G
Gast5050
Gast
Also auch wieder die 1350 auf 1200 setzen ja?
^^ ja ^^
Ich habs auf ondemand
^^ ja ^^
Ich habs auf ondemand
Folge dem Video um zu sehen, wie unsere Website als Web-App auf dem Startbildschirm installiert werden kann.
Anmerkung: Diese Funktion erfordert derzeit den Zugriff auf die Seite über den integrierten Safari-Browser.
So ging es mir auch. Seit gestern experimentiere ich rum und der Stromverbrauch bei der v20 ist nicht weniger geworden. Deshalb bin ich zurück auf v19 den da war alles ok und die Akkulaufzeit viel besser. Alles gleicher Kernel und gleiche Einstellungen. Nichts verändert.Angosa schrieb:Ich muss aber sagen, dass ich von der Akkulaufzeit dabei echt enttäuscht bin. Ich schaffe es in 18 Stunden nicht mal auf 3 Stunden Screen on. Mit dem AOKP Stock Kernel habe ich es in der Zeit auf 4 Stunden geschafft.
[KERNEL][GPL][25 JAN - #14] franco.Kernel | 4.0.3 | OC | UV | INIT.D | COLORCONTROL - Page 412 - xda-developers25th January - nightly #14
* Bumped up some bus frequencies
* Added two omap patches that fix two false IO wakeup detections
* GPU Overclocked again to 384
* Changed the voltages file to increase stability
* C4 state patch reverted - according to Colin Cross (Google Android team) it can only save 2mAh max during idle, but it could cause some instability, thats why they reverted it as well. I want stability, so this stays reverted
* VM_Readahead values hardcore - now it's set to 512kbs
* CIFS added as a module
* Writeback values reverted back to stock - they seemed to perform extremely well in conjunction with the Writeback IO backport from 3.2
* Conservative governor tuned again - what I did this time is pretty simple, this governor now only jumps between min frequency and max frequency. Why this? Because I want to minimize any kind of lag that "can" happen when polling and then ramping up step by step to max frequency. To balance the battery/performance ratio I increased the ramp_up threshold from 60% to 80% CPU load, which means that while it doesn't use the 920 step (in the 700min-1200max scenario, since it's the default and the one that I recommend) it will ramp less times to max frequency. This is wielding the best results in the last 24 hours for me and my testers didn't complain at all, quite the contrary
I don't understand this AndroidPIT people. My application ain't, for some reason, in their databases so I can't list it. I send them 5 emails, got one answer that they are gonna look into it. No more answer for almost a week. Contacted their support through the site, that says they come back to me in 24 hours - false, been 24 hours already not a single shit of an answer. This is how I'm starting to feel: fuck them.
It's simple. Yesterday I did a change in tuna-display drivers trying to reduce more of the yellow tint, so I commented the 8500k to D65 color changes in all the three structs that defined the display shits. I noticed no changes in my screen, but I forgot to revert that, and since I had no issues I never thought of it. Then I realized that I forgot it, and I only reverted that patch from #14 to #14.1. It's the only thing that changed But the scenario you pictured is not very far from reality. I wonder how my testers still talk to me in a daily basis, as I usually send them a few kernels per day and I bug them around for hours asking for input, bugs, battery etc etc. They deserve as much as credit as myself when building this kernels, make no fun of it