You were probably getting sick of me saying stuff like "we'll get to that when we talk about buffers". Well now's the time to finally talk about buffers, but first...
## What is a buffer?
A buffer is a blob of data on the GPU. A buffer is guaranteed to be contiguous, meaning that all the data is store sequentially in memory. Buffer's generally are used to store simple things like a struct or an array, but it can store more complex stuff such as graph structures like a tree (provided all the nodes are stored together and don't reference anything outside of the buffer). We are going to use buffer's a lot, so let's get started with two of the most important one's: the vertex buffer, and the index buffer.
## The vertex buffer
Previously we've stored vertex data directly in the vertex shader. While that worked fine to get our bootstraps on, it simply won't do longterm. The types of objects we need to draw will very in size, and recompiling the shader whenever we need to update the model would massively slow down our program. Instead we are going to use buffers to store the vertex data we want to draw. Before we do that though we need to describe what a vertex looks like. We'll do this by creating a new struct.
```rust
// main.rs
#[repr(C)]
#[derive(Copy, Clone, Debug)]
struct Vertex {
position: [f32; 3],
color: [f32; 3],
}
```
Our vertices will all have a position and a color. The position represents the x, y, and z of the vertex in 3d space. The color is the red, greed, and blue values for the vertex. We need the `Vertex` to be copyable so we can create a buffer with it.
Next we need the actual data to will make up our triangle. Below `Vertex` add the following.
We arrange the vertices in counter clockwise order: top, bottom left, bottom right. We do it this way partially out of tradition, but mostly because we specified in the `rasterization_state` of the `render_pipeline` that we want the `front_face` of our triangle to be `wgpu::FrontFace::Ccw` so that we cull the back face. This means that any triangle that should be facing us should have it's vertices in counter clockwise order.
Another thing to note is that the y coordinate is negative for up and positive for down. This may seem weird to those of you coming from an OpenGL perspective, or even more traditional mathematics, but it's actually quite common for down to be positive on the y axis in graphics. We can change this later using a projection matrix, but we'll leave it as is for now.
Now that we have our vertex data, we need to store it in a buffer. Let's add a `vertex_buffer` field to `State`.
You'll note that we're using [bytemuck](https://docs.rs/bytemuck/1.2.0/bytemuck/) to cast our `VERTICES`. The `create_buffer_with_data()` method expects a `&[u8]`, and `bytemuck::cast_slice` does that for us. Add the following to your `Cargo.toml`.
We're also going to need to implement two traits to get `bytemuck` to work. These are [bytemuck::Pod](https://docs.rs/bytemuck/1.2.0/bytemuck/trait.Pod.html) and [bytemuck::Zeroable](https://docs.rs/bytemuck/1.2.0/bytemuck/trait.Zeroable.html). `Pod` indicates that our `Vertex` is "Plain Old Data", and thus can be interpretted as a `&[u8]`. `Zeroable` indicates that we can use `std::mem::zeroed()`. These traits don't require us to implement any methods, so we just need to use the following to get our code to work.
```rust
unsafe impl bytemuck::Pod for Vertex {}
unsafe impl bytemuck::Zeroable for Vertex {}
```
Finally we can add our `vertex_buffer` to our `State` struct.
We need to tell the `render_pipeline` to use this buffer when we are drawing, but first we need to tell the `render_pipeline` how to read the buffer. We do this using `VertexBufferDescriptor`s and the `vertex_buffers` field that I promised we'd talk about when we created the `render_pipeline`.
A `VertexBufferDescriptor` define's how a buffer is layed out in memory. Without this, the render_pipeline has no idea how to map the buffer in the shader. Here's what the descriptor for a buffer full of `Vertex` would look like.
```rust
use std::mem;
wgpu::VertexBufferDescriptor {
stride: mem::size_of::<Vertex>() as wgpu::BufferAddress, // 1.
step_mode: wgpu::InputStepMode::Vertex, // 2.
attributes: &[ // 3.
wgpu::VertexAttributeDescriptor {
offset: 0, // 4.
shader_location: 0, // 5.
format: wgpu::VertexFormat::Float3, // 6.
},
wgpu::VertexAttributeDescriptor {
offset: mem::size_of::<[f32; 3]>() as wgpu::BufferAddress,
shader_location: 1,
format: wgpu::VertexFormat::Float3,
}
]
}
```
1. The `stride` defines how wide a vertex is. When the shader goes to read the next vertex, it will skip over `stride` number of bytes. In our case, stride will probably be 24 bytes.
2.`step_mode` tells the pipeline how often it should move to the next vertex. This seems redundant in our case, but we can specify `wgpu::InputStepMode::Instance` if we only want the change vertices when we start drawing a new instance. We'll cover instancing in a later tutorial.
3. Vertex attributes describe the individual parts of the vertex. Generally this is a 1:1 mapping with a structs fields which it is in our case.
4. This defines the `offset` in bytes that this attribute starts. The first attribute is usually zero, and any future attributes are the collective `size_of` the previous attributes data.
5. This tells the shader what location to store this attribute at. For example `layout(location=0) in vec3 x` in the vertex shader would correspond to the position field of the struct, while `layout(location=1) in vec3 x` would be the color field.
6.`format` tells the shader the shape of the attribute. `Float3` corresponds to `vec3` in shader code. The max value we can store in an attribute is `Float4` (`Uint4`, and `Int4` work as well). We'll keep this in mind for when we have to store things that are bigger than `Float4`.
For you visually learners, our vertex buffer looks like this.
Let's create a static method on `Vertex` that returns this descriptor.
Before we continue, we should change the `render_pass.draw()` call to use the number of vertices specified by `VERTICES`. Add a `num_vertices` to `State`, and set it to be equal to `VERTICES.len()`.
Before our changes will have any effect, we need to update our vertex shader to get it's data from the vertex buffer.
```glsl
// shader.vert
#version 450
layout(location=0) in vec3 a_position;
layout(location=1) in vec3 a_color;
layout(location=0) out vec3 v_color;
void main() {
v_color = a_color;
gl_Position = vec4(a_position, 1.0);
}
```
We'll want to update the fragment shader to use `v_color` as well.
```glsl
// shader.frag
#version 450
layout(location=0) in vec3 v_color;
layout(location=0) out vec4 f_color;
void main() {
f_color = vec4(v_color, 1.0);
}
```
If you've done things correctly, you should see a triangle that looks something like this.
## The index buffer
We technically don't *need* an index buffer, but they still are plenty useful. An index buffer comes into play when we start using models with a lot of triangles. Consider this pentagon.
It has a total of 5 vertices, and 3 triangles. Now if we wanted to display something like this using just vertices we would need something like the following.
You'll note though that some of the vertices are used more than once. C, and B get used twice, and E is repeated 3 times. Assuming that each float is 4 bytes, then that means of the 216 bytes we use for `VERTICES`, 96 of them are duplicate data. Wouldn't it be nice if we could list these vertices once? Well we can! That's were an index buffer comes into play.
Basically we store all the unique vertices in `VERTICES` and we create another buffer that stores indices to elements in `VERTICES` to create the triangles. Here's an example of that with our pentagon.
Now with this setup our `VERTICES` take up about 120 bytes and `INDICES` is just 18 bytes given that `u16` is 2 bytes wide. All together our pentagon is 132 bytes in total. That means we saved 84 bytes! It may not seem like much, but when dealing with tri counts in the hundreds of thousands, indexing saves a lot of memory.
There's a couple of things we need to change in order to use indexing. The first is we need to create a buffer to store the indices. In `State`'s `new()` method create the `index_buffer` after you create the `vertex_buffer`. Also change `num_vertices` to `num_indices` and set it equal to `INDICES.len()`.
We don't need to implement `Pod` and `Zeroable` for our indices, because `bytemuck` has already done that for us. That means we can just add `index_buffer` and `num_indices` to `State`.
2. When using an index buffer, you need to use `draw_indexed`. The `draw` method ignores the index buffer. Also make sure you use the number of indices (`num_indices`), not vertices as you model will either draw wrong, or the method will `panic` because there are not enough indices. Last thing to note about this method is that the second parameter specifies what index to start at in the buffer. This could allow you to store multiple sets of indices in one buffer. The last parameter is the size of the buffer. Passing 0 tells wgpu to use the entire buffer.
Create a more complex shape than the one we made (aka. more than three triangles) using a vertex buffer and an index buffer. Toggle between the two with the space key.
