Ubuntu 12.10 Sets To Make ARM Even Stronger

[TobiSGD]

Well in progress of making Ubuntu better for ARM they are also making Linux better for ARM by working upstream with Linaro. They are in no way trying to make it harder for other distributions to support ARM.

But they also don't make it easier. Be real, the amount of upstream contributions coming from Canonical/Ubuntu is ridiculously low for the amount of users and developers they have in their community.

[Teho]

But they also don't make it easier. Be real, the amount of upstream contributions coming from Canonical/Ubuntu is ridiculously low for the amount of users and developers they have in their community.

How could they make it easier? By compiling packages for other distributions? I do understand the critique that Canonical is getting for the crap they are pulling with Gnome/GTK+ but this has nothing to do with it. Canonical/Ubuntu does a lot to advance Linux on ARM servers and contribute directly to upstream so I don't see what the problem is here if there's any. Sure they could do more in many departments but I doubt that ARM is one of them.

[uid313]

Some regressions, but also some very impressive performance improvements which are twice as fast!

How about ARMv8 (64-bit) support?

[johnc]

I'm a bit surprised to see the idle power so high. I thought these things idled at like zero watts. I guess it depends on the governor of course.

[popper]

I was wrong, here it is: http://wiki.debian.org/ArmHardFloatPort

it's alright darkbasic armhf had only really been officially part of Debian! since Sat Nov 26, 2011 where markos and Hector did the hard work on/for the Freescale cortex A8/Genesi Efika MX (armhf) kit.

http://www.powerdeveloper.org/forums...pic.php?t=2212
and so all the other Linaro membership also took it up etc .

im surprised Michael still hasn't got around to testing a Linaro Optimized build yet even if most of markos's (aka Konstantinos Margaritis Senior Software Engineer, NEON optimizations Genesi USA Inc" NEON optimizations are not widely used yet and so not getting best speed given you are going to get better speed/more spare cycles from any app using NEON SIMD rather than just generic hardfloat generally.

[JanC]

I'm a bit surprised to see the idle power so high. I thought these things idled at like zero watts. I guess it depends on the governor of course.

I doubt a full Ubuntu desktop system that is also polling an USB device can really "idle".

Also, while the SoC might (almost) idle, the Pandaboard also contains a whole bunch of peripheral chips that draw power too.

[popper]

I'm a bit surprised to see the idle power so high. I thought these things idled at like zero watts. I guess it depends on the governor of course.

i imagine Michael is using a monster hard drive to boot and run the tests and so drawing that 4watts just with that drive connected and running or his power meters crap OC or maybe the PandaBoard ES design is less than optimal ,it was primarily for development when it was released after all.

[popper]

Some regressions, but also some very impressive performance improvements which are twice as fast!

How about ARMv8 (64-bit) support?

the Linaro membership do have some ARMv8 sample/dev boards ready and are working some patches but don't expect to see them any time soon as its mostly server related patches etc right now as i understand it, its not like you can buy any ARM v8 kit until next year or perhaps late December at best anyway.