Accelerating DirectFB

Hi,

Where you are planning to use directfb? Can you please provide some details on your targeted use case?

You can also refer to some of the below articles for directfb Neon optimizations-

http://blogs.arm.com/software-enablement/624-optimizing-directfb-with-arm-neon/

https://blueprints.launchpad.net/linaro-multimedia-project/+spec/engr-multimedia-directfb-neon-optimization

With regards to SGX, Qt can be accelerated & also we have Xorg DRI acceleration supported with our graphics SDK releases.

http://processors.wiki.ti.com/index.php/RN_4_08_00_01#What_is_supported

Thanks,

Prathap.

Hi Prathap,

this project is for a home appliance. It was originally developed on a Freescale's ARM9 and now I'm porting it to an AM3352. To do this work I'm using an AM335x-SK, but I've changed the orignial TFT/TouchScreen by a TFT with capacitive touchscreen.

On the ARM9, with RGB565 graphs, worst case CPU load was around 75%. On the AM335x-SK,  I'm getting a worst case load around 52%. My target was to get better numbers on the AM335x-SK, as this will be our first release and I'm sure things will get more CPU demandant in the future, so was thinking on using some kind of hardware acceleration.

At this moment we cannot think on migrating from DirectFB - the software is already running (some minor tweaks pending) and we have no time to afford QT, GTK+ or whatever else.

I'm used to ARM9 processor but I'm new to TI & A8 processors, so thank you very much for your links. At first look they are plenty of valuable information for me, so I'll read them more deeply.

Thanks,

Gonzalo.