabling JIT
Sent from my 2.2.1 Beta4 using XDA App

My score for mflops is always around 39 and my highest is 40+
Sent from my X10i using XDA App

xxcysxx said:
Hello folks,
I just tried to benchmark my x10 with linpack and got only 6.8 or 6.9 mflops. is this normal, is there anything i can do to make it get a higher score?
I'm using the latest .435 firmware and rooted and have all the junk pps from the 'x10 junk apps' thread removed from my phone.
Thanks for any help.
Click to expand...
Click to collapse
whether you updated via flashtool or through seus or pc companion, to optimize your phone is very simple especially if not wanting to root or install custom roms. download latest flash tool, once installed and up and running. make sure your phone is set on debug mode. connect to usb and plug into comp. though you may not want the apps that are in the custom folder. hit the customize button, you can remove them later. you will now notice that the optimize button is now highlighted in black. click on that button and let it install the files. This installs JiTv2 Files. this will bring mflops up and quad scores up. it however will not give very high scores. quad will be around 800-850 tops. mflops will be increased to about 20-30 maybe 35. notice on quad when running the rotating planets test. you might hit 33fps!!! instead of the supposed 30 that its only suppose to get. if you run neocore test you might only peg and 25-28. Good Luck!!

-OP
try running the test again after the unit has settled down (no activity for a while). you might get different results.

hit the customize button, you can remove them later. you will now notice that the optimize button is now highlighted in black. click on that button and let it install the files. This installs JiTv2 Files
Click to expand...
Click to collapse
Hi SErooted,
thank you for your helpfull respond. i pressed the customize and then the optimize button like you said and the phone rebooted and the result was better. linpack now get about 32 and some change, quadrant about 757, and neocore about 21fps, shouldn't neocore get better frame rate than this?
can i get it even faster than this?
i have heard about clearing the dalvik chache when installing JIT for the first time. well i didn't clear any cache when i press that customize or optimizer button. did it do it automatically or do i still have to do it manually?
what i wanted was to have the interface and games run fluid and smooth. like no studder at all.
does the processor throttle down when the system is at idle?
should i overclock?
also, i have notice the video recording in hd mode have less studdering now. though its still there, but less. do you think it probably have anything to do with the memory card performance. i'm still using the stock 8gb that originally came with the phone.
My score for mflops is always around 39 and my highest is 40+
Click to expand...
Click to collapse
hello Michael7168,
thank you for your respond to my message. can you please elaporate more on your statement. is this an x10 you using? i would like to get this result too.
furthermore, i prefer not to use custom firmware at the moment as i use this phone frequently day-to-day and i need to have full functionality of it and reliably. custom firmware are still buggy and lacks some functionality of the phone that i still find usefull for my daily task. but if there is a better custom firmware that really does utilize all of the phone functionality, then please point me in that direction.

xxcysxx said:
Hi SErooted,
thank you for your helpfull respond. i pressed the customize and then the optimize button like you said and the phone rebooted and the result was better. linpack now get about 32 and some change, quadrant about 757, and neocore about 21fps, shouldn't neocore get better frame rate than this?
can i get it even faster than this?
i have heard about clearing the dalvik chache when installing JIT for the first time. well i didn't clear any cache when i press that customize or optimizer button. did it do it automatically or do i still have to do it manually?
what i wanted was to have the interface and games run fluid and smooth. like no studder at all.
does the processor throttle down when the system is at idle?
should i overclock?
also, i have notice the video recording in hd mode have less studdering now. though its still there, but less. do you think it probably have anything to do with the memory card performance. i'm still using the stock 8gb that originally came with the phone.
hello Michael7168,
thank you for your respond to my message. can you please elaporate more on your statement. is this an x10 you using? i would like to get this result too.
furthermore, i prefer not to use custom firmware at the moment as i use this phone frequently day-to-day and i need to have full functionality of it and reliably. custom firmware are still buggy and lacks some functionality of the phone that i still find usefull for my daily task. but if there is a better custom firmware that really does utilize all of the phone functionality, then please point me in that direction.
Click to expand...
Click to collapse
Hello xxcysxx,
yes neocore normally runs higher scores. round about 27-28 avg. i personally havent ran a stock firmware since right around mid december. i have alternated with a few and have finally settled on z.s 2.2b4. and have checked all roms available to find which is best. and custom firmware is sometimes shere you get the desired result. i have done some testing on my x10 with various settings and roms and will be happy to share with you what i have going on. my phone is very responsive. quad scores as of ten minutes ago was 2011 without having overclock widget installed. if yo install overclock widget and max out on screen and off screen min and max to 998mhz, you can achieve slightly higher quad scores. but really not needed if you follow settings below. i have a free app from market sysctl. it works!! not sure if you have read the thread on it yet but if not i will have to post there as i have found perfect settings for performance but at cost of battery drain. install if you have root, if not, the call to root is yours.
if interested.... download sysctl from market.
min free kb: 900000
dirty ratio:500000
dirty background:200000
vfs cache pressure:10
Oom allocating: checked
like i said performance not battery. i know for battery saving settings set these numbers close to what the default are. the lower the better battery, the higher, performance.
if you are really worried about daily performance and custom roms.... when/if you decide. i recommend eminence for 2.1 theme of your choice but i really dig bitsnackers theme its awsome and who wouldnt think a tron looking buzzlightyear isnt cool? its awsome. has a cool bootlogo too. all phone funstionality is there video, camera, flash. EVERYTHING!!
BELOW>>> Optimized, xrecovery, root<<<< ENABLED
if you are interested in a 2.2 rom which really does boost speed on phone and what not. i recommend z's 2.2beta4 as video is not working as of yet you might hold off. if you dont use video you should be ok with camera for now and that works. install gappsfreex10, z's data2sd, and theme of your choice. i like xfx10 cant remember name of creator (sorry) but this theme is great! all of theme are great to be honest this is just the one i use.
2.3 roms are out and they are snappy alone without having any adjustments, however there are things that dont work, camera is really the only thing i can think of right now. but by this time it might be fixed, cant say havent checked status but will when done replying.
you can install and run sysctl on these roms as well and gives great boost.
update: not sure if this is a result of sysctl or problems due to network overrun... mflops as of now are 3.2 i normally run 38+. Mmmmm, I wonder?
update 2: double checked after a little while and low mflops was due to network overload not settings. also once upon reboot from my settings i have a quad score of 2233. screen shot below. mflops picked back up but yet are not as high as usual 34.xx
as well to answer on dalvik cache... yes clear cache and also install cache cleaner from market. i have mine set to clear every twelve hours. when installing custom rom not always a case but recommend to wipe user data as well. a few minutes to reset your phone and make sure everything smooth is worth doing this. as well, clear battery stats.
either way, let me know what you guys think and if these settings help.
SErooted

SErooted said:
Hello xxcysxx,
yes neocore normally runs higher scores. round about 27-28 avg. i personally havent ran a stock firmware since right around mid december. i have alternated with a few and have finally settled on z.s 2.2b4. and have checked all roms available to find which is best. and custom firmware is sometimes shere you get the desired result. i have done some testing on my x10 with various settings and roms and will be happy to share with you what i have going on. my phone is very responsive. quad scores as of ten minutes ago was 2011 without having overclock widget installed. if yo install overclock widget and max out on screen and off screen min and max to 998mhz, you can achieve slightly higher quad scores. but really not needed if you follow settings below. i have a free app from market sysctl. it works!! not sure if you have read the thread on it yet but if not i will have to post there as i have found perfect settings for performance but at cost of battery drain. install if you have root, if not, the call to root is yours.
if interested.... download sysctl from market.
min free kb: 900000
dirty ratio:500000
dirty background:200000
vfs cache pressure:10
Oom allocating: checked
like i said performance not battery. i know for battery saving settings set these numbers close to what the default are. the lower the better battery, the higher, performance.
if you are really worried about daily performance and custom roms.... when/if you decide. i recommend eminence for 2.1 theme of your choice but i really dig bitsnackers theme its awsome and who wouldnt think a tron looking buzzlightyear isnt cool? its awsome. has a cool bootlogo too. all phone funstionality is there video, camera, flash. EVERYTHING!!
BELOW>>> Optimized, xrecovery, root<<<< ENABLED
if you are interested in a 2.2 rom which really does boost speed on phone and what not. i recommend z's 2.2beta4 as video is not working as of yet you might hold off. if you dont use video you should be ok with camera for now and that works. install gappsfreex10, z's data2sd, and theme of your choice. i like xfx10 cant remember name of creator (sorry) but this theme is great! all of theme are great to be honest this is just the one i use.
2.3 roms are out and they are snappy alone without having any adjustments, however there are things that dont work, camera is really the only thing i can think of right now. but by this time it might be fixed, cant say havent checked status but will when done replying.
you can install and run sysctl on these roms as well and gives great boost.
update: not sure if this is a result of sysctl or problems due to network overrun... mflops as of now are 3.2 i normally run 38+. Mmmmm, I wonder?
update 2: double checked after a little while and low mflops was due to network overload not settings. also once upon reboot from my settings i have a quad score of 2233. screen shot below. mflops picked back up but yet are not as high as usual 34.xx
as well to answer on dalvik cache... yes clear cache and also install cache cleaner from market. i have mine set to clear every twelve hours. when installing custom rom not always a case but recommend to wipe user data as well. a few minutes to reset your phone and make sure everything smooth is worth doing this. as well, clear battery stats.
either way, let me know what you guys think and if these settings help.
SErooted
Click to expand...
Click to collapse
holy sh*t. is that benchmark test at the bottom for your X10?

TidusWulf said:
holy sh*t. is that benchmark test at the bottom for your X10?
Click to expand...
Click to collapse
no the one at the top where it says your device. 2233
or how about now after i modded my sysctl settings.screen shot below taken just seconds before this post

xxcysxx said:
Hello folks,
I just tried to benchmark my x10 with linpack and got only 6.8 or 6.9 mflops. is this normal, is there anything i can do to make it get a higher score?
I'm using the latest .435 firmware and rooted and have all the junk pps from the 'x10 junk apps' thread removed from my phone.
Thanks for any help.
Click to expand...
Click to collapse
Install JIT from here.

batman_112 said:
Install JIT from here.
Click to expand...
Click to collapse
Read post on first page

i have alternated with a few and have finally settled on z.s 2.2b4, and have checked all roms available to find which is best. and custom firmware is sometimes shere you get the desired result..
Click to expand...
Click to collapse
thank you for your respond again.
your generic test result has me sold to custom firmware. i want to try it out to see what i'm really missing.
so to your experience, z's 2.2 b4 has all functionality of the device working with only the exception of the video recording? is that correct?
perhaps i can live with that for now. i frequently record my baby playing around the house and playing outside, that's why i prefer to keep the camcorder.
perhaps i will find an old phone laying around in the house just to get by while i'm going on an adventure on custom firmware.
if yo install overclock widget and max out on screen and off screen min and max to 998mhz, you can achieve slightly higher quad scores.
Click to expand...
Click to collapse
will keep this overclocking widget in mind. i just recently learned that you cant clock the processor past 1Ghz without sucessfully cracking the bootloader first. so overclocking is not going anywhere for now.
i have a free app from market sysctl. it works!!
Click to expand...
Click to collapse
yes, i done this already. but still learing on settings.
i have found perfect settings for performance but at cost of battery drain. install if you have root, if not, the call to root is yours.
if interested.... download sysctl from market.
min free kb: 900000
dirty ratio:500000
dirty background:200000
vfs cache pressure:10
Oom allocating: checked
Click to expand...
Click to collapse
on your x10 running at these settings, can you go at least a full day of work on a single charge with frequent text messaging, internet browsing, and phone calls?
no video games or leisure activity involve. full day of work as in taking the phone off the charger in the moning and comming home around eight or nine o'clock in the evening.
i recommend eminence for 2.1 theme of your choice but i really dig bitsnackers theme its awsome and who wouldnt think a tron looking buzzlightyear isnt cool? its awsome. has a cool bootlogo too. all phone funstionality is there video, camera, flash. EVERYTHING!!
Click to expand...
Click to collapse
i'm confused here. is eminence 2.1 a custom firmware or is it a theme for android 2.1 firmware?
if it is a modified sony ericsson 2.1 stock firmware with a different theme, then i'm not interested.
however, if it is a fully ported open source 2.1 android firmware, (not tamper with by sony ericsson) then this is good news. i really hope it is this!
BELOW>>> Optimized, xrecovery, root<<<< ENABLED
Click to expand...
Click to collapse
i have a question about xrecovery.
are you familiar with bare metal restore, acronis true image, or norton ghost.
i use these software to customize my laptop for different occations. whether i use my laptop to play games with my buddies at break hours durring schools, or do homework assignments in the library, or other general work. all i have to do is lay the image at boot and i have a completely different pc. it can be any os using any software customize in anyway, and took about five to ten minutes to lay down the image.
now back to the xrecovery question. can i do a "full backup" or a complete snap shot of the flash of my current android os and then restore it back at a later time in the same state and condition it was in just before i left it?
this is just incase i might have a remorse when i try the custom firmware.
i put a lot of time in making all the settings and adjustments just to the way i like it and i prefer not to have to do it all over again.
I have another question about the JIT software.
how many jit versions are out there?
i'm using the x10 flash tool version 0.2.6 to install the jit software on my phone. do you know what version of jit was distributed with this flash tool version?
thanks again, your time and help is very much appreciated.

your generic test result has me sold to custom firmware. i want to try it out to see what i'm really missing.
Bold step... always worth a shot.
so to your experience, z's 2.2 b4 has all functionality of the device working with only the exception of the video recording? is that correct?
perhaps i can live with that for now. i frequently record my baby playing around the house and playing outside, that's why i prefer to keep the camcorder.
perhaps i will find an old phone laying around in the house just to get by while i'm going on an adventure on custom firmware.
i always keep several older phones for camera use, or other things so this is always a good idea.
will keep this overclocking widget in mind. i just recently learned that you cant clock the processor past 1Ghz without sucessfully cracking the bootloader first. so overclocking is not going anywhere for now.
i agree in holding off, really overclock widget should be utilizes when operating stock 1.6. anything after that 2.1 2.2 2.3 their optimized with jit or are optimized enough without it
on your x10 running at these settings, can you go at least a full day of work on a single charge with frequent text messaging, internet browsing, and phone calls?
no video games or leisure activity involve. full day of work as in taking the phone off the charger in the moning and comming home around eight or nine o'clock in the evening.
these settings, im actually getting stats on this today as i was modding my phone alot yesterday and did not much have chance to let charge until last night. will let you know how well those settings work through out the day.
i'm confused here. is eminence 2.1 a custom firmware or is it a theme for android 2.1 firmware?
if it is a modified sony ericsson 2.1 stock firmware with a different theme, then i'm not interested.
however, if it is a fully ported open source 2.1 android firmware, (not tamper with by sony ericsson) then this is good news. i really hope it is this!
to my knowledge is has not been tampered with by sony, cant recall if the dev had placed any se apks in theme. but it is a great rom. so is blue sparks. if going 2.1 anyway. these are roms and not just themes overlay. so should be as open source as the dev made it.
i have a question about xrecovery, can i do a "full backup" or a complete snap shot of the flash of my current android os and then restore it back at a later time in the same state and condition it was in just before i left it?
this is just incase i might have a remorse when i try the custom firmware.
i put a lot of time in making all the settings and adjustments just to the way i like it and i prefer not to have to do it all over again.
yes you can. when in xrecovery, go to back up and restore. and backup. this will save all settings, phone book, contacts text message, everything! if not happy with custom software you have installed. just go back to xrecovery>backup and restore> restore. back to 2.1 software before the custom rom
I have another question about the JIT software.
how many jit versions are out there?
i'm using the x10 flash tool version 0.2.6 to install the jit software on my phone. do you know what version of jit was distributed with this flash tool version?
the JiT version that is with the new flashtool is JiTv2 i believe.. dont quote me on that. i believe it is as i have looked at the files in the optimizer and compared them to the copy of JiTv2 i have saved on my comp. to my knowledge, there is only two versions. again, dont quote me on it. i havent seen an chatter going on about newer versions so this might be all we get. not sure though
thanks again, your time and help is very much appreciated.[/QUOTE]
always happy to try and help.

http://forum.xda-developers.com/showthread.php?t=935919
Refer this
Cheers!
AdY

SErooted said:
your generic test result has me sold to custom firmware. i want to try it out to see what i'm really missing.
Sorry took so long to kget back to you. hope you catch this.
Sysctl settings really dont make large change in how fast all functions of the phone operate. most increase in performance for settings is shutdown and power on time. as well as a little better scrolling of menus.
basically running the 2.2 software from z or trip, adding the data2sd.zip will get you well into having quad scores of 2+++. if aiming for higher closer to around 3000. which i am very near doing. ill show when i have the proof not till then
anyway, getting mid 2+++ scores and higher, do require using something like set cpu or Overclock widget, in order to get your phone to run at top speed while doing the test..
after that run a test then another and so on until you have achieved what is perfect W/O for you.
Click to expand...
Click to collapse

i did tried the setting parameters you suggested earlier before your latest post and i found the improvement is not that noticable, but the battery did die very quick. i found the original configuration work just fine at this point.
i'm still doing some reading on the custom firmware to make sure i make the proper approach without any hickups. i still have a lot of school work to finish and a part time job so the custom firmware is moving along quite slowly.

[Q] V6 SuperCharger on MyTouch 4G?

It sounds like everyone would benefit from using this script, but there's no specific guidance for the MyTouch 4G. I assume option 8 or 9 would be best since the phone has 512MB of ram.
Are there some roms that we shouldn't be using V6 SuperCharger with?
Link to V6 SuperCharger: http://forum.xda-developers.com/showthread.php?t=991276

You're wrong on 2 points (possibly more):
1) This script is just the same as Autokiller app, with a small addition - it can (or can't) keep launcher in memory. Nothing new and revolutionary. This app exists for a couple of years.
2) This phone has 768MB of RAM. It won't benefit from a low memory killer (or actually, different settings for an existing in OS lowmemkiller), because it has TONS of memory. I just took a look at my phone, ~100MB of memory free, and ~300MB of remaining memory is taken by CACHED apps. If you don't know what it means - please read up on Android memory management, and I'll give you the short version - it's the same as free memory, but better.

It states in the first post:
Also Note: Nothing else does what The V6 SuperCharger does!
................Not AutoKiller Memory Optimizer, Not Auto Memory Manager, Not Minfree Manager...
The Nook Color has 512 MB of ram and people have noticed a big difference using this script on Cyanogenmod.

but since the Glacier has 768MB of ram, you won't notice much change.

saranhai said:
but since the Glacier has 768MB of ram, you won't notice much change.
Click to expand...
Click to collapse
I have found it useful on ports and ROMs that aren't tweaked specifically for this phone. For instance, if TDJ made the ROM, V6 is useless. In fact, it will only hurt.
Sent from my HTC Glacier using XDA App

I use this script on my evo and its great so I just rooted this phone and flashed the mik runny ROM, I was using it for a day and only stayed at about 100mb of RAM so I decided to use v6 with option 8 and now it stays around 200mb and is running super smooth.
IDK if that helps any but I always loved v6 and know a few people that use it on a few different phones and it always works for the better.
Sent from my HTC Glacier using XDA App

Why do people that know nothing or close to nothing about OS internals, decide that they have better knowledge of memory management should be, then OS and phone designers? The same people who don't know the difference between cached and active apps, and the only number they understand is the (useless) amount of free memory? I see it all over the forums, and it amazes me each time. How do people actually try to judge if something works well or not, without getting at least some basic understanding of how the things work?
Oh well, here it comes again:
http://forum.xda-developers.com/showthread.php?t=678205
Read this. Maybe you will understand something.
Few, if any, 512MB phones and no 768MB phones need this script, or any kind of tweaking for lowmemkiller values, especially not since Gingerbread and when not running Sense (which retains the ability to cache apps, removed in Gingerbread mostly to ease running the OS on older devices). The only thing it does, is to make garbage collector work harder and kick in earlier. It doesn't make your phone "smoother", and whoever think it does - should check the meaning of the word "placebo" in the nearest dictionary. The number that stands for "free memory" means something between "close to nothing" and "absolutely nothing".
I know I shouldn't be surprised, people always tend to have strong opinions on everything, even things they sometimes don't know a thing about. But still, it's XDA-Developers, not XDA-Phone-users, so at least something should be done about it. Even if the education attempts will fail, like they mostly do.

Jack_R1 said:
Why do people that know nothing or close to nothing about OS internals, decide that they have better knowledge of memory management should be, then OS and phone designers? The same people who don't know the difference between cached and active apps, and the only number they understand is the (useless) amount of free memory? I see it all over the forums, and it amazes me each time. How do people actually try to judge if something works well or not, without getting at least some basic understanding of how the things work?
Oh well, here it comes again:
http://forum.xda-developers.com/showthread.php?t=678205
Read this. Maybe you will understand something.
Few, if any, 512MB phones and no 768MB phones need this script, or any kind of tweaking for lowmemkiller values, especially not since Gingerbread and when not running Sense (which retains the ability to cache apps, removed in Gingerbread mostly to ease running the OS on older devices). The only thing it does, is to make garbage collector work harder and kick in earlier. It doesn't make your phone "smoother", and whoever think it does - should check the meaning of the word "placebo" in the nearest dictionary. The number that stands for "free memory" means something between "close to nothing" and "absolutely nothing".
I know I shouldn't be surprised, people always tend to have strong opinions on everything, even things they sometimes don't know a thing about. But still, it's XDA-Developers, not XDA-Phone-users, so at least something should be done about it. Even if the education attempts will fail, like they mostly do.
Click to expand...
Click to collapse
I do understand that in some cases you don't need a memory manager like if your running a stock ROM or an aosp ROM that doesn't take up as much memory.
Now I haven't had this phone long enough to say if this script is all that good for this phone but I know on the evo running a sense 3.0-3.5 ROM that wasn't meant for the phone and hugs up every little bit of memory that the phone has to offer, this scrip makes those ROMs usable.
Without it or something like it the phone can't handle doing simple tasks like using an app without fc something else like the launcher.
So you could say what you want and yes maybe this phone doesn't need it since it has more RAM and ROM but I'll still try things like this to try and see if it will better the phones performance.
Sorry if that doesn't make sense I'm still half asleep.
Sent from my HTC Glacier using XDA App

The instances where i have noticed that this works is while doing benchmarking with quadrant. It has shown increased framerates for me after running the script and I also get higher scores on quadrant, about 500-1k more than without. I dont know if its usefull for much other than benchmarking though. I think the phone runs fine without it though.

What you have to understand is performance is not measured via syntactic benchmarks (ex: Quadrant). The biggest issue with people is that they don't know enough to know that they don't know, so they compare it with silly numbers (ex: score) they can't comprehend what they see, much less put numbers behind real life activities that's not applicable in controlled environment.
Now far as V6-SC script goes its almost obsolete now due to few things. 1) Hardware advancement where now minimum spec requirements for "SmartPhone" are 1ghz single core proc with 512mb ram. But so called "SuperPhone" now has dual core 1ghz-1.5ghz with 768mb-1gb ram. So it make no sense as we don't use 256-566mhz proc with 64-256mb ram because we are more then enough hardware adequate for heavy daily usage. 2) OS development which elements most of it as hardware is more and more powerful. But on software level mostly all custom base rom (ex: CyanogenMod) is highly optimized and tweaked to run on optimal performance.
Now is it all placebo effect? Mostly, but not all. But does it mean it can't be tweaked any further? (Rhetorical) No. How do I know? We (scope outside of XDA) tweaked it to the next level. How you ask?
1) Optimized ext4fs: reduced r/w rate (healthy NAND lifespan), improved journaling (corrupted data writeback integrity) = Which improves the IOPs and performance access rate.
2) HC3.x fugu binaries, patched sqlite libraries, mSD read ahead buffer fix.
3) Modified VM: OOM (Out Of Memory), LMK (Low Memory Killer), VM heap (Virtual Machine), DRA (Dirty Ratio), DBR (Dirty Background Ratio), DWC (Dirty Writeback Centisecs), DEC (Dirty Expired Centisecs), SWP (Swap), VCP (VFS Cache Pressure).
4) Increased minfree value: Background, Foreground, Empty, Hidden, Visible, Secondary, Content.
5) Optimized cache: File and Drop cache, Forced cache (resident loop).
6) Custom kernel: OC/UC, UV/SVS/VDD, BFS/CFS, RSU/VR/SP supported.
7) Custom ROM: Optimized Rom script and props (ex: CyanogenMod).
I bet my superior MT4G can own your inferior MT4G. Cuz you can't touch this as its tweaked to THE next level. I'll stick with AOSP2.3.7GB until ICS4.X is more stable and we understand more as most memory grouping and adjustments might be changed.
Sent from my HTC Glacier

Jack_R1 said:
You're wrong on 2 points (possibly more):
1) This script is just the same as Autokiller app, with a small addition - it can (or can't) keep launcher in memory. Nothing new and revolutionary. This app exists for a couple of years.
2) This phone has 768MB of RAM. It won't benefit from a low memory killer (or actually, different settings for an existing in OS lowmemkiller), because it has TONS of memory. I just took a look at my phone, ~100MB of memory free, and ~300MB of remaining memory is taken by CACHED apps. If you don't know what it means - please read up on Android memory management, and I'll give you the short version - it's the same as free memory, but better.
Click to expand...
Click to collapse
Obviously, you never tried it lol
Here... you may learn something new...
http://www.rt-embedded.com/blog/archives/linux-memory-consumption/
http://forum.xda-developers.com/showpost.php?p=20163493&postcount=6695
Below a certain threshold of free ram (ie. not enough cached), the device WILL gag...
Hundreds if not thousands of users with 1 GB ram devices use it (Atrix, SGSII, etc.) and I know your phone stutters from time to time with a slight delay when pressing buttons from time to time since that's what my friend's Atrix does.
In fact, the biggest difference he notices is in the use of google maps... never a stutter.
So you're missing out.

zeppelinrox said:
[1] Obviously, you never tried it lol
[2] I know your phone stutters from time to time with a slight delay when pressing buttons from time to time since that's what my friend's Atrix does.
[3] So you're missing out.
Click to expand...
Click to collapse
First of all as dev of V6-SC you would be very defensive but at same times your not charging money to normal folks for it is a good thing, so thank you. Which I can say less about other folks editing same value claiming it new. Now I don't know about Jack but let's be clear on few points.
1) I did try your so called script and didn't like the whole script manger + busybox cast AFTER the OS startup. Which normally you can achieve via daemon or init.d script after kernel is initialized by declaring and using native shell. So no need for force apply afterwards as it was utilized before it was initiated via script manager. Also V6-SC couldn't keep the selected category minfree value which changed. But in short I didn't notice anything revolutionary as it was fully optimized long before I randomly landed on Android General section and saw your post claiming it maximize the devices performance. Which I was spectacle about as from your post you did seem to have basic knowledge hopefully not from wiki/google but *nix usr exp before landing on to Android.
2) Like I said I don't know about Jacky Boy but I can GRANTEE you I have NEVER had this so called "button delay" you specified. But I did modify the sampling rate and pressure density accommodated by tweaking transition speed. But now I run min:368mhz/max:1027mhz/gov:SmartAssV2. But even when I was battery conscience before I had MP1650mAh I ran on min:230mhz/max:768mhz/gov:SmartAssV1 with custom -75 to -100 VDD using ~14mA idle and ~60-90mA active per unit scale. I never had lag with 200mb used RAM running at least 18-20pcs and 14-15svc. So what your friend is running (Atrix) is irrelevant also isolated.
3) O-RLY am I really missing out? I think ill stick to my own. But don't take this post personal as it was ment for it to be argumentative. Difference is I actually know what I'm talking about as I have strong backgrounds on...
Sent from my HTC Glacier

zeppelinrox said:
Obviously, you never tried it lol
Here... you may learn something new...
http://www.rt-embedded.com/blog/archives/linux-memory-consumption/
http://forum.xda-developers.com/showpost.php?p=20163493&postcount=6695
Below a certain threshold of free ram (ie. not enough cached), the device WILL gag...
Hundreds if not thousands of users with 1 GB ram devices use it (Atrix, SGSII, etc.) and I know your phone stutters from time to time with a slight delay when pressing buttons from time to time since that's what my friend's Atrix does.
In fact, the biggest difference he notices is in the use of google maps... never a stutter.
So you're missing out.
Click to expand...
Click to collapse
I didn't need to try it to know. I tried Autokiller, I played with lowmemkiller settings and watched the results, and I did it on Nexus One with 512MB of memory. It never needed anything since Gingerbread, and unless I made the settings super-aggressive, Autokiller actually failed to make any difference whatsoever - the apps were killed based on their age and never dropped by replacing apps.
In the current system, I have 100MB free + 250MB cached apps (which is just the same as free - theoretically and practically). The main difference you're not accounting for is - Android isn't a Linux distro, it's a Linux-derived OS, with many changes for mobile activity, especially on the kernel level, especially in the memory management area. "Linux memory consumption" isn't Android memory consumption, since they manage things differently. Linux isn't build to kill running apps, its lowmemkiller can't do it. Linux doesn't have concurrent garbage collector. Many Linux examples are irrelevant. Cached apps in Android aren't cached pages in Linux, freeing cached pages in Linux isn't killing cached apps in Android, and the most important - "performance degradation" doesn't exist in Android, since you ALWAYS have enough memory for any size of task (the largest loading task requires 50MB of memory, and there's 100MB free on my phone), and concurrent garbage collection is ALWAYS present in the system, the only thing you're doing - is calling it earlier, making it actually work more and getting the system more laggy than it could be.
I understand that you want to protect your creation, but in this case, you're wrong, sorry. You won't convince me.
And yes, I don't know what "button lag" are you talking about.

HTC Glacier said:
First of all as dev of V6-SC you would be very defensive but at same times your not charging money to normal folks for it is a good thing, so thank you. Which I can say less about other folks editing same value claiming it new. Now I don't know about Jack but let's be clear on few points.
1) I did try your so called script and didn't like the whole script manger + busybox cast AFTER the OS startup. Which normally you can achieve via daemon or init.d script after kernel is initialized by declaring and using native shell. So no need for force apply afterwards as it was utilized before it was initiated via script manager. Also V6-SC couldn't keep the selected category minfree value which changed. But in short I didn't notice anything revolutionary as it was fully optimized long before I randomly landed on Android General section and saw your post claiming it maximize the devices performance. Which I was spectacle about as from your post you did seem to have basic knowledge hopefully not from wiki/google but *nix usr exp before landing on to Android.
2) Like I said I don't know about Jacky Boy but I can GRANTEE you I have NEVER had this so called "button delay" you specified. But I did modify the sampling rate and pressure density accommodated by tweaking transition speed. But now I run min:368mhz/max:1027mhz/gov:SmartAssV2. But even when I was battery conscience before I had MP1650mAh I ran on min:230mhz/max:768mhz/gov:SmartAssV1 with custom -75 to -100 VDD using ~14mA idle and ~60-90mA active per unit scale. I never had lag with 200mb used RAM running at least 18-20pcs and 14-15svc. So what your friend is running (Atrix) is irrelevant also isolated.
3) O-RLY am I really missing out? I think ill stick to my own. But don't take this post personal as it was ment for it to be argumentative. Difference is I actually know what I'm talking about as I have strong backgrounds on...
Sent from my HTC Glacier
Click to expand...
Click to collapse
Sounds like you never got it working properly.
Also, if you have init.d support no need to run anything on boot with script manager.
Maybe the rom's kernel was applying settings late.
And no my friends atrix is not isolated there is a rather big thread in the atrix forums.
SGSII users see benefits too so seems there is always room for improvement.
Jack_R1 said:
I didn't need to try it to know. I tried Autokiller, I played with lowmemkiller settings and watched the results, and I did it on Nexus One with 512MB of memory. It never needed anything since Gingerbread, and unless I made the settings super-aggressive, Autokiller actually failed to make any difference whatsoever - the apps were killed based on their age and never dropped by replacing apps.
In the current system, I have 100MB free + 250MB cached apps (which is just the same as free - theoretically and practically). The main difference you're not accounting for is - Android isn't a Linux distro, it's a Linux-derived OS, with many changes for mobile activity, especially on the kernel level, especially in the memory management area. "Linux memory consumption" isn't Android memory consumption, since they manage things differently. Linux isn't build to kill running apps, its lowmemkiller can't do it. Linux doesn't have concurrent garbage collector. Many Linux examples are irrelevant. Cached apps in Android aren't cached pages in Linux, freeing cached pages in Linux isn't killing cached apps in Android, and the most important - "performance degradation" doesn't exist in Android, since you ALWAYS have enough memory for any size of task (the largest loading task requires 50MB of memory, and there's 100MB free on my phone), and concurrent garbage collection is ALWAYS present in the system, the only thing you're doing - is calling it earlier, making it actually work more and getting the system more laggy than it could be.
I understand that you want to protect your creation, but in this case, you're wrong, sorry. You won't convince me.
And yes, I don't know what "button lag" are you talking about.
Click to expand...
Click to collapse
I was tempted to stop reading when you admit to not even using it.
If it did the same as AKMO or Auto Memory Manager why on earth would anybody bother.
I sure as hell wouldn't bother writing a 4500+ line script lol.
I totally agree that Android memory is not the same as linux (see my sig) but the similarities are there and the article I posted applies 100%.
Its not about free ram.
Its about the right balance.
In fact, many report LESS free ram, ie. better multitasking, along with better performance and smoother performance.
Because I don't think Android memory should work the same as linux memory either.
Also, you tried AKMO because you felt there could be improvement and it didn't work.
THAT'S why I wrote a 4500+ line script that blows anything else out of the water

zeppelinrox said:
Sounds like you never got it working properly.
Also, if you have init.d support no need to run anything on boot with script manager.
Maybe the rom's kernel was applying settings late.
And no my friends atrix is not isolated there is a rather big thread in the atrix forums.
SGSII users see benefits too so seems there is always room for improvement.
Click to expand...
Click to collapse
Well my highly optimized as is but I am aware of V6 and others using it but personally I would stick to my.
Sent from my HTC Glacier

[Q] Nexus 10 using 1gb+ RAM at all times?

Doesn't matter what Rom I'm on, cleanrom, aokp, stock, even after I close all apps and services all task managers tell me I am using at least 1gb
ALWAYS
Is this normal?

Use the search function on these forums

I'd try 4.2.2 stock, I have around 400-500MB usage idle. On 4.2.1 (stock and custom) I had way higher usage though (700MB-1GB).
On 4.2.2 though I seem to have only 1.2GB of usable RAM whereas 4.2.1 had 1.6GB

I'm no pro, but if RAM 'disappeared' maybe it got reserved as GPU memory? And the GPU using it's own pool now would explain why you have less used up in general.

bee55 said:
I'm no pro, but if RAM 'disappeared' maybe it got reserved as GPU memory? And the GPU using it's own pool now would explain why you have less used up in general.
Click to expand...
Click to collapse
Well the N10 is advertised to have 2GB of RAM. Prior to 4.2.2, we had 1.6GB usable, with the other 400MB or so (from what I hear) reserved for the GPU already. Now after 4.2.2, there's 1.2GB usable.
May as well just sold it as a 1GB device at this point... lol

espionage724 said:
Well the N10 is advertised to have 2GB of RAM. Prior to 4.2.2, we had 1.6GB usable, with the other 400MB or so (from what I hear) reserved for the GPU already. Now after 4.2.2, there's 1.2GB usable.
May as well just sold it as a 1GB device at this point... lol
Click to expand...
Click to collapse
Lol but almost all devices usable memory for the user is about half of the advertised amount
As you can see my 1gb ram HTC amaze only has about half usable
Sent from my Amaze 4G using xda app-developers app

filthygoat223996 said:
Lol but almost all devices usable memory for the user is about half of the advertised amount
As you can see my 1gb ram HTC amaze only has about half usable
Sent from my Amaze 4G using xda app-developers app
Click to expand...
Click to collapse
Hmm, the only other device that comes close to this behavior that I've seen is the Galaxy Tab 2 7.0 (stated 1GB, has around 600MB usable). The display isn't nearly HD on that (1024x600 I think) so I really can't think of an actual reason for the 400MB loss there.
A Galaxy Prevail I have shows about 235MB usable, with 384MB stated.
Can't recall numbers on the Nexus 7 or Galaxy Tab 7 though, but I really don't recall memory usage/free being of a concern.

espionage724 said:
Well the N10 is advertised to have 2GB of RAM. Prior to 4.2.2, we had 1.6GB usable, with the other 400MB or so (from what I hear) reserved for the GPU already. Now after 4.2.2, there's 1.2GB usable.
May as well just sold it as a 1GB device at this point... lol
Click to expand...
Click to collapse
And the device does have 2gb RAM. You should know by now that the RAM in phones and tablets is versatile and used for more than just user space.
One could easily argue the counterpoint of yours, which is that 2gb of RAM means more for the GPU while still keeping a respectable amount for the user. But obviously, you need to do your research when buying.
Sent from my Nexus 10 using xda premium

With some simple init.d and sysctl.conf tweaks you can have complete control over how the RAM is used. Search is your friend. Or at least it should be. LOL

It has to share ram with the GPU. This is monster resolution screen... Did you think it would have as much ram usable as a 720p screen? There's a lot going on at all times with this hardware because of the screen. Remember that ram is worthless if its not being used.
Sent from my Nexus 4 using xda premium

They just released the updated kernel source for 4.2.2 for the N10 and there is a change where they up the ram reserved for the GPU by 384mb. The comment says they did this as otherwise when the GPU needed more ram it would cause fragmentation of the user space ram.
Geoff
Sent from my Nexus 10 using Tapatalk HD

Sounds reasonable enough.

...uhm...
styckx said:
Sounds reasonable enough.
Click to expand...
Click to collapse
After having upgraded to 4.2.2... I have the same behaviour ....ram decreased to 1.3...in total before with 4.2.1 was 1.6...something strange, if the extra ram has been allocated for gpu....why with 4.2.1 wasn't like now?...in my opinion...they didn't free up the ram after upgrading..it's just locked...
I don't see notable differences between 4.2.2 and 4.2.1.... So that strange...could it be a bug?

borgy said:
After having upgraded to 4.2.2... I have the same behaviour ....ram decreased to 1.3...in total before with 4.2.1 was 1.6...something strange, if the extra ram has been allocated for gpu....why with 4.2.1 wasn't like now?...in my opinion...they didn't free up the ram after upgrading..it's just locked...
I don't see notable differences between 4.2.2 and 4.2.1.... So that strange...could it be a bug?
Click to expand...
Click to collapse
The answer to why the total ram availible has decreased has been solved - Read gzub post.
Called it re-allocated, free'd up or locked, The GPU now has 800mb in 4.2.2 all to its self, where as before it had half that. So in running apps you will only see a total of around 1.2gb.
moving on...
I have a similar issue as the OP.
Sitting idle its running at around 400-500mb used but can, and often does, spike up to over a gb which causes it to not function correctly (Chrome especially seems to throw a massive wobbler and never stay open). I have nothing different installed than I have on my N7 or S3. I mainly use it for browsing, email. If I check running apps there is nothing else there above what was there when it was idling at around 400-500mb.
I have tried a factory reset and the issue still persists. Next step is to roll back to 4.2.1 to see if I can get the stability back as I didn't have these issues on 4.2.1.

VanCatRabbit said:
The answer to why the total ram availible has decreased has been solved - Read gzub post.
Called it re-allocated, free'd up or locked, The GPU now has 800mb in 4.2.2 all to its self, where as before it had half that. So in running apps you will only see a total of around 1.2gb.
moving on...
I have a similar issue as the OP.
Sitting idle its running at around 400-500mb used but can, and often does, spike up to over a gb which causes it to not function correctly (Chrome especially seems to throw a massive wobbler and never stay open). I have nothing different installed than I have on my N7 or S3. I mainly use it for browsing, email. If I check running apps there is nothing else there above what was there when it was idling at around 400-500mb.
I have tried a factory reset and the issue still persists. Next step is to roll back to 4.2.1 to see if I can get the stability back as I didn't have these issues on 4.2.1.
Click to expand...
Click to collapse
Sounds you might be suffering from the same thing as I have since I got my N10.
I don't know how many have this or has noticed it but after a while the amount of free memory just starts disappearing and no matter what you do, you don't seem to be able to recover it unless you reboot. And when the memory gets low enough, the device soft reboots itself. I still haven't found the cause of this but I have finally found where all the memory is going. It turns out that it's the surfaceflinger process that hogs the memory. I started checking when I was running low on memory and surfaceflinger was at that point using over 400MB of it. My suspicion was confirmed when I rebooted and noticed that it went down to around 25MB and now after a while of usage, it has slowly climbed to 50MB.
It would be great if some developer could take a look at this in order to try to fix it.

CazeW said:
Sounds you might be suffering from the same thing as I have since I got my N10.
I don't know how many have this or has noticed it but after a while the amount of free memory just starts disappearing and no matter what you do, you don't seem to be able to recover it unless you reboot. And when the memory gets low enough, the device soft reboots itself. I still haven't found the cause of this but I have finally found where all the memory is going. It turns out that it's the surfaceflinger process that hogs the memory. I started checking when I was running low on memory and surfaceflinger was at that point using over 400MB of it. My suspicion was confirmed when I rebooted and noticed that it went down to around 25MB and now after a while of usage, it has slowly climbed to 50MB.
It would be great if some developer could take a look at this in order to try to fix it.
Click to expand...
Click to collapse
Sorry for my ignorance, but what is surfaceflinger? I don't see anything similarly named on my device.
I've just downloaded 'usage timelines free' to monitor memory usage in more detail on all my devices and compare.
I'll try to update if I see anything interesting.

VanCatRabbit said:
Sorry for my ignorance, but what is surfaceflinger? I don't see anything similarly named on my device.
I've just downloaded 'usage timelines free' to monitor memory usage in more detail on all my devices and compare.
I'll try to update if I see anything interesting.
Click to expand...
Click to collapse
It's a system process out of /system/bin/. Check your app to see if it's hiding system processes. Most of the process viewing apps/loggers hide system processes by default.
SurfaceFlinger is Androids compositor: http://stackoverflow.com/questions/5699107/android-surfaceflinger
SurfaceFlinger shouldn't stay open though. Everytime I've seen it in OS Monitor the process lasts less than a second and disappears. Something he's running, or the ROM he's using is keeping it open for whatever reason..

styckx said:
It's a system process out of /system/bin/. Check your app to see if it's hiding system processes. Most of the process viewing apps/loggers hide system processes by default.
SurfaceFlinger is Androids compositor: http://stackoverflow.com/questions/5699107/android-surfaceflinger
SurfaceFlinger shouldn't stay open though. Everytime I've seen it in OS Monitor the process lasts less than a second and disappears. Something he's running, or the ROM he's using is keeping it open for whatever reason..
Click to expand...
Click to collapse
I noticed it with the "ps" terminal command. Currently I'm running Aokp but before that I was running stock. I'm suspecting the aosp browser or flash being the cause but I can't confirm it.
EDIT: surfaceflinger is shown in OS Monitor all the time but with ps I can see the amount of memory it is using.

After some monitoring I found nothing. I did a full refresh install of 4.2.2 from the factory image and have installed nothing on top other than applying updates for what comes as standard. After two days of browsing it started to slow down again and show as low memory. I installed a terminal and ran 'PS surfaceflinger', it was showing as using over 600mbs. Stupidly I didn't take a screen dump, but I imagine it will happen again.
Anyone able to shed any light on what's going on here?

Having the same issue
I too have seen surfaceflinger using almost 600MB of RAM. My used RAM was over 1GB and apps were running very sluggish. My SwiftKey keyboard constantly does a soft restart in the middle of typing when the tablet is in this state as well as other apps acting as others have described in this thread. Time for a fix Google!

Dalvik vs. ART - Benchmark Results on the Droid X

I've never seen benchmark results for the DX between Dalvik and ART, so I decided to do it myself. I'm running @Aaahh, @palmbeach05 and @m.ksy's latest CM11 build without GApps, and I ran the test with AnTuTu.
To ensure the integrity of the results, I wiped the caches before testing each runtime and disabled background processes via the developer options. Time was allowed after each boot for it to finish applying Aero Control's settings and everything else. (Thanks to @Oggie7797 for those tips)
All CPU frequencies are stock except the last one - 1.2GHz at 68vsel. The GPU is overclocked at 266MHz (default is 200MHz). Nothing was changed between tests aside from the runtime. The first two screenshots are for ART, the last two are for Dalvik.
The winner is -- ART!
But not by a miraculous amount. In my experience with ART, it hangs occasionally but does offer slightly better performance. Although, as glitchy as it is, as well as the number of apps that simply will not run on it, I'd stick to Dalvik. It's default on Kitkat for a reason. Only use it if you don't care about said apps and the occasional crash.
If you're like me and @Oggie7797 and have been wondering which one you should use (or otherwise found this info useful), please hit "Thanks"

@argentwarrior20 I still like Dalvik cuz I'm an Xposed fan but nice post!
Edit:
Unless I'm mistaken, @sd_shadow didn't develop any parts of cm11 but he has done some amazing work helping out the community!! If anyone should be given credit for cm11 I'd tag @palmbeach05 since he done a lot at the beginning.
Tap'd from my Unicorn covered Bionic

Low memory killer values

Hello my nexus 6 always slow down after days of use, too much apps on memory and low free ram, so I started stuying other devices oom settings, and found that two of fastest devices I can observe have this settings, so try it if you want , for me these fixed slugginess after days without reboot.

lol!
btw, rebooting your device every once in a while will also keep it fast.

fedef12evo said:
Hello my nexus 6 always slow down after days of use, too much apps on memory and low free ram, so I started stuying other devices oom settings, and found that two of fastest devices I can observe have this settings, so try it if you want , for me these fixed slugginess after days without reboot.
Click to expand...
Click to collapse
which kernel are you using ?
normaly the stocl lmk values are good enough to not be laggy

Dead-neM said:
which kernel are you using ?
normaly the stocl lmk values are good enough to not be laggy
Click to expand...
Click to collapse
All the kernels have the same lmk values, that are too low

fedef12evo said:
All the kernels have the same lmk values, that are too low
Click to expand...
Click to collapse
yes these are the stock value for 3gb of ram i guess but they're not too low. the android will kill old apps when necessary to free up the ram.
with gravitybox i can see that on recent tab my memory can be around 130mb free ( not feeling laggy) and when reopen the recent tab again there is around 400mb free and i don't touch lmk value
i think it's a kernel related problem this is why i ask which one he's using.

fedef12evo said:
All the kernels have the same lmk values, that are too low
Click to expand...
Click to collapse
No... not really...
I highly doubt it truly has to do with your lmk... and probably more to do with vmpressure
I bet you have that either set too low... or its broken all together... which means garbage collections isn't happening when it's supposed to...
If you were to drop_cache... your lag would be gone...
Messing with lmk without proper knowledge just leads to an unstable device

rignfool said:
No... not really...
I highly doubt it truly has to do with your lmk... and probably more to do with vmpressure
I bet you have that either set too low... or its broken all together... which means garbage collections isn't happening when it's supposed to...
If you were to drop_cache... your lag would be gone...
Messing with lmk without proper knowledge just leads to an unstable device
Click to expand...
Click to collapse
Maybe yes you are right, I am using aicp rom that has some tweaks about swappiness and vm cache pressure, what are the best values?

fedef12evo said:
Maybe yes you are right, I am using aicp rom that has some tweaks about swappiness and vm cache pressure, what are the best values?
Click to expand...
Click to collapse
Swappiness: no swap file(why do you need one with 3 GB of ram?) No change...
Vmpressure:
This is kinda grey...
There are 2 schools of thought...
HIGH: run between 100 and 200... your device will drop caches as necessary... with error on the side of... dump early dump often...
LOW: run at like 20... or even 0... but then run a cron.d job every 2 hours or so to "manually" drop caches... and when you realize you can make it longer between drops... you do...
Pros & Cons... I have no bloody idea...
Me personally... I like running between 150 and 200... it keeps my phone happy enough...

rignfool said:
Swappiness: no swap file(why do you need one with 3 GB of ram?) No change...
Vmpressure:
This is kinda grey...
There are 2 schools of thought...
HIGH: run between 100 and 200... your device will drop caches as necessary... with error on the side of... dump early dump often...
LOW: run at like 20... or even 0... but then run a cron.d job every 2 hours or so to "manually" drop caches... and when you realize you can make it longer between drops... you do...
Pros & Cons... I have no bloody idea...
Me personally... I like running between 150 and 200... it keeps my phone happy enough...
Click to expand...
Click to collapse
On aicp stock value is swap 0 and cache pressure 20 , so I am going to try 200 thank you

i used to set my low memory killer values, a long time ago. but ive found that rebooting every day or two does a better job.

I know it's not exactly on topic but what about scrolling cache? I hear if you disable it, scrolling in the UI is glassy smooth.
I don't know how to disable except I did see a setting once when I was running Dirty Unicorns.
Don't know if this was only valid on previous OS version or if it still works for MM.
Sent from my Nexus 6 using XDA Labs

MrBrady said:
I know it's not exactly on topic but what about scrolling cache? I hear if you disable it, scrolling in the UI is glassy smooth.
I don't know how to disable except I did see a setting once when I was running Dirty Unicorns.
Don't know if this was only valid on previous OS version or if it still works for MM.
Sent from my Nexus 6 using XDA Labs
Click to expand...
Click to collapse
Already tried, without scrolling cache, device is very laggy, dont know why, but having it enabled is better