can't lower nema buffer pool memory usage

[Jcoop.1]

Hi,

On using svg's on an U5f9, i try to activate the hardware acceleration but this demands to much free memory of my project. 

the nema buffer create pool function demands 614K of memory, independent of the MEM_POOL_SIZE or the STENCIL_POOL_SIZE. is there a way to lower this usage or is it connected to the display resolution (614K is exactly half of my framebuffersize)

Thanks in advance

[GaetanGodart · Accepted solution]

Hello @Jcoop.1 ,

The stencil buffer is required and cannot be reduced.

It seems you want to use quite a big display (1024x600 or 960x640) which is ok but it obviously need a lot of memory for the framebuffer.

What you could do is to use a single framebuffer strategy. If you use LTDC display, since they can "track" which line is being written, it can act as a smart framebuffer where the lines that have already been written can be updated without tearing effect issue.

Another solution to reduce memory usage would be to reduce the pixel color depth to 24bpp or 16bpp.

I hope this helps! :smiling_face_with_smiling_eyes:
If this comment answers your question, I invite you to select it as "best answer".

Regards,

[Jcoop.1]

Hi Gaetan,

I use indeed a ultra wide screen 1280 x 480 screen,

and this in internal memory only,

so pretty constraint.

will try the singlebuffer and see what this gives.

Thanks for the info.

Regards

Jan

[GaetanGodart]

I am glad it helped!

If when using single framebuffer you find the performance not good enough, you can either reduce computation graphics such as SVG, transparency, etc or go the pixel bpp route.

Regards,