About the future of SFOS
Before answering this question, we should agree what is SFOS and what is not. However, I wish to skip this premise and go straight to the point with this image:
The original high resolution image is here.
What’s this image shows? A sleeping system when untouched that consume 5%/h of battery which means 20h of stand-by. It is not a great achievement because it should be 100h of stand by for an optimised system (here).
On the other side, we notice that CPUs stopped to be pollute and fall asleep and awaken constantly (sleep, awake, sleep, awake) apparently without a real need. BTW, the system consumed 5%/h of battery - while untouched with the CPUs which NEVER goes to sleep. This means that apart the display consumption, 4G and WiFi connections, the system was running at its 100% of the time.
Update: now, with the 4G data mobile and WiFi tethering connections active, it sucks 6%/h. Therefore the changes tested are more suitable for a daily calm working session rather than keeping the smartphone near our bed while we are sleeping (check here).
How did I achieved this?
This is the log of that day and changes:
-
in the morning I developed this patch to fix few udev rules and applied to the system - reloading and restarting the systemd-udevd and related services would have been enough but I did a reboot for be 100% sure to not have bad surprises after.
-
during the day, I used it with different application and for several tasks also with
Android Supportand Android apps. -
Before leaving the smartphone untouched for 4h, I did a swap offload using the script conveyed by this patch about
zRAMand applied some power management rules (not published yet) about CPUs and internalSSDflash devices.
Why this achievement matters?
From a point of view of the performances, it is not a game changer. From the PoV of bringing the OS part of SailFish OS under strict control - in particular about power management vs performances - it is a huge advancement.
After all, the great business that Google and Apple did was not about their OS but about their Market revenues. In fact, there is no hopes for apps developers to earn good money as long as the operative system below their app is out of control and does not provide a consistently good performances.
Which is the reason because the Linux kernel won the global challenge of being adopted. Also in this case, the great business is not about the kernel itself but about all the applications that can run because that kernel delivers a consistent good QoS (quality of service).
Think about this, because it is about the future of SFOS much more than everything else.