Database Info

[UTIL] PowerSet - Power saving settings replacement

Hi
After some time using orginal Toshiba power saving settings I've found it's terrible slow and upractical using fingers, so I've came with an idea to make simple fast and finger friendly app that will allow me to change basic settings like:
CPU Speed
Wi-Fi speed
Backlight while on battery power
If someone wants to make better graphics for eg. buttons or icon feel free to contribute in this thread and I will think about using them

Nice app ! Hope you'll get someone to make some graphics because I think you desserve it
Maybe you could ask LeSScro

But I'm not sure how much slower it will be with added graphics
I'm not a good programer... Now loking for a way to control cpu via DVS1: driver so I can disable autoscaling

The app seems to be very useful!
I'll install it on my TG01 after I flash a new Rom on it.
Thanks for that app!

Does limiting the CPU actually make the battery last any longeR? When I had it throttled down, I didn't notice a difference :S

"Now loking for a way to control cpu via DVS1: driver so I can disable autoscaling"
I don`t think `Snapdragon`CPU have the ability to auto scale like `Intel Xscale`CPU can. I used Asus P535 before and I don`t need to touch the CPU setting for high or low CPU frequency it does it automatically, for example if the CPU needs more power it will throttle up and if its idle it will throttle down the CPU automatically.

@BIG K in TG01 it looks like this if You'll set Low performance then cpu intensive apps work slower but battery drain is lower if you'll set it to high then cpu reaches max speed and get's higher battery drain it's good for apps that always uses 100% cpu
@elder T Snapdragon CPU does have autoscaling and can be overclocked (look at Snapdragon Android devices forum and at HTC Leo (HD2) fo HD2 there is app to disable autoscaling and overclock cpu but for some unknown reasons works on Acer F1 and LG Expo but doesnt work on TG01.

nico101 said:
but for some unknown reasons works on Acer F1 and LG Expo but doesnt work on TG01.
Click to expand...
Click to collapse
Everybody has reported their oc'ed phones overheating. I wouldn't try it even if the app worked in tg01. Many use it to downclock for better battery life in which case we have the toshiba tool. Or nico101's

can you make a version how our enable wlan and bloototh? and a way to change skin? greets alcatras

I think it's possible to add enabling bt and wlan but it's rather pointles cause most of today replacements etc have this function with a quick access and with skins problem is that I'm not a good in making better graphics but if someone wants to do them I can replace ugly buttons with graphical ones
I know about overheating and I know that toshiba have even overheating warning in drivers that can warn or shutdown the phone... But disabling autoscaling and controling cpu frequency can be more exact not only setting our cpu to High perf. so it will autoscale from low to high when needed but making it stay on high perf.

Comes there no update from your great app?

Thx!
'Dzięki'
Very useful application!

Due to lack of time I've stopped development to find a way to disable autoscaling
Anyway program works like it should and there are no reasonable propositions what should I try to add

I don't think this program works.
I set the CPU speed to low, but it is indeed still actually set to high. I use this program primarily for back light purposes now. Is there any other program out there that can actually set the CPU to low mode? Using the Toshiba program is soo slow!
Just adding some additional information. When I use the Toshiba power utility and I set the cpu at slow from being previously at high, then when I play a movie it is choppy. When I set my cpu to low in this utility, the movie is still smooth. This utility doesn't do anything cpu related. I tell ya, If someone can produce a good CPU power/backlight utility, that would be really benafictial. We really, really need that, the Toshiba utility is SOOOOOO slow.
-Kilihari.

Weird... Try to change settings in my tool then check if Toshiba settings shows the same... From what I've researched TG01 takes CPU mode from registry values and I"m using and changing the sam values

Yes, it is true.... but still ....
It is indeed true that if you change the cpu mode to low or that it will show up that way in the Toshiba power management utility, however, it definitely does not change the actual CPU speed. Man, I really wish this utility worked, we so need a utility like this for the TG01, something that is finger friendly and fast.

I will look into powersave.exe again but first I need to finish polish version of WM6.5.5

Parameters to run as a script
Hello Nico,
thanks for this helpful and battery life time saving application. I am running it without any problem on my Toshiba TG01 for a weeks.
I have a question if there are (or you plan to add) parameters for starting up this application? I would like to set night script to set Low CPU automatically every night, how to achieve this?

Hi i have just tried this app and i gotta say it is amazing!! I use it as a Spb shortcut and i use it frequently to save battery! . Its fantastic and fast. I throttle down the power when i am not using the phone for intense apps (standby, music) than when i am browsing i use full speed. With this method, i have saved alot of battery.! thanks veey much!

Unfortunaetly i am sad to say i have the same problem as kilihari. The app doesnt seem to work. It works after a reboot, or after i change the settings, i have to open up the toshiba settings app, close it then the changes take place. I used benchmark test to test it and i also notice a lag in spb. Its a shame as i highly anticipated this app for fasy access to toggling swithces.
i hope nico could provide us with a update, but unfortunately he has moved on to android.
Anyone want to take the job up?
Sent from my GT-P1000 using XDA App

[To "lagfix" or Not to "lagfix" This is the question]

This will be the start of a hopefully short lived thread as I only intend to use it to post my findings about the affects of a lagfix. Alright if you are going to come in here to start a flame war about how you know it works or doesn't work just "because" do not even post just leave now! if you have actual measurable data or intelligent input then please contribute.
So a lot of folks state that the advantage lagfix has for them is general responsiveness of applications like touchdown or the market or even jsut opening speedtest.net app. I already no that the direct read/write speed to the FS where the OS resides shows the Captivate much faster than the Nexus one and there is no general feeling that the N1 needs any lagfix. So we can not measure any advantage of a lagfix just by testing the read/write speed of the FS. what then? if it's application responsiveness that is the difference then we have to create environment in the form of a measurable benchmark. So right now I have a ver VERY comprehensive I/O bench script running. it creates an application level benchmark environment where several to many processes are each running various mixed mode I/O streams (read/write/cache) doing this with the exact same standards on a non-lagfix FS and a lagfix FS will show even the smallest most tiny variations in the different types of I/O. it will let us know if perhaps there is a certain type of stream (buffer size, block size, file size) that a lagfix might be improving upon. once the results finish here in about half an hour I will sift through the results, build a graph then post it! I am in no way on one side or the other of a lagfix, I have used both and would rather keep my opinion to my self right now. check back later today for results. I welcome all constructive input/ideas/do it better benchmark (currently 1429 Hawaii standard time)

This is also a question that came in to my head before installing a lag fix.
Well i notice some lags in a stock non-lag fixed captivate. Examples are in games like asphalt (when you win/lose a race) and in N.O.V.A. (progress save) and sometimes in google search widget (when you tap on the text bar, a little sec before the virtual keyboard appear).
I thought that these are just bugs on software or the app itself, but everything is gone when i install lag fix. Yes there are still lags but not as much as before.

This is awesome, can't wait to read it. Are you considering releasing the batch script so others can attempt to contribute?

Already got moved but it doesn't matter. I for one can not wait to see your results cruiser. If you do decide you want to let someone else try your benchmark I'll gladly run it as well.

phishie said:
This is awesome, can't wait to read it. Are you considering releasing the batch script so others can attempt to contribute?
Click to expand...
Click to collapse
Absolutely! I take no credit for these tools and credit will be given where due once the graph is up. I only wrote a script to create the benching environment I wanted. using these tools you can write a very comprehensive script to create the benching environment you want/need for the specific aspect that is being benched. these tools don't just test hardware as hardware is nothing without great software. so using the application level software components to perform these benches and bouncing them off the hardware benches you can see what areas of the software environment are choking what part of the systems performance and vice the versa.

There are a few different lagfixes out there - so don't leave any out. Also, some attack one partition, others attack more then one (/dbdata) - so you need to take this into account. Also Touchdown, and i believe stock email app, benefit because they store data on dbdata in a database - so database throughput should be part of the equation.
Benchmarks are nice, but i have tried with and without and see the difference deleting emails using the stock client. That being said, I rolled back to stock because one-click odin does a poor job of reformatting everything - I don't want to get stuck with an EXT4 filesystem and stock kernel with no way to go.

I found with 2.2 it is way faster than 2.1. But 2.1 with any lag fix is still faster. Now you could run 2.2 with no lag fix and it's definetly fast enough but with a lag fix it's amazing. It's like buying a 300hp sports car then the sales guy asks you if you want the 400hp model for a free upgrade and it gets the same gas milage and you saying no because 300hp is plenty to get you around town. If theres even a chance a lag fix makes a difference why not do it, it's not gonna make it any slower thats for sure.
Sent from my SAMSUNG-SGH-I897

Just my 2 cents, I do think that lag fix will give you better performance with certain apps however, it is not that significant and with Froyo, I had problems with apps force closing and battery drain (probably from all the background crashes) with it enabled. I do not think it is needed and may cause more problems than it is worth. Maybe later once the official Froyo is out for a while and the lag fix apps are improved, it may be worth another look. But for now, I do not think it is worth it. But your mileage may vary and if you think your phone is too slow without it, you may have some other underlying issue that needs more attention than trying to fix it with a lag fix.

Duty calls, this is taking longer than I anticipated so I'll still be running this on different file systems tomorrow after work. very interesting results so far though! Spoiler alert! there are obvious gains from database read/writes, only milliseconds but I suppose that counts when it comes to responsiveness of certain programs.

I can't wait to see the results.

I wanted to say that I am interested in what you are trying to do here and appreciate the time you are putting into making this an unbiased effort at answering this question.
I also wanted to provide you with my own real world experience based on three different captivates all presently using cognition 2.2b5.5...until later tonight...lol
I have never installed a lag fix on two of the phones (my wife and duaghter's) and held off on reinstalling the ext lagfix on mine again until yesterday as I wanted to see if it would benefit the performance prior to updating again to beta6
On the phones that my wife and daughter use, I can see where the lagfix is not necessary and they don't complain about any hangups. However, on mine I definitely see a performance improvement with the lagfix installed and i am quite sure that the reason is I am addicted to widgets almost as bad as flashing and I cannot seem to get away from them. After installing the lagfix I do not get the occasional "freezes" that occurred without it and music never skips when opening other apps.
Also, all three phones are running launcher pro, using at least a few widgets, such as friends, messaging, weather, etc. Not sure how you would reflect this in your testing but I do see a difference here though I would love, and certainly tried, to not use the lagfix.

It seems people have gotten bored of this issue. I will say this:
When I ODINed directly from 2.2Beta2 to 2.2Beta5.5, lag was undeniably horrible and OCLF made a huge difference.
When I master-cleared 5.5, ODINed to stock 2.1, then ODINed to 2.2Beta6, everything became zippy fast again without any lagfix. I'm afraid to even try to put OCLF on there now.
Anyone experiencing lag when others say they shouldnt: go through most or all of the steps I laid out for battery drain. Quirks abound, and your lag and lagfix situation could be an indirect side-effect, and not a direct result, of the ROM you've flashed to.

I agree. Some of the claims are down right odd about lag issues so I too feel a lot has to do with "left overs".

My Experiences
i didn't install the lag fix because of the horror stories in the forums about applying an update and having to roll everything back. Normally it's not so bad but I've noticed some of the behavior mentioned. Are these issues addressed in the Froyo leaks (hopefully official) or will these always be a problem.
-Load time of games (using Asphalt 5 and comparing the load times to an IP4 running the same game). The Captivate load times were very slow.
-The appearance of freezing when you win / lose a race in Asphalt 5?
It may not be a fair comparison for load times but it's obvious signs something is wrong when I'm trying to brag to my friends about how my phone is better then theres
The biggest issue I have is with the music stuttering though. I listen to music at work (normally the 13GB data partition is full of music) and the email client is running in the background. The music stutters A LOT when it's downloading the emails. Is this issue no longer a problem in Froyo?

Hopefully this settles it.
Here's a vid I made. I don't know how to split-screen, but I think this makes the point regardless.
I master cleared, then stock 2.1, then Cog 2.2B6. I didn't restore any backups, all fresh installs and configs.
When I ODINed directly from Beta3 to Beta6, the lag was horrible. Doing a master clear, stock, to 6 was much better. But even then, as seen below, there is blatantly noticeable lag. That only leads me to believe that the issue relates to something like fragmentation, and that if people used beta6 for a few weeks without a lagfix, they'll end up with really bad lag too.
For the record, I doubt any of this has much to do with DG's work. I don't think DG is responsible for lag. DG is only responsible for giving us awesome alternatives to choose from when we don't like what the factory gave us and we're willing to take the risk. However, the lag is not a placebo, and your benchmarks are deceiving you.
I picked Touchdown because it was an easily reproduced example, but I notice a difference all over.
The stopwatch is to add legitimacy. The iPhone, to add irony. There's a glass of beer off-camera you can't see, but it's responsible for most of the background noise.

Alright folks!
Those of us in the Military know how your work schedule can tends to change randomly causing you to change your plans a lot. Anyway as I don't see myself with the free time to spend making a cool but not very practical PowerPoint right now I am going to tell ya'll what I found and post the DHT tools so you can play around with it yourselves.
First to those who wish to try the tools, read the vague overview file in there it's as good as it gets as far as information from DHT. you HAVE to run the IO bench several time in both read, write ADN mixed modes. Make sure you run each mode a few times specifying different files sizes block sizes AND numbers of IO streams! each specific setup should be run 5 times and averaged, if run in each set is 5% or more from the rest it should be scrapped and run again. NOTE: some modes require you to have already deployed a file or DB to run the test on!
Alright so I found the difference in every area except one to be non-substantial! and by that I mean 5%+/- (within the considered error margin)
I ONLY HAVE HAD THE TIME TO TEST OCLF! all benches were run on Cognition 2.2B5.5.
When you start getting more than TWO IO streams you start to see a substantial difference in the DB IO streams. in my last run in mixed mode multi-stream bench of 5 runs (all within the acceptable margin) it showed a 19% increase in throughput over the stock FS! I know that doesn't sound like a lot but milliseconds count when it means the responsiveness of an application! One thing I did not do which I wont get to for a while is look into how many IO streams are running and in what modes while running some common applications! this could give us an idea of what kind of bench environment to set up and perhaps give us a base (just for information sake) % throughput increase with a certain FS. just to give you an idea (I am sure most people already know this but wth) of what types of application might take advantage of an increased IO throughput on DB read/writes. A lot more applications do DB IO streams than you might think, some email applications/exchange contact syncs, even some system settings are stored in DB anyways I say it's up to the person! I wont personally use a FS patch "lagfix" until A at least as stable as the current COG2.2B6 Voodoo kernel is done for the captivate. I don't feel like messing with the loopback stuff and whatnot, I can live with it like it is as I don't feel it's bad enough to HAVE to have one, but again everyone is different! However when Voodoo comes standard in a good ROM like COG I say what the hell~! even if it's only 10% gain and only in certain situations! if it's stable and cooked in then it can't hurt me right! ok enough now hope you guys can figure out the tools well enough. it is clear that no matter what anyone says it is factual that no matter how small in SOME situations/applications there IS an IO throughput gain when using OCLF and I am sure any other EXTx FS. weather or not the gain % is worth the possible side effects, well like everyone keeps saying (well almost everyone, some do just cry "MUST HAVE IT") that's up to you!

Ok great video mate.
I also noticed one HUGE lag difference which now made me a believer.
I use Launcher Pro and when adding applications to the lower scrolling bar there was always lag the first time accessing the applications.
I timed it at 17 seconds without lag fix and no delay with lag fix...that is BIG!

That was one of my bigger lag issues too...lp seeming to freeze but after finally giving up on touchwiz a while ago and removing it I have been able to get by without the lagfix though I am probably going to use one again as soon as voodoo is available for 2.2
Posted from my phone using some technology I don't understand

shaolin95 said:
Ok great video mate.
I also noticed one HUGE lag difference which now made me a believer.
I use Launcher Pro and when adding applications to the lower scrolling bar there was always lag the first time accessing the applications.
I timed it at 17 seconds without lag fix and no delay with lag fix...that is BIG!
Click to expand...
Click to collapse
I originally intended to include a few other tests in the video:
* Adding an app shortcut
* Opening the app drawer
* fast scrolling through the 3D app drawer and some widget-heavy home screens
* boot times (particularly initializing widgets on the home screen)
However, in prepping for the video (and most particularly after removing TW), I determined that I could not reproduce the lag effect as consistently in any of the above. I've definitely seen what you describe, and noticed lag differences OCLF vs. not OCLF in all of the above, but I've also ran some tests where I couldn't reproduce it. Touchdown was one of the easiest to demo and that never fails to present an obvious difference.
My current theory is that while Froyo has performance tweaks, a lot of the reason people think there is no need for lagfix anymore is because more people were master-clearing and flashing stock before flashing Cog betas. This gives you a clean slate on your fs, with little chance of fragmentation or other inefficiencies that crop up in a well-used file table. Like I said, flashing without going stock or master-clearing, or restoring lots of backups afterwards, seem to be what bring back the lag. A well-managed cache like you get with a lagfix helps that a lot. I still say that even those who claim to see no lag, if they put some heavy use and load some heavy databases and throw some I/Os at them, then over time they will see more and more of the same lag others see. I don't know that for sure, but I do know for sure what you can see in the video. And that's on a recent setup of master-clear, stock, beta, and no backup restores whatseover.

This probably in the wrong section, but I recently purchased the captivate through an upgrade, and read about and installed the lagfix. Yes it did speed up my phone considerably. My quadrant score was about 2386. I was pretty damn impressed. But then I wanted to go back to stock, and just did a factory reset. I don't think the lagfix is installed anymore. But I just want to know if I didn't hurt my device by doing so. Thanks

Forced GPU 2D rendering

There is an option in ICS to force HW (GPU) 2D rendering in applications.
Do you guys use it? Is there any obvious performance boost in apps?
I have to admit I don't see anything changed at all in my apps.
For example, when I flick G+ posts, there is a little bit of "lag", also in Feedly and other such apps.
Is this because of poor app development, heavy pictures content and graphics (don't see why this would be a problem in feedly as there is mainly text, G+ has much bigger pictures), or something else (all known IO problem.)
After using a tablet for less then a week, having tons of apps, some widgets etc...i can see drastic performance drop in overall.
I tried to install SetCpu to see if with different scheduler there would be any difference, but I have no option for SIO, only noop. Maybe I am missing something here.
Nevertheless, I think tablet should run super smooth in apps like g+ and others. Is this really related to IO?
Thx

Svashtar said:
There is an option in ICS to force HW (GPU) 2D rendering in applications.
Do you guys use it? Is there any obvious performance boost in apps?
I have to admit I don't see anything changed at all in my apps.
For example, when I flick G+ posts, there is a little bit of "lag", also in Feedly and other such apps.
Is this because of poor app development, heavy pictures content and graphics (don't see why this would be a problem in feedly as there is mainly text, G+ has much bigger pictures), or something else (all known IO problem.)
After using a tablet for less then a week, having tons of apps, some widgets etc...i can see drastic performance drop in overall.
I tried to install SetCpu to see if with different scheduler there would be any difference, but I have no option for SIO, only noop. Maybe I am missing something here.
Nevertheless, I think tablet should run super smooth in apps like g+ and others. Is this really related to IO?
Thx
Click to expand...
Click to collapse
Forced Hardware acceleration isn't necessary for the newer spps they should already use it on their own.
However it csn cause FCs especially in games if you activate it. So actually no need to activate it.
Sent from my Galaxy Nexus using xda premium

Svashtar said:
There is an option in ICS to force HW (GPU) 2D rendering in applications.
Do you guys use it? Is there any obvious performance boost in apps?
I have to admit I don't see anything changed at all in my apps.
For example, when I flick G+ posts, there is a little bit of "lag", also in Feedly and other such apps.
Is this because of poor app development, heavy pictures content and graphics (don't see why this would be a problem in feedly as there is mainly text, G+ has much bigger pictures), or something else (all known IO problem.)
After using a tablet for less then a week, having tons of apps, some widgets etc...i can see drastic performance drop in overall.
I tried to install SetCpu to see if with different scheduler there would be any difference, but I have no option for SIO, only noop. Maybe I am missing something here.
Nevertheless, I think tablet should run super smooth in apps like g+ and others. Is this really related to IO?
Thx
Click to expand...
Click to collapse
Could be, but currently no-one can tell. I see no benefit whatsoever from turning 2D GPU rendering on, and indeed it has been clearly estabished lots of games (and some apps) suffer severily when it is enbled, so I keep it disabled.
On the stock ROM/kernel combo, only noop is available as a I/O scheduler. You can use ATP Tweak and the loose sio scheduler file to make it available/select it.

Its helps with the Netflix app that's about it
Sent from my SAMSUNG-SGH-I727 using xda premium

[I9001][ICS][GUIDE]Boost up your ICS based phone!

I create this topic for information and stop asking the same questions on developers threads. Becouse many of people asking how boost up his phone and get more free memory.
In many threads I can see questions like "My phone gets laggy when it wake up.", "My phone after install ICS are slowed". That's can be, because many of us use app's like task killer or phone is just littered. Task killers in fact dosen't boost up your phone - it's only placebo effect. I'll try to explain something.
This is nonsense! Task Killer does not help significantly extend battery life, no use to speed up application startup or most of the services to which the theory, as described, has come in handy.
Among the members there is a mistaken belief that only free RAM will speed up your system. This legend is a denizen of the times of old systems such as Windows 98 or XP, which actually had some influence. New systems such as windows 7 or Android intentionally cache memory to speed up the service most frequently used programs. At the time when we want to take advantage of another and not for the memory system will complete the other processes - in anything he does not need help.
Manually release the RAM can not be anything here. In a moment, the system will use the free memory with another program, for example - a completely unnecessary and contrary to our purpose - Task Killer, which we used to release the RAM after all do not give us a needed effect!
Moreover, by doing so quickly, it appears that the Task Killer himself consumes more memory and the battery than our smartphone applications running in the background, that we wanted to get rid of! This vicious circle can be especially annoying if you decide to including the option of automatic removal of applications running in the background. I recommend skeptics at the end of the day to see how much battery consumed by Task Killer... Effects can be really intresting. So you still want to use Task Killers? Yes? At the end I'll show you something - several facts of RAM in Android:
Android intentionally caches RAM applications and services
Manual removal processes free the memory for a moment - it will soon be allocated to another program
Improper removal of automation of processes in Task Killerz may result in more rapid wear of the battery and unnecessary processor usage
Resource usage when running Android itself will close several processes to allow the newly launched operate at full capacity
Click to expand...
Click to collapse
Task Killer may be helped when we need to check which application consume battery and processor - but not in the daily work.
In ICS memory management and performance is really good - we don't need to use Task killers at all.
Next is lags when phone wake up from deep sleep mode. Yep - that's in fact that deep sleep mode in our ICS gets phone laggy sometimes. But here is simple solution which I write many times in other topics you just must:
Go to Settings -> Performance -> Processor, and now change grovernor to ONDEMAND or INTERACTIVE (in many phones INTERACTIVE work perfect), then you must set minimum CPU frequency to 368MHz. That's all! Your phone when wake up will have no more lags. Simple, isn't it?
Click to expand...
Click to collapse
I'll try to update the topic to date. Cheers.

reversed for V6SuperScriptCharger description for our phones and ICS

but the battery drain will be worse if i set min cpu freq at 368 mhz ?!

task killers may be counterproductive though it is wrong to say, that androids ram management is good. when it needs memory it can kill apps you may need while it thinks other apps you don't really need right now are too important to kill. so it happens that too much ram is used or the system doesn't give enough ram to an app and the phone starts to lag. in this situation it really helps to kill some background apps which i do manually.

thanks for share.

haribo318 said:
but the battery drain will be worse if i set min cpu freq at 368 mhz ?!
Click to expand...
Click to collapse
No. In deep sleep phone don't use it. 368MHz is really optimal setting. I have it on my phone and battery is really god. In normal use my phone withstand 2-3 days. Wifi, calling, messaging, sometimes small games and internet browsing.

Harrocan said:
No. In deep sleep phone don't use it. 368MHz is really optimal setting. I have it on my phone and battery is really god. In normal use my phone withstand 2-3 days. Wifi, calling, messaging, sometimes small games and internet browsing.
Click to expand...
Click to collapse
ok nice i will try which rom are you using.
cm9 by arco and which build or by ivendor ?

haribo318 said:
ok nice i will try which rom are you using.
cm9 by arco and which build or by ivendor ?
Click to expand...
Click to collapse
I'm use CM9 build6 by arco. Now i'm gonna test beta1 by Ivendor

Harrocan said:
I'm use CM9 build6 by arco. Now i'm gonna test beta1 by Ivendor
Click to expand...
Click to collapse
ok could you test which you think is smoother and which hast the longest battery time ?

haribo318 said:
ok could you test which you think is smoother and which hast the longest battery time ?
Click to expand...
Click to collapse
I'll try.

Thanks for the post, it's really interesting.
I realized myself battery drain is lower when not using the task killer app.
I adjusted my settings and i'll see if "interactive" works better than "ondemand".
I'll wait for more info in this thread.

Alpha120 said:
Thanks for the post, it's really interesting.
I realized myself battery drain is lower when not using the task killer app.
I adjusted my settings and i'll see if "interactive" works better than "ondemand".
I'll wait for more info in this thread.
Click to expand...
Click to collapse
interactive should be a little bit more battery friendly and faster.

Sorry, my bad. In first post should be interactive. I change this.
Sent from my GT-I9001 using xda app-developers app

Great Guide, it is definitly needed

tomsteel1 said:
The way linux works is slightly like Java, It does a rubbish collection once RAM Fills up. So Ive noticed once your system goes below 50mb's free of RAM, It will kill unnecessary apps, or wipe the cache.
The only real way to speed up our system (RAM wise) would be to either make ZRAM fully functional, Or a kernel with the RAM Swap feature.
Click to expand...
Click to collapse
Yes, you have right. But on this time we can't have this so i decide to write this guide. I'll add any good posts to 1st post to explain how it can be only possible or just for information

In my opinion setting the minimum frequency to a higher state can't be a solution. Maybe we should optimize the governors configuration like arco did a time ago while the wakeup lag was more aggressive with the interactive governor. I guess there is much more potential in those values if we adjust them correctly.
Unfortunately I don't want to screw my phone with wrong settings so I won't change those values as long as I can't be sure that they're safe.
If someone want to google or experiment by his own - feel free with no warranty
These are the default values for interactive and ondemand:
Governor: interactive
Values:
hispeed freq (1401600)
go hispeed load (85)
above hispeed delay (80000)
min sample time (80000)
timer rate (80000)
boost (0)
boostpulse (-1)
Governor: ondemand
Values:
sampling rate (50000)
up threshold (90)
down differential (3)
sampling down factor (1)
ignore nice load (0)
powersave bias (0)
io is busy (0)
I'll try to test something if I have the time

On demand !!
Harrocan said:
I create this topic for information and stop asking the same questions on developers threads. Becouse many of people asking how boost up his phone and get more free memory.
In many threads I can see questions like "My phone gets laggy when it wake up.", "My phone after install ICS are slowed". That's can be, because many of us use app's like task killer or phone is just littered. Task killers in fact dosen't boost up your phone - it's only placebo effect. I'll try to explain something.
This is nonsense! Task Killer does not help significantly extend battery life, no use to speed up application startup or most of the services to which the theory, as described, has come in handy.
Among the members there is a mistaken belief that only free RAM will speed up your system. This legend is a denizen of the times of old systems such as Windows 98 or XP, which actually had some influence. New systems such as windows 7 or Android intentionally cache memory to speed up the service most frequently used programs. At the time when we want to take advantage of another and not for the memory system will complete the other processes - in anything he does not need help.
Manually release the RAM can not be anything here. In a moment, the system will use the free memory with another program, for example - a completely unnecessary and contrary to our purpose - Task Killer, which we used to release the RAM after all do not give us a needed effect!
Moreover, by doing so quickly, it appears that the Task Killer himself consumes more memory and the battery than our smartphone applications running in the background, that we wanted to get rid of! This vicious circle can be especially annoying if you decide to including the option of automatic removal of applications running in the background. I recommend skeptics at the end of the day to see how much battery consumed by Task Killer... Effects can be really intresting. So you still want to use Task Killers? Yes? At the end I'll show you something - several facts of RAM in Android:
Task Killer may be helped when we need to check which application consume battery and processor - but not in the daily work.
In ICS memory management and performance is really good - we don't need to use Task killers at all.
Next is lags when phone wake up from deep sleep mode. Yep - that's in fact that deep sleep mode in our ICS gets phone laggy sometimes. But here is simple solution which I write many times in other topics you just must:
I'll try to update the topic to date. Cheers.
Click to expand...
Click to collapse
Please continue use ON DEMAND...i used interactive and that will kill Deep Sleep.....I set Cpu to 368 Mhz and it works perfect...

kingsnoopy said:
Please continue use ON DEMAND...i used interactive and that will kill Deep Sleep.....I set Cpu to 368 Mhz and it works perfect...
Click to expand...
Click to collapse
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
I use Interactive..

Me too... What about "conservative"?
Obviously, I wont ask for "performance" nor "power save" as they are pretty self-explanatory.

Alpha120 said:
Me too... What about "conservative"?
Obviously, I wont ask for "performance" nor "power save" as they are pretty self-explanatory.
Click to expand...
Click to collapse
Conservative:
A slower Ondemand which scales up slowly to save battery. The conservative governor is based on the ondemand governor. It functions like the Ondemand governor by dynamically adjusting frequencies based on processor utilization. However, the conservative governor increases and decreases CPU speed more gradually. Simply put, this governor increases the frequency step by step on CPU load and jumps to lowest frequency on CPU idle. Conservative governor aims to dynamically adjust the CPU frequency to current utilization, without jumping to max frequency. The sampling_down_factor value acts as a negative multiplier of sampling_rate to reduce the frequency that the scheduler samples the CPU utilization. For example, if sampling_rate equal to 20,000 and sampling_down_factor is 2, the governor samples the CPU utilization every 40,000 microseconds.

Games Crashing With Flickering

Games like Strikefleet Omega and Shadowgun crash with the screen flickering. Does anyone know why this happens?
Sent from my ASUS Transformer Pad TF700T

i have a tf300 and riptide crash like this too, looks like every one is having this problem.

Bump had this problem on prime with mc3 and now on the infinity. It has something to do with Tegra 3 I believe

In my two weeks of using my Infinity I came across this problem twice. Restarting the tablet always fixed it.

Even after restarting games do the same thing
Sent from my GT-I9300

I'm guessing that this issue is related to the OpenGL Out Of Memory issues that seem to plague ICS on the Transformer devices.
The only way that I know of to "fix" the issue after it occurs is to reboot the device.
Not the best workaround, but until Asus fixes it, I'm not sure what else you can do.
If you want to verify if this is related to the OpenGL issue, just open a terminal window and type this:
logcat | grep -i opengl
If you see OpenGL out of memory errors, then it's probably related...
Sent from my ASUS Transformer Pad TF700T using Tapatalk 2

Yeah I get consistent screen flickering almost guaranteed anytime I play a game. Is there a rom that fixes this?
Sent from my ASUS Transformer Pad TF700T using xda premium

I found a fix for this, at least it worked for me.
First power off the device.
Then hold the power button and volume down button.
You should see some text in the top left corner, let go and press the volume up button.
Now clear cache and reboot.
Everything should now work fine!

Something strange going on there because all of the games mentioned have played perfectly for me over the last several weeks. Six Guns is the only game I've been unable to play because of graphical issues.

No issues with game play for me, except for Modern Combat 3. I would constantly get connection loss during game play. Although not as often when I had my Prime.
Sent from my ASUS Transformer Pad TF700T using xda app-developers app

I had this issue as well, however, since i removed all of my third party launchers and went back to the stock launcher I haven't had a problem. There was another thread in here someplace where someone else said the same thing. Another thing I did was remove the apps that mess with the scheduler and cpu (ATP tweak, no frills cpu, etc), and of course froze alot of bloatware. Ever since I did that the tablet has no problems and absolutely screams. I would recommend stripping it down to as close to stock (what you would see if you just did factory reset) then work your way back up testing apps as you go. I have a feeling the launcher apps and/or the scheduler/cpu tweaking apps are not quite optimized for the infinity.

hshah said:
Does anyone know why this happens?
Click to expand...
Click to collapse
I have had the same thing happen to me. The cause is lack of memory, more specific, contiguous memory. Especially when playing a graphics intensive game, it will want large chunks of memory. This would have been less of a problem if all apps used the java VM, which takes care of the memory management and reduces memory fragmentation. However to get the most out of the system, most games do not (exclusively) use the java VM, but are written in native code. This however puts the job of memory management in the hands of the developer, being one myself, I know it is hard to keep track of every piece of memory the app uses and thus, should release again after use. Not doing so can cause memory leaks and memory fragmentation, so the next app you run will have less memory at its disposal.
The solution that works for me before starting a "big" game is to make sure no unnecessary apps run in the background and check the available memory. I have found FMR Memory Cleaner to be quite useful. If the game still crashes or causes the screen to flicker, reboot. If that doesn't help, uninstall/freeze/disable those apps that FMR lists, but you can do without, then try again, and maybe even add a reboot for good measure.
The drawback of having such a high resolution screen on the TF700 is that it requires more memory to run when compared to the 1280x800 screens. Unfortunately ASUS did not increase the amount of memory...
Sent from my ASUS Transformer Pad TF700T using xda app-developers app

xcal321 said:
I have had the same thing happen to me. The cause is lack of memory, more specific, contiguous memory. Especially when playing a graphics intensive game, it will want large chunks of memory. This would have been less of a problem if all apps used the java VM, which takes care of the memory management and reduces memory fragmentation. However to get the most out of the system, most games do not (exclusively) use the java VM, but are written in native code. This however puts the job of memory management in the hands of the developer, being one myself, I know it is hard to keep track of every piece of memory the app uses and thus, should release again after use. Not doing so can cause memory leaks and memory fragmentation, so the next app you run will have less memory at its disposal.
The solution that works for me before starting a "big" game is to make sure no unnecessary apps run in the background and check the available memory. I have found FMR Memory Cleaner to be quite useful. If the game still crashes or causes the screen to flicker, reboot. If that doesn't help, uninstall/freeze/disable those apps that FMR lists, but you can do without, then try again, and maybe even add a reboot for good measure.
The drawback of having such a high resolution screen on the TF700 is that it requires more memory to run when compared to the 1280x800 screens. Unfortunately ASUS did not increase the amount of memory...
Sent from my ASUS Transformer Pad TF700T using xda app-developers app
Click to expand...
Click to collapse
This might illustrate my lack of knowledge when it comes to these things, but, is this something that Asus can fix by software upgrade or will it be a perpetual problem.