I’m not sure if this will work. But with flexio + pll + dma i run 4096 leds & 178 fps. This uses 3 pins, 1 for data to external 32bit shifter. You can output to more than one bit, so you can try to output 8 bits. There are 3 flexio, but only 1 and 2 can use dma. So that maybe thats 16 bits...