Replace `nalgebra` with `glam`

`glam` compiles much faster and leverages SIMD nicely.

4 files changed, 20 insertions, 25 deletions

diff --git a/wgpu/Cargo.toml b/wgpu/Cargo.toml
index 30f9224f..04fae248 100644
--- a/wgpu/Cargo.toml
+++ b/wgpu/Cargo.toml
@@ -13,5 +13,5 @@ wgpu = { version = "0.3", git = "https://github.com/gfx-rs/wgpu-rs", rev = "ed2c
 wgpu_glyph = { version = "0.4", git = "https://github.com/hecrj/wgpu_glyph", rev = "954ac865ca1b7f6b97bf403f8c6174a7120e667c" }
 raw-window-handle = "0.3"
 image = "0.22"
-nalgebra = "0.18"
+glam = "0.8"
 log = "0.4"
diff --git a/wgpu/src/image.rs b/wgpu/src/image.rs
index c42e1cd4..75cfa166 100644
--- a/wgpu/src/image.rs
+++ b/wgpu/src/image.rs
@@ -222,11 +222,9 @@ impl Pipeline {
         bounds: Rectangle<u32>,
         target: &wgpu::TextureView,
     ) {
-        let matrix: [f32; 16] = transformation.into();
-
         let transform_buffer = device
             .create_buffer_mapped(16, wgpu::BufferUsage::COPY_SRC)
-            .fill_from_slice(&matrix[..]);
+            .fill_from_slice(transformation.as_ref());
 
         encoder.copy_buffer_to_buffer(
             &transform_buffer,
diff --git a/wgpu/src/quad.rs b/wgpu/src/quad.rs
index 6365e117..bfbd7e2d 100644
--- a/wgpu/src/quad.rs
+++ b/wgpu/src/quad.rs
@@ -23,14 +23,12 @@ impl Pipeline {
                 }],
             });
 
-        let matrix: [f32; 16] = Transformation::identity().into();
-
         let transform = device
             .create_buffer_mapped(
                 16,
                 wgpu::BufferUsage::UNIFORM | wgpu::BufferUsage::COPY_DST,
             )
-            .fill_from_slice(&matrix[..]);
+            .fill_from_slice(Transformation::identity().as_ref());
 
         let constants = device.create_bind_group(&wgpu::BindGroupDescriptor {
             layout: &constant_layout,
@@ -169,11 +167,9 @@ impl Pipeline {
         bounds: Rectangle<u32>,
         target: &wgpu::TextureView,
     ) {
-        let matrix: [f32; 16] = transformation.into();
-
         let transform_buffer = device
             .create_buffer_mapped(16, wgpu::BufferUsage::COPY_SRC)
-            .fill_from_slice(&matrix[..]);
+            .fill_from_slice(transformation.as_ref());
 
         encoder.copy_buffer_to_buffer(
             &transform_buffer,
diff --git a/wgpu/src/transformation.rs b/wgpu/src/transformation.rs
index 53583e7e..b0d14cc8 100644
--- a/wgpu/src/transformation.rs
+++ b/wgpu/src/transformation.rs
@@ -1,29 +1,30 @@
-use nalgebra::Matrix3;
+use glam::{Mat4, Vec3, Vec4};
 use std::ops::Mul;
 
 /// A 2D transformation matrix.
 #[derive(Debug, Clone, Copy, PartialEq)]
-pub struct Transformation(Matrix3<f32>);
+pub struct Transformation(Mat4);
 
 impl Transformation {
     /// Get the identity transformation.
     pub fn identity() -> Transformation {
-        Transformation(Matrix3::identity())
+        Transformation(Mat4::identity())
     }
 
     /// Creates an orthographic projection.
     #[rustfmt::skip]
     pub fn orthographic(width: u16, height: u16) -> Transformation {
-        Transformation(nalgebra::Matrix3::new(
-            2.0 / f32::from(width), 0.0, -1.0,
-            0.0, 2.0 / f32::from(height), -1.0,
-            0.0, 0.0, 1.0
+        Transformation(Mat4::from_cols(
+            Vec4::new(2.0 / f32::from(width), 0.0, 0.0, 0.0),
+            Vec4::new(0.0, 2.0 / f32::from(height), 0.0, 0.0),
+            Vec4::new(0.0, 0.0, -1.0, 0.0),
+            Vec4::new(-1.0, -1.0, 0.0, 1.0)
         ))
     }
 
     /// Creates a translate transformation.
     pub fn translate(x: f32, y: f32) -> Transformation {
-        Transformation(Matrix3::new_translation(&nalgebra::Vector2::new(x, y)))
+        Transformation(Mat4::from_translation(Vec3::new(x, y, 0.0)))
     }
 }
 
@@ -35,14 +36,14 @@ impl Mul for Transformation {
     }
 }
 
+impl AsRef<[f32; 16]> for Transformation {
+    fn as_ref(&self) -> &[f32; 16] {
+        self.0.as_ref()
+    }
+}
+
 impl From<Transformation> for [f32; 16] {
-    #[rustfmt::skip]
     fn from(t: Transformation) -> [f32; 16] {
-        [
-            t.0[0], t.0[1], 0.0, t.0[2],
-            t.0[3], t.0[4], 0.0, t.0[5],
-            0.0, 0.0, -1.0, 0.0,
-            t.0[6], t.0[7], 0.0, t.0[8]
-        ]
+        t.as_ref().clone()
     }
 }