animartrix speedup: crank up math optimizations
may reduce accuracy, but gives up to 20% speedup in animartrix effects
This commit is contained in:
@@ -54,8 +54,10 @@
|
||||
#if defined(ARDUINO_ARCH_ESP32) && defined(ESP_IDF_VERSION)
|
||||
#if ESP_IDF_VERSION >= ESP_IDF_VERSION_VAL(4, 4, 0)
|
||||
// this pragma temporarily raises gcc optimization level to "-O3", to avoid internal error conditions
|
||||
// -fsingle-precision-constant = all literals are float (default is "double" = very slow)
|
||||
// -ffast-math = additional math optimizations that may reduce accurcy
|
||||
#pragma GCC push_options
|
||||
#pragma GCC optimize ("O3")
|
||||
#pragma GCC optimize ("O3,single-precision-constant,fast-math")
|
||||
#endif
|
||||
#endif
|
||||
|
||||
|
||||
Reference in New Issue
Block a user