Fastest codepath for per-pixel updates with double buffering?

[psy0rz]

For my project, i'm updating an entire 64x32 matrix at 60fps.

Would it be faster to directly call

ESP32-HUB75-MatrixPanel-DMA/src/ESP32-HUB75-MatrixPanel-I2S-DMA.cpp

Line 328 in 447254d

void IRAM_ATTR MatrixPanel_I2S_DMA::updateMatrixDMABuffer(uint16_t x_coord, uint16_t y_coord, uint8_t red, uint8_t green, uint8_t blue)

(its a protected function)

instead of going through drawPixelRGB888(), which also doesn't have IRAM_ATTR?

I already use NO_GFX and NO_CIE1931, since animations are created on a server and color correction is also done there.

Also: Is this busyloop needed and a significant amount?

ESP32-HUB75-MatrixPanel-DMA/src/platforms/esp32/esp32_i2s_parallel_dma.cpp

Line 613 in 447254d

while (!previousBufferFree);

Perhaps we can make it skippable so we can preprocess the next frame?

(This is how i'm using the library: https://github.com/psy0rz/ledstream/blob/hub75/main/leds_hub75.cpp )

[psy0rz]

Also its pretty impressive what this library can do with a ESP32 and a HUB75 display, just by using clever DMA tricks instead of extra hardware. :)

[J-Swift]

Found this on a google search wondering sort of the same. What did you end up figuring out @psy0rz?

[psy0rz]

Nothing special :)

[J-Swift]

Haha ok well thanks for posting your code, im sure ill get some pointers from it