@yakabmarci Thank you for sharing. Maybe the state of this library is the reason why it is not shared. To stay competitive ST had to immediately add GPU to their products, top executives shook hands and ST purchased what appeared to be a good choice, maybe based on the promise that quality will be managed and now they are ashamed to release it because it is far from the release-ready state while vendor (NEMA) believes otherwise so ST has to come up with nonsense excuses simply because releasing this masterpiece software would do ST more harm than not releasing it.
business has no shame, I think that it is a conscious functionality limitation. Here it may makes a bit of sense as it favours TouchGFX. But have you seen the neural accelerator in N6? It is a very powerful thing and apart of neural network acceleration could be very efficiently used for signal processing. It contains so many multipliers so could calculate FIR in a single cycle! But no, it may be used only with a neural net model from a limited number (3) of suppliers. Just like the GPU it has no HAL drivers and cannot be used freely. So @TDJ I am sure that it is conscious.
