[KERNEL][AOSP/TW][27 Dec] War Kernel r3 // colder than ever

[gleb_grid]

@GeeckoDev , I afraid yet use TRIM, is it FULL safety for my note?

 

[bladecgn]

I noticed something but it's not that big, do you remember back on your R7 build there was a post regarding microstutter I'm having that issue. I read back on your post before that you actually asked to push the gralloc for slimsaber will that be the same reason why there's that small lag? but everything is good it's actually not a big problem. will do some testing for the next days to check if there would be any major problems. Again good job and thank you for the hard work! :good:

Same here.. slight microstutter with slimsaber (that is not there when using slimsabers built in kernel), otherwise perfect, thank you!!

 

[kazuna69]

@GeeckoDev,
sorry,
would it be possible to make a r9 version for slimkat/slimsaber ?
sorry if it too much to ask for,
I really appreciate your effort :highfive: :victory:

 

[f0xy]

@GeeckoDev do you have any plans with the i9300 pal?

Sorry for offtopic N7000 users x

 

[7esain]

does r9 works on NightOwl BETA3 ??

 

[GeeckoDev]

@GeeckoDev , I afraid yet use TRIM, is it FULL safety for my note?

Yes. I've got 400 downloads and zero complaints.

Same here.. slight microstutter with slimsaber (that is not there when using slimsabers built in kernel), otherwise perfect, thank you!!

I'll look into that. Maybe the slimsaber kernel performs better because of many compiler optimizations I shoud take a look at.

would it be possible to make a r9 version for slimkat/slimsaber ?

I believe that the Omni flavor works with slimsaber.

@GeeckoDev do you have any plans with the i9300 pal?

I'm not sure what you're talking about? HWC?

does r9 works on NightOwl BETA3 ??

I guess not, you should upgrade to beta4.

 

[kazuna69]

Thanks! :victory:
Will try it soon!~ :silly:

edit: tested, getting bootloop, reinstall default kernel/recovery

 

[dev_freak]

@GeeckoDev what can I use fstrim for? Can I use it on another kernel too?

 

[zeu55]

i tried forest v4 , bauners stock beta 4 and raw9 kernel on beta 4 rom. on raw r9 kernel i am getting more vibrant display colour compared to other two. may be for the refresh rate ...secondly my battery is working great with this kernel . trim is working. overall smooth experience on beta 4 .

sent from N7000 powered by bauner's CM12 beta 4

 

[GeeckoDev]

i tried forest v4 , bauners stock beta 4 and raw9 kernel on beta 4 rom. on raw r9 kernel i am getting more vibrant display colour compared to other two. may be for the refresh rate ...secondly my battery is working great with this kernel . trim is working. overall smooth experience on beta 4 .

sent from N7000 powered by bauner's CM12 beta 4

That's because of the s3cfb update that hasn't been merged to CyanogenMod kernels.

 

[I_android]

@GeeckoDev
CM12 beta4 kernel has OC oled as well?
And any specific reason to overclock to 59hz and not 60hz?

 

[User_99]

@GeeckoDev what can I use fstrim for? Can I use it on another kernel too?

For speed up your file system - like a defragmentation. Newer kernel do this automatically so no need to do that manually. For our note only few kernel have trim feature included...

 

[dev_freak]

For speed up your file system - like a defragmentation. Newer kernel do this automatically so no need to do that manually. For our note only few kernel have trim feature included...

And can I use it on as example forest kernel if I want to?

 

[User_99]

Yes

 

[dev_freak]

Yes

Thanks. I think xda is the best forum for help and modding etc.

 

[GeeckoDev]

@GeeckoDev
CM12 beta4 kernel has OC oled as well?
And any specific reason to overclock to 59hz and not 60hz?

Because 60hz makes the screen flicker at boot and breaks S-Pen. On Touchwiz only the second issue appears.

 

[KalZakath]

@GeeckoDev

are you still workin' at kernel with i9300 sorcess?
Any progress?

 

[I_android]

Because 60hz makes the screen flicker at boot and breaks S-Pen. On Touchwiz only the second issue appears.

thanks for explanation
Does CM12 beta4 kernel has OC oled as well?

 

[I_android]

@bauner @GeeckoDev @forest1971 @Lanchon
Here I am attaching screenshots from AndroBench app benchmark results with bauner, GeeckoDev and forest1971 kernels.
But before that I'd summarize my findings:
1. With bauner's stock kernel the scores are good before Trimming (and after Trimming on forest1971 and GeeckoDev Trim enabled kernels), I did not notice any huge difference in the scores before and after trimming.
2. Scores seems to be poor with forest1971 and GeeckoDev kernels before and after trim.
3. Not only scores, I also noticed that on forest1971 kernel apps took long time to install as compared to bauner's stock kernel.
4. Well I know all respected devs may say that I am the one who won't gain from trimming and may be trimming is not for me. But my question is that why trim enabled kernels (forest1971 and GeeckoDev) are giving so low scores? They should at least give equally good scores as bauner's kernel.
Is there anything that forest1971 and GeeckoDev need to look into their kernels.
Could @forest1971, @GeeckoDev and all the amazing people checking this thread please verify these results on your device and report to devs so that we can have enough stats to conclude.
Thanks a lot for all the amazing work done by @bauner @forest1971 @GeeckoDev and @Lanchon.

Screenshot from @forest1971 kernel after trimming on beta3:


Screenshot from @bauner stock kernel after returning from forest1971 kernel on beta3 (there were almost similar results on bauner stock kernel before installing forest1971 kernel and trimming):


Screenshot from @GeeckoDev kernel after trimming on beta4:

 

[Lanchon]

@bauner @GeeckoDev @forest1971 @Lanchon
Here I am attaching screenshots from AndroBench app benchmark results with bauner, GeeckoDev and forest1971 kernels.
But before that I'd summarize my findings:
1. With bauner's stock kernel the scores are good before Trimming (and after Trimming on forest1971 and GeeckoDev Trim enabled kernels), I did not notice any huge difference in the scores before and after trimming.
2. Scores seems to be poor with forest1971 and GeeckoDev kernels before and after trim.
3. Not only scores, I also noticed that on forest1971 kernel apps took long time to install as compared to bauner's stock kernel.
4. Well I know all respected devs may say that I am the one who won't gain from trimming and may be trimming is not for me. But my question is that why trim enabled kernels (forest1971 and GeeckoDev) are giving so low scores? They should at least give equally good scores as bauner's kernel.
Is there anything that forest1971 and GeeckoDev need to look into their kernels.
Could @forest1971, @GeeckoDev and all the amazing people checking this thread please verify these results on your device and report to devs so that we can have enough stats to conclude.
Thanks a lot for all the amazing work done by @bauner @forest1971 @GeeckoDev and @Lanchon.

Screenshot from @forest1971 kernel after trimming on beta3:
View attachment 3112642

Screenshot from @bauner stock kernel after returning from forest1971 kernel on beta3 (there were almost similar results on bauner stock kernel before installing forest1971 kernel and trimming):
View attachment 3112645

Screenshot from @GeeckoDev kernel after trimming on beta4:
View attachment 3112648

firstly, andobench results are hardly repeatable.

also, trim capable kernels are absolutely NOT in any way different than noncapable ones, except when you actually execute a trim command (ie: it works).

if you want to compare apples to apples, compare one of my trim kernels against my corresponding fpbug-only non-trim version.

some ramdisks might be mounting partitions with the DISCARD mount option. this transparently trims on a trim kernel and not on others, introducing differences. DISCARD should not be used in android 4.3+, as it decreases performance and android 4.3+ has an "offline" trim feature. my kernels do not mount with DISCARD.

but again, the highest factor is probably androbench's randomness.