How can I understand from the name, FDE.AI - hints for the presence of the simplest artificial intelligence. The meaning of the AI is that it will understand the scenarios of using the device and based on this to customize the various system settings in real time.
No personal information is going to, not processed and is not transmitted anywhere. All actions are written in chat (log).
What is implemented: GPU turbo: When starting the AI determines whether the GPU Turbo can work on your device and write about it in the log. AI will observe the level of load on the video screen. If there is a high load for a long time, then the AI will raise the frequency of the accelerator to a maximum of ~ 30 seconds before the next check. As soon as the load falls, returns stocks. This will not increase the framerate, but will make it more stable, because the frequency controller of the accelerator does not have to "jump" in frequencies when changing the load, which ultimately should have a positive effect on FPS stability. In logic there is no binding to the application name - the GPU load is measured directly, it does not matter what loads it (game, benchmark, etc.), making the logic universal. GPU Turbo will not conflict with the logic from Huawei with the same name. When the device is raised, the GPU Turbo trigger threshold increases to avoid overheating the device. Logic is active only when the screen is on. When the frequency screen is off, the stock value is returned.
Supported on most Adreno and Mali video director.
Legacy GPU Turbo: This mode is determined if the driver in the kernel does not provide information about the load of the GPU. Difference from the usual GPU Turbo is that the acceleration will occur relying on the GPU load in real time, but for the presence of the 'heavy' process / game from the #GameList list. Consequently, to automatically work this mode, you need to activate the option of the forced booster of heavy games in the settings. You can also activate the Permanent Legacy GPU TURBO mode manually switting the AI mode to performance. Logic is active only when the screen is on. When the frequency screen is off, the stock value is returned.
Supported on some MALI and VIVANTE video speakers.
Frame-buffer boost: When starting the AI determines whether Framebuffer Boost can work on your device and write about it in the log. At the moment, this logic will only be used if there is support for GPU Turbo. Framebuffer Boost is an acceleration of frame buffer. When accelerated, the frequencies of the personnel buffer are raised. With the active use of the GPU (load more than 30%), Framebuffer Boost will be activated automatically, and as a result of operation, the screen update rate will be maximally stable. Logic is active only when the screen is on. When the frequency screen is off, the stock value is returned.
Supported on most devices with Soc Qualcomm.
CPU turbo: When starting the AI determines whether CPU Turbo can work on your device and write about it in the log. AI will observe the load level of the CPU. With a high load for a long time, the AI will raise the responsiveness of the CPU controller and will force it to keep the frequency above 'High Speed Freq.' (This is about half from the maximum CPU frequency) by ~ 30 seconds before the next check. In logic there is no binding to the application name - the CPU load is measured directly, it does not matter that it loads it (game, benchmark, etc.), making the logic universal. The higher the device temperature is the higher the CPU TURBO trigger threshold to avoid overheating the device. Logic is active only when the screen is on. When the screen is turned off, the parameters are returned to the initial values.
Works with Interactive, OnDemand, ScheduTut and the CPU regulators based on them.
Machine learning: At the moment, machine learning (MO) is used only if there is support for CPU and / or GPU TURBO. MO will study how you use the device and customize some system parameters on the go. So, if you play a lot - II will customize the system for better performance. If you just rewrite and watch the video - AI will customize the system for better energy saving. If you often do both, and the other things will decide what to do, for this and made machine learning. In the FDE application, you can configure the constant preference for choosing MO towards performance or energy saving.
Dynamic customization VM: Each ~ 30 seconds the use of RAM is checked. Depending on the data obtained, some VM settings are adjusted. As a result, the cashem management should be improved in RAM. Logic is active only when the screen is on.
Force DOZE: The option is available for devices running version 6.0 and higher (OFF by default). When the screen device is forced into the deep sleep mode (Doze), then, after 3 seconds, the forced operation mode is canceled and Doze continues in the normal mode, leaving the device to sleep. Not recommended for use on heavily modified OS (e.g., MIUI, EMUI, ColorOS etc.)
throttling switch: The option can be activated in the settings (default). It turns on and disables the trottling for the CPU and the video screen. When trottling is disabled - the device may overheat, but the performance will be maximum even with long loads.
Temperature controller: AI checks the temperature of the device and notifies in the log if it overheats. CPU / GPU triggering thresholds are calculated relying on data from this function. If the device will overheat the CPU / GPU turbo will cease to raise frequencies or will do it much less often.
Charge Controller: AI checks the device is charging or not. If it is charged, AI adjusts VM performance. After disconnecting from the charging AI will return the previous regime which he 'learned'. Force Doze (if enabled) is also dependent on this function, preventing forced deep sleep mode while charging. This option does not increase the frequency and any changes only VM settings, so it is safe and does not cause additional heating devices.
The extension of the cycle life of the battery: The option can be activated in the chat AI if the device is supported by the device (default). When the option is enabled, the AI monitors the battery level of the device and stops the battery charging at the level of 90% (default). The trigger threshold can be configured manually. Charging is activated again if the device is still connected to charging and the battery level fell by 5% of the trigger threshold. Without a charge battery 100%, you can extend the life of lithium batteries.
"Chat" with AI: (Text on a single line - 96 characters)
Starting with version 7 application has the opportunity to correspond with the AI. This can be viewed either as another method of setting the AI, or just for fun. OCR is searched by keywords. The option is still in development. Settings in the "chat" reflect the settings in the application menu. What can an assistant:
What can an assistant:- Display of the message (#help)
- this message is displayed (#help)
- Display brief statistics (#stats)
- Set AI for ext. performance (#performance)
- Set AI for ext. Energy Saving (#powersave)
- Set AI in the 'balance' (#adaptive) mode
- Activate constant boost for heavy games (#hardgaming)
- Display a list of games, appreciated in AI as 'heavy' (#gamelist)
- Add your games to the List 'Heavy' (#ADDGame 'name. Package. Game')
- Clear the entire custom list of added 'heavy' games (#PurgeList)
- Empty the chat / log (#clean #chat)
- Close background applications (#clean #ram)
- Clear the cache files of applications (#clean #cache)
- Clearing Cache Shaders GPU (#CleangPU)
- Optimize cache ART (#art)
- Switching startup script runoff Qualcomm settings at OS startup (#dontqcom / #doqcom)
- Switching application build.prop tweaks at OS startup (#dontbprop / #dobprop)
- Switching closing background applications when OFF option screen (#killbgapps / #dontkillbgapps)
- Skip the use of additional tweaks build.prop for the GPU during the boot process (#skipgputweak / #gputweakback)
- Switching operation sensor / sensor device in Doze mode (#toodozed / #dozesensor)
- Switching method of forced mode Doze (#alternatedoze / #normaldoze)
- 'Refreshing' file system partitions all devices - to check for errors and TRIM (#fsrefresh)
- Copy the current log FDE.AI the root of the internal memory as a text file (#dumplog)
- Switching vibration when performing FDE.AI (#novibro, #vibrate)
- switch the function of extending the battery life cycle (#GreenBattery 'optional, trigger threshold', #fullBattery)
- Answer to the hi / until some other words on the subject (#hi, #bye)
- Loves hvalbu
- He knows who Daddy
- Other.