2D Graphics Processing Unit (GPU) on i.MX Processor

[mohamed_gaseen]

Hi,

I am currently working on 2D graphics data processing using the Graphics Processing Unit (GPU) on the Android operating system.

I would like to know if there is any recommended method or framework to stream or render 2D graphical models efficiently using the GPU on i.MX processors. The platforms I plan to implement this on include:

• i.MX93 EVK

• i.MX8M Plus (i.MX8MP)

• i.MX8QXP

My goal is to offload 2D graphical rendering tasks from the CPU to the GPU to take advantage of its parallel processing capabilities. While GPUs are typically used for graphics rendering, they are also increasingly being used for compute-intensive tasks such as machine learning and scientific processing.

Could you please let me know if there is any official documentation, SDK, or procedure available from your end to support GPU-accelerated 2D graphics processing on these platforms?

Looking forward to your guidance.

[Bio_TICFSL]

Hello,

From NXP you can find the Machine Learning user guide document:

https://www.nxp.com/docs/en/user-guide/IMX-MACHINE-LEARNING-UG.pdf

But for guidelines about parallel processing we don't have any documentation, you can do it with the NPU and GPU, but for this matter search on the web if somebody has an article about it.

Regards

[mohamed_gaseen]

Hello @Bio_TICFSL ,

I'm asking you that is there any drivers needed to be enabled for GPU, found reference nowhere.

Is there any detailed documentation or instruction notes regarding GPU available out there??

There's no such details about the GPU in the Machine Learning UG link.

My work is on Android 14.0.0_2.0.0 BSP, Also by the way coming to the NPU. I've referred the below source & it have mentioned that the NPU only supports for iMX8MP & IMX95 processors.

Is that mean the above mentioned processors doesn't work right??

https://www.nxp.com/docs/en/application-note/AN14411.pdf

Kindly clarify & sort out my issue ASAP.

Thank you in advance.

[Bio_TICFSL]

Hello,

If you intalles the full-image there is no need to install an external driver, the NPU works on Mx8MP, MX93 and MX95, and the application note you are seeing is applied to MX8MP and MX95EVK and it working nicely.

Regards

[mohamed_gaseen]

Hi @Bio_TICFSL,

I came across the NXP application note titled “Enabling eIQ Core NPU Delegates for i.MX Android Applications” (AN14411), which states that NPU support on Android OS is currently limited to the i.MX8MP and i.MX95 application processors.

Given this, I would like to clarify whether NPU acceleration is officially supported on the i.MX93 processor running Android. Could you please verify this?

Additionally, I believe there might have been some misunderstanding regarding my earlier request. My goal is to enable both the GPU and NPU on the Android platform, not on a full Linux image. Kindly confirm the available support and any relevant steps for enabling these components on Android for i.MX93.

But my requirement board is actually IMX 8MP.

Looking forward to your response and support.