Aleš Katona

So from what I understand there are a few issues here: - `display-interface` forcing 64 bit buffer size - `display-interface` forcing endianness runtime conversion - `display-interface` using dynamic dispatch Optionally...

One thing that `mipidsi` can do here is to introduce something more than just `draw_pixels` to the model. If we allowed something like `write_pixels_raw_u8()` with `&[u8]` and exposed or hint-switch...

@rfuest Lmk what you think about https://github.com/almindor/mipidsi/pull/143 @fu5ha Thanks I'll have a look!

@akx could you please try the `Display::set_pixels_from_buffer()` (need to use latest master) to see how much of a difference it makes?

> Not a great improvement That's unfortunate. I'd expect a much better improvement. I got myself esp32-C6, once I have some time I'll try to experiment myself and see where...

Ok so for me, using the `esp32-C6` with SPI set to `80Mhz` and a `320x240` `ST7789` display with CS: * I can get to `~30ms` fullscreen draw time when using...

I did some more testing and if I deconstruct the SPI back after sending all the init stuff and prepping a data window, I'm able to send the raw buffer...

@rfuest @akx Found it! Note I used only `esp-hal` not the IDF/std abstraction. The main issues are: 1. CPU default of esp-hal is 80Mhz, so make sure to set it...

I'll look at what I get with my C6, I've been doing some work with it lately. When it comes to async, it's hard. I did some experimenting and found...

I just tested with my esp32-C6 and I get ~18ms for full screen clear on a `ST7789`. I use this code ```rust #![no_std] #![no_main] // use display_interface_spi::SPIInterface as SpiInterface; use...