fillrate.md with some test details

Signed-off-by: Emil Muratov <gpm@hotplug.ru>
This commit is contained in:
Emil Muratov 2020-12-07 16:41:33 +03:00
parent e8efa3d29d
commit f92de5e81e

25
fillrate.md Normal file
View file

@ -0,0 +1,25 @@
## Estimating fillrate
Here are some results of simple tests on filling DMA buffer with data.
Filling DMA buffer requres lot's of memory operations on a bit level rather than doing simple byte/word wide store and copy. And it looks like it's quite a task both for esp32 core and compiler.
I've done this while optimizing loops and bit logic along with testing compiler results.
So the testbed is:
- Matrix modules: 4 x FM6126A based 64x64 modules chained in 256x64
A simple sketch:
- allocating single DMA buffs for 256x64
- allocating (NUM_LEDS*3) bytes for CRGB buffer
- measuring microseconds for the following calls:
- clearScreen() - full blanking
- fillScreenRGB888() with monochrome/gray colors
- filling some gradient into CRGB buff
- painting CRGB buff into DMA buff with looped drawPixelRGB888()
|Pattern |Reference|Ref+SPEED|Optimized|Optimized+SPEED|
|--|--|--|--|--|
|fillScreenRGB888()|14570|14570|13400 (8.5% faster)|5520 (164% faster)|
|CRGB buff fill|760|760|760|760|
|updateMatrixDMABuffer(CRGB)|7700|32080|55780 (38% faster)|33500(+4.2% slower)|
to be continued...