View Single Post
olf's Avatar
Posts: 305 | Thanked: 1,246 times | Joined on Aug 2015
#139
Originally Posted by Zeta View Post
[...] So they are great and powerful in embedded systems. iMx6 are still on the older Cortex-A9 architecture (one to 4 cores), and iMx8 should be a lot more powerful with newer generation 4 cores Cortex-A53 (ARM-V8 64 bits, against ARM-V7 32 bits). [...]
One should not expect too much performance gain by going from a ARM's low-end 32bit core (A9) to their low-end 64bit cores (A53 & A55): This is mostly 32 vs. 64bits, e.g. allowing for 4+ GBytes of RAM.
Even though ARM also designed a class of cores below that (i.e. the very-low-end 32bit A7 and 64bit A35), their midrange cores (32bit: A12 = A17, 64bit: A57 & A73) and especially their high-end cores (32bit: A15, 64bit: A72 & A75) are way faster.
And the Vivante GPUs are definitely not anywhere near Adreno: In terms of performance (while Vivante surely suffices for a 3D-accelerated GUI, 3D gaming will be limited) and FOSS driver support (Freedreno has been in MESA releases for some time already, while Etnaviv is still on its way and currently by far not as mature as Freedreno; but supposedly that will be achieved by 2019)!

But while I wonder why many people are getting so excited about the Librems's hardware, the real conundrum to me is the software stack:
Showing out a GNU/Linux desktop on an ARM developer board is nothing new, but without a large set of "touch"-capable applications for miniature screens, this will presumably be in 2019, where Openmoko's software distributions ended (or where a "naked" Mer/Nemo is): Not really what one expects as a Smartphone, today (or in 2019).

Has anybody heard or read about specific plans for the basic software stack and for addressing the lack of "mobile" apps?

Last edited by olf; 2017-10-10 at 17:32.
 

The Following 5 Users Say Thank You to olf For This Useful Post: