zenfg 0.1.0

A FrameGraph compiler and transient-resource executor for wgpu
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166

use std::{env, hint::black_box, time::Instant};

use zenfg::{
    BufferDesc, BufferRange, CompileOptions, FrameGraph, RootReason, TextureDesc, TextureViewDesc,
    WriteContents,
};

fn main() {
    let scenario = env::args().nth(1).unwrap_or_else(|| "linear-chain".into());
    let node_count = env::args()
        .nth(2)
        .and_then(|value| value.parse::<usize>().ok())
        .unwrap_or(64)
        .max(2);
    let iterations = env::args()
        .nth(3)
        .and_then(|value| value.parse::<usize>().ok())
        .unwrap_or(1_000);

    let start = Instant::now();
    for _ in 0..iterations {
        black_box(run(&scenario, node_count));
    }
    let elapsed = start.elapsed();
    println!(
        "scenario={scenario} nodes={node_count} iterations={iterations} total_ms={:.3} mean_us={:.3}",
        elapsed.as_secs_f64() * 1_000.0,
        elapsed.as_secs_f64() * 1_000_000.0 / iterations as f64
    );
}

fn run(scenario: &str, node_count: usize) -> usize {
    match scenario {
        "linear-chain" => linear_chain(node_count),
        "buffer-ranges" => buffer_ranges(node_count),
        "texture-subresources" => texture_subresources(node_count),
        "allocation-aliasing" => allocation_aliasing(node_count),
        _ => panic!(
            "unknown scenario {scenario}; expected linear-chain, buffer-ranges, texture-subresources, or allocation-aliasing"
        ),
    }
}

fn linear_chain(node_count: usize) -> usize {
    let mut graph = FrameGraph::new();
    let mut frame = graph.begin_frame();
    let buffers: Vec<_> = (0..node_count)
        .map(|index| {
            frame
                .create_buffer(BufferDesc::new(format!("b{index}"), 64))
                .unwrap()
        })
        .collect();
    for index in 0..node_count {
        let mut pass = frame.compute_pass(format!("p{index}"));
        if index > 0 {
            let _ = pass
                .storage_buffer_read(buffers[index - 1], BufferRange::whole())
                .unwrap();
        }
        let _ = pass
            .storage_buffer_write(
                buffers[index],
                BufferRange::whole(),
                WriteContents::Overwrite,
            )
            .unwrap();
        pass.finish().unwrap();
    }
    frame
        .mark_buffer_root(
            *buffers.last().unwrap(),
            BufferRange::whole(),
            RootReason::Output,
        )
        .unwrap();
    frame
        .compile(CompileOptions::default())
        .unwrap()
        .retained_node_count()
}

fn buffer_ranges(node_count: usize) -> usize {
    let mut graph = FrameGraph::new();
    let mut frame = graph.begin_frame();
    let buffer = frame
        .create_buffer(BufferDesc::new("ranges", (node_count * 4) as u64))
        .unwrap();
    for index in 0..node_count {
        let mut pass = frame.compute_pass(format!("write-{index}"));
        let _ = pass
            .storage_buffer_write(
                buffer,
                BufferRange::new((index * 4) as u64, 4),
                WriteContents::Overwrite,
            )
            .unwrap();
        pass.finish().unwrap();
    }
    let mut read = frame.command_pass("read-all");
    let _ = read
        .storage_buffer_read(buffer, BufferRange::whole())
        .unwrap();
    read.finish().unwrap();
    frame
        .compile(CompileOptions::default())
        .unwrap()
        .retained_node_count()
}

fn texture_subresources(node_count: usize) -> usize {
    let layers = node_count.min(1_024) as u32;
    let mut graph = FrameGraph::new();
    let mut frame = graph.begin_frame();
    let mut desc = TextureDesc::new_2d("layers", 16, 16, wgpu::TextureFormat::Rgba8Unorm);
    desc.size.depth_or_array_layers = layers;
    let texture = frame.create_texture(desc).unwrap();
    for layer in 0..layers {
        let view = frame
            .create_texture_view(
                texture,
                TextureViewDesc {
                    label: format!("layer-{layer}"),
                    base_array_layer: layer,
                    array_layer_count: Some(1),
                    ..Default::default()
                },
            )
            .unwrap();
        let mut pass = frame.compute_pass(format!("write-{layer}"));
        let _ = pass
            .storage_texture_write(view, WriteContents::Overwrite)
            .unwrap();
        pass.finish().unwrap();
    }
    let mut read = frame.command_pass("read-all");
    let _ = read.sampled_texture(texture).unwrap();
    read.finish().unwrap();
    frame
        .compile(CompileOptions::default())
        .unwrap()
        .retained_node_count()
}

fn allocation_aliasing(node_count: usize) -> usize {
    let mut graph = FrameGraph::new();
    let mut frame = graph.begin_frame();
    for index in 0..node_count {
        let buffer = frame
            .create_buffer(BufferDesc::new(format!("scratch-{index}"), 1_024))
            .unwrap();
        let mut pass = frame.compute_pass(format!("write-{index}"));
        let _ = pass
            .storage_buffer_write(buffer, BufferRange::whole(), WriteContents::Overwrite)
            .unwrap();
        pass.finish().unwrap();
        frame
            .mark_buffer_root(buffer, BufferRange::whole(), RootReason::Output)
            .unwrap();
    }
    frame
        .compile(CompileOptions::full_report())
        .unwrap()
        .allocations()
        .len()
}