View Single Post
Posts: 323 | Thanked: 180 times | Joined on Oct 2009 @ Gent, Belgium
#19
I could think of a few things that would help performance on our poor N900 : Thumb2, using the swap on the eMMC instead of the uSD card (aka flopswap), gcc 4.8 compiler.

But the first and last are difficult, as the whole build infrastructure set-up needs to support it:

Thumb2 looks to be difficult as the N900 is claimed to be not performant enough to run fluid GUIs, of 50-60 fps, when using the SoC thumb erratas that are provided by Pali and other in their uboot/kernel/glibc combination. This due to the now slower context switches. Also the QEMU VM based build infrastructure seems to have a bug when trying to compile Thumb for N900. Btw, these are stskeeps words, not mine.

The latest gcc is also not implemented yet (as far as I know).

So yes, optimizations are still possible but the number of N900 users is falling. And priorities are not always there to for the N900, as expected. Jolla / Sailfish / General Mer is most important now on the infra side, together with the x86 HW adaptations. N9xx HW adaptations were v interesting when there were no devices available but since June there should be at least some Jolla prototypes. Also there was not enough community support testing/developing for Nemo. Think of the whole X11 -> Wayland transition and -related- the Qt4 to Qt5 transition and the new Glacier GUI transition. Everything is in transition

Anyway, there are still plenty of Jolla people willing to help Nemo progress on the N9xx, by promising to help out with the new Nemo Glacier UI and other things.
So it's certainly not dead! But they need help, all they can get! Install, test, raise bugs, help fixing.

Just my 2 cents.
 

The Following 3 Users Say Thank You to Netweaver For This Useful Post: