[ROM][GEN2]CyanogenMod 7 (Android 2.3.7)

[Guest targetbsp]

There is no configuration, it's preconfigured in the kernel and can't be adjusted by apps. Whether you have said kernel depends on which KANG you downloaded. If you have Sejs or Burstlams then it's already included. If you have mine then it's not but can be added as an option by flashing http://blade.windows98.co.uk/downloads/cm7-undervolt-kernel.zip

if you have another one then ermm... see that persons release notes cause I don't know. :D

[Guest sej7278]

Ok, just to test it, here's what I did

--- a/policy/src/com/android/internal/policy/impl/KeyguardViewManager.java

If it works, it's really just a hack as I'm assuming the animation will still run in background but the screen is refreshed (lockscreen disabled) right away rather than giving it 500ms to show the animation. I'll see if it even makes a difference, I might as well be truely on the wrong track here :)

have you tried it? i was looking at frameworks/base/core/java/com/android/internal/widget/RingSelector.java to try and disable the animation in the first place, but can't figure it out. there's some constants in there that seem to define the animation time.

[Guest vareBlade]

Hey guyz, what happened to the nightlies? which rom should i install now?

sorry was absent for a long time so idk my way around anymore :(

[Guest gruad23]

There is no configuration, it's preconfigured in the kernel and can't be adjusted by apps. Whether you have said kernel depends on which KANG you downloaded. If you have Sejs or Burstlams then it's already included. If you have mine then it's not but can be added as an option by flashing http://blade.windows...volt-kernel.zip

if you have another one then ermm... see that persons release notes cause I don't know. :D

Hmmm .... This rises more questions ...

I flashed a file called "update-cm-7.2.0-RC0-Blade-KANG-2012-01-26-signed" and as Mod-Version it says CyanogenMod-7.2.0-RC0-Blade-KANG-26jan and Build-Number is GWK74.

Any way to tell from this if its Sejs, Burstlams or your kernel? For the occasional visitor its hard to keep track on the versions and whats the difference but I understand that this is mainly my problem and not your task cause your are doing fine but making the kernels. But - its hard to keep track of the versions and changes :)

Second question: undervolting should mean that I can configure somewhere when and how it should undervolt. Like it does with overclock? Or is undervolt just a "on/off-switch" where I tell the CPU that it can undervolt itself when its feeling understressed?

thnx,

p

[Guest GodveX]

just flashed sej`s cm7 build worked great tested for a few hours so far n bugs just wondering is it possible to center the clock in the middle of the status bar ? if so how may i do this please

[Guest sm4tik]

have you tried it? i was looking at frameworks/base/core/java/com/android/internal/widget/RingSelector.java to try and disable the animation in the first place, but can't figure it out. there's some constants in there that seem to define the animation time.

Nah, had to go to bed before it finished building.Will post results later today.

[Guest sej7278]

I flashed a file called "update-cm-7.2.0-RC0-Blade-KANG-2012-01-26-signed" and as Mod-Version it says CyanogenMod-7.2.0-RC0-Blade-KANG-26jan and Build-Number is GWK74.

Any way to tell from this if its Sejs, Burstlams or your kernel? For the occasional visitor its hard to keep track on the versions and whats the difference but I understand that this is mainly my problem and not your task cause your are doing fine but making the kernels. But - its hard to keep track of the versions and changes :)

Second question: undervolting should mean that I can configure somewhere when and how it should undervolt. Like it does with overclock? Or is undervolt just a "on/off-switch" where I tell the CPU that it can undervolt itself when its feeling understressed?

if the mod-version includes the 26jan its probably either mine or burgerz's, i don't think burstlam or targetbsp change the mod-version. my kernel version will say simon@hydra #1

the undervolting for the blade/turies is not configurable, its just the tables, not the full driver, so no app required, it just works.

[Guest targetbsp]

As a sneak preview for the results due in an hour... for the third different comparative test I've carried out... undervolting once again makes bugger all difference. :D

And it seemed such a nice idea. :(

[edit]

OK, after an overnight charge and 14 hours of stability testing at a fixed 245Mhz in airplane mode with the phone forced awake in tasker the normal voltage kernel finished on 43% battery (19% after a power off and on)

Undervolted finished 49% (24% after a power off and on)

So undervolting may have a slight edge - or that could just be a reasonable margin of error for the test. Lets say it did have a benefit. That's a pretty insignificant difference for 14 hours of being forced at the absolute ideal condition to see said difference? I can't see it making much real world difference.

Edited January 30, 2012 by targetbsp

[Guest emc02]

well, thats a very low energy consumption

but why is the phone getting warm (hot) when playing games? I don't think, that the screen uses more power than the CPU/GPU...

maybe something went wrong with the test?

[Guest sej7278]

As a sneak preview for the results due in an hour... for the third different comparative test I've carried out... undervolting once again makes bugger all difference. :D

And it seemed such a nice idea. :(

[edit]

OK, after an overnight charge and 14 hours of stability testing at a fixed 245Mhz in airplane mode with the phone forced awake in tasker the normal voltage kernel finished on 43% battery (19% after a power off and on)

Undervolted finished 49% (24% after a power off and on)

So undervolting may have a slight edge - or that could just be a reasonable margin of error for the test. Lets say it did have a benefit. That's a pretty insignificant difference for 14 hours of being forced at the absolute ideal condition to see said difference? I can't see it making much real world difference.

so probably not worth the hassle of a custom kernel and possible stability issues.

i've not noticed any low wifi strength problems, but in theory there could be issues from undervolting the wifi chip, and low wifi strength means higher battery load in the first place!

plus it doesn't work at all when idle or under heavy load, only when you're doing something middle-of-the-road.

but why is the phone getting warm (hot) when playing games? I don't think, that the screen uses more power than the CPU/GPU...

maybe something went wrong with the test?

pretty sure the screen uses more power than the rest of the phone put together, especially the TFT models.

Edited January 30, 2012 by sej7278

[Guest targetbsp]

yeah I don't think it's worth including as standard with a rom. Especially if it's giving a few people stability issues. Now it exists, there's no reason not to keep it around as an optional download I guess.

The test seemed sound to me, between them tasker (keeping the phone awake) and stability test used almost all the power. Stability test had logged loads of runs (vs when it did go wrong in post 12203). the phone was laggy as heck at the times I woke it to check it was still going.

For power consumption, I'd rank them (in an unscientific way!) as:

wifi downloading

screen (oled min brightness)

cpu in use

wifi not downloading

Not sure where the gpu goes as you can't easily test it independant of cpu and screen. :D

Edited January 30, 2012 by targetbsp

[Guest targetbsp]

This was totally posted in the wrong tab!

Edited January 30, 2012 by targetbsp

[Guest saukkiird]

Hiya guys,

I'm running sej's 21jan at the moment. If I want to leap up to sej's 26jan, do I have to go through the whole backup everything-wipe/factory reset-wipe cache-flash ROM-flash gapps-restore everything-routine again?

[Guest targetbsp]

Nope, just flash the rom and wipe the dalvik cache. It'll only take a few seconds and all your existing data and settings will be intact. Though it's wise to still do the backup portion before hand if you're not used to upgrading roms!

[Guest saukkiird]

Nope, just flash the rom and wipe the dalvik cache. It'll only take a few seconds and all your existing data and settings will be intact. Though it's wise to still do the backup portion before hand if you're not used to upgrading roms!

Ok, Just to be sure, do you mean the backup option in recovery mode? Do I still need to backup my apps, sms' and call logs with something like mybackup root? And what does "wipe dalvik cache" mean? Is it the wipe cache option in recovery mode also?

[Guest targetbsp]

I normally use Rom Manager. It's easier as it will wipe the dalvik and take the backup for you. But yes just use the backup in recovery, you can then restore your phone completely if it goes wrong. Not much point in backing up individual items and wiping unless you are changing roms rather than upgrading.

Wipe dalvik cache is under advanced and is different from wipe cache.

[Guest shmizan]

targetbsp, weren't you suppose to check how long it takes? and not how much battery left in the end of the benchmark?

[Guest targetbsp]

targetbsp, weren't you suppose to check how long it takes? and not how much battery left in the end of the benchmark?

Didn't have the time as the CPU wasn't draining the battery enough. There won't be any substantial difference in draining the battery to a certain amount and timing it than draining over a certain fixed time and noting the drain. Any significant difference would be clear in either case. The insignificant difference either means its so little its not worth bothering with or as i said it could be within the margin of error of either an inaccurate battery read (but as I stated, I measured before and after a reboot to force a recheck of the battery level and the difference was the same in each case) or variance in the test despite my attempts to minimise that as much as possible.

Like I said, this is my third different test to compare battery drain all of which say UV isn't worth it so I'm personally confident enough with that conclusion now.

[Guest sej7278]

Like I said, this is my third different test to compare battery drain all of which say UV isn't worth it so I'm personally confident enough with that conclusion now.

based on the "scientific method" of most QA guys i've worked with, testing something works once is enough to prove anything, testing multiple times is positively OCD.

most of them shared an office with the "if it builds, ship it" developers.

and before you ask, no i've not worked with ZTE :D

[Guest Davidoff59]

That's not QA at all. In some industries, a failure that results in say, machine downtime can be charged back to the manufacturer so an attitude like that would be very costly and result in you losing your job. I worked in QA for 6 years.

Saying that, when it comes to electronics nowadays, its very apparent that consumers are beta testers. Its prob much cheaper to fix a bug after release than to delay a release.

It happened to me after buying a sat navigation. Brand new and would not fix to a satellite. They would not refund or replace as it had been opened. So I was forced to buy an extra pack for £50 so that I could upgrade the software. After a lot of hassle, I got my 50 quid back, and a free lead from the manufacturer.

[Guest sm4tik]

Ok, my test for disabling the unlock animation failed.. back to the drawing board :(

edit:

have you tried it? i was looking at frameworks/base/core/java/com/android/internal/widget/RingSelector.java to try and disable the animation in the first place, but can't figure it out. there's some constants in there that seem to define the animation time.

Looking at the file, here's my next test. If this does what I think it will, there will be no blowup but the fade will still be present. I can live with that :) Starting from line 963:

   void startAnimating() {

        mAnimating = true;

        // final Animation trans1, trans2;

        final Animation trans2;

        final AnimationSet transSet;

        final Ring ring = mCurrentRing;


        // trans1 = new ScaleAnimation(1.0f, 7.5f, 1.0f, 7.5f, Animation.RELATIVE_TO_SELF, 0.5f, Animation.RELATIVE_TO_SELF, 0.5f);

        // trans1.setDuration(ANIM_DURATION);

        // trans1.setInterpolator(new AccelerateInterpolator());


        trans2 = new AlphaAnimation(1.0f, 0.2f);

        trans2.setDuration(ANIM_DURATION);

        trans2.setInterpolator(new AccelerateInterpolator());


        transSet = new AnimationSet(false);

        transSet.setDuration(ANIM_DURATION);

        transSet.setAnimationListener(mAnimationDoneListener);

        // transSet.addAnimation(trans1);

        transSet.addAnimation(trans2);

        transSet.setFillAfter(true);


        ring.hideTarget();

        ring.startAnimation(transSet);

    }

If it works, maybe someone can clean it up as I dont think it's necessary to have an AnimationSet if there's only one type of animation left (alpha). ..I would do it if I knew howto ;)

Edited January 30, 2012 by sm4tik

[Guest shmizan]

I don't know what I'm doing wrong.

fresh install, wiped everything (done it 10214 times before), "Shoot Bubble Deluxe" runs laggy as hell!

tried disabling every possible process and turning airplane mode on. tried rebooting 5 times after the install, it just runs slow.

what changed from nightly 220 to this nightly that I'm not aware of? I feel really dumb. I should upload a video but I'm away from home until friday.

[Guest targetbsp]

fresh install, wiped everything (done it 10214 times before), "Shoot Bubble Deluxe" runs laggy as hell!

Just tried it and yeah that game runs terrible! Must be something wrong with it as most games run fine. I like Bubble Maniac myself

[Guest shmizan]

nothing wrong with it.. it runs smoothly on nightly 220 (exact same version). and as stupid as it may sound it's the only game I play (and my girlfriend too) so without it I just keep going back to nightly 220.

also the game menu animations on "Backgammon" from "AI factory" are slower on the latest KANG, so it gotta be something wrong with the KANGs, as it happens with not only 1 game.

also I'll be glad if someone else could test it to verify.

Edited January 31, 2012 by shmizan

[Guest Sulthekk]

http://db.tt/ZArUa7a6

Grabbed from Targetbsp KANG

I am using this with sej's latest build because i found that the uv kernel makes cpu less stable on higher freqs.

Flash it from adb or terminal emulator with flash_image boot path-to-boot.img then reboot.