fast math 2

crates/bevy_pbr/src/atmosphere/functions.wgsl

@@ -1,5 +1,7 @@
 #define_import_path bevy_pbr::atmosphere::functions
 
+#import bevy_render::maths::{PI, HALF_PI, PI_2}
+
 #import bevy_pbr::atmosphere::{
     types::Atmosphere,
     bindings::{
@@ -35,9 +37,6 @@
 
 // CONSTANTS
 
-const PI: f32 = 3.141592653589793238462;
-const TAU: f32 = 6.283185307179586476925;
-const FRAC_PI: f32 = 0.31830988618379067153; // 1 / π
 const FRAC_3_16_PI: f32 = 0.0596831036594607509; // 3 / (16π)
 const FRAC_4_PI: f32 = 0.07957747154594767; // 1 / (4π)
 

crates/bevy_pbr/src/atmosphere/multiscattering_lut.wgsl

@@ -5,7 +5,7 @@
         bindings::{atmosphere, settings},
         functions::{
             multiscattering_lut_uv_to_r_mu, sample_transmittance_lut, isotropic, 
-            get_local_r, get_local_up, sample_atmosphere, FRAC_4_PI, TAU,
+            get_local_r, get_local_up, sample_atmosphere, FRAC_4_PI,
             max_atmosphere_distance, rayleigh, henyey_greenstein
         },
         bruneton_functions::{
@@ -14,6 +14,8 @@
     }
 }
 
+#import bevy_render::maths::PI_2
+
 
 const PHI_2: vec2<f32> = vec2(1.3247179572447460259609088, 1.7548776662466927600495087);
 
@@ -26,7 +28,7 @@ fn s2_sequence(n: u32) -> vec2<f32> {
 
 //Lambert equal-area projection. 
 fn uv_to_sphere(uv: vec2<f32>) -> vec3<f32> {
-    let phi = TAU * uv.y;
+    let phi = PI_2 * uv.y;
     let sin_lambda = 2 * uv.x - 1;
     let cos_lambda = sqrt(1 - sin_lambda * sin_lambda);
 

crates/bevy_pbr/src/render/fast_math.wgsl

@@ -1,24 +1,33 @@
 #define_import_path bevy_pbr::fast_math
 
-#import bevy_render::maths::PI;
+#import bevy_render::maths::{PI, HALF_PI};
 
-// Reference: XEngine
-// https://github.com/ShawnTSH1229/XEngine/blob/main/Source/Shaders/FastMath.hlsl
+fn fast_sqrt(x: f32) -> f32 {
+    return bitcast<f32>(0x1fbd1df5 + (bitcast<i32>(x) >> 1u));
+}
+
+//slightly less accurate than fast_acos_4, but much simpler.
+fn fast_acos(in_x: f32) -> f32 {
+    let x = abs(in_x);
+    var res = -0.156583 * x + HALF_PI;
+    res *= fast_sqrt(1.0 - x);
+    return select(PI - res, res, in_x >= 0.0);
+}
 
 // 4th order polynomial approximation
 // 4 VGRP, 16 ALU Full Rate
 // 7 * 10^-5 radians precision
 // Reference : Handbook of Mathematical Functions (chapter : Elementary Transcendental Functions), M. Abramowitz and I.A. Stegun, Ed.
-fn fast_acos(x: f32) -> f32 {
+fn fast_acos_4(x: f32) -> f32 {
     let x1 = abs(x);
     let x2 = x1 * x1;
     let x3 = x2 * x1;
-    var s;
+    var s: f32;
 
     s = -0.2121144 * x1 + 1.5707288;
     s = 0.0742610 * x2 + s;
     s = -0.0187293 * x3 + s;
-    s = sqrt(1.0 - x1) * s;
+    s = fast_sqrt(1.0 - x1) * s;
 
 	// acos function mirroring
     return select(PI - s, s, x >= 0.0);
@@ -37,7 +46,7 @@ fn fast_atan2(y: f32, x: f32) {
 
     t3 = select(t3, (0.5 * PI) - t3, abs(y) > abs(x));
     t3 = select(t3, PI - t3, x < 0);
-    t3 *= sign(y)
+    t3 = select(-t3, t3, y > 0);
 
     return t3;
 }

crates/bevy_pbr/src/ssao/ssao.wgsl

@@ -10,7 +10,7 @@
 // Source code base on SSRT3 implementation
 // https://github.com/cdrinmatane/SSRT3
 
-#import bevy_pbr::ssao_utils::fast_acos
+#import bevy_pbr::fast_math::fast_acos
 
 #import bevy_render::{
     view::View,

crates/bevy_pbr/src/ssao/ssao_utils.wgsl

@@ -11,14 +11,3 @@ fn ssao_multibounce(visibility: f32, base_color: vec3<f32>) -> vec3<f32> {
     let x = vec3<f32>(visibility);
     return max(x, ((x * a + b) * x + c) * x);
 }
-
-fn fast_sqrt(x: f32) -> f32 {
-    return bitcast<f32>(0x1fbd1df5 + (bitcast<i32>(x) >> 1u));
-}
-
-fn fast_acos(in_x: f32) -> f32 {
-    let x = abs(in_x);
-    var res = -0.156583 * x + HALF_PI;
-    res *= fast_sqrt(1.0 - x);
-    return select(PI - res, res, in_x >= 0.0);
-}