enginerenderer 0.0.1

A zero-dependency offline rendering engine in pure Rust — CPU path tracing, BVH acceleration, 16-band spectral rendering, PBR materials, animation & video export.
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

use std::thread::available_parallelism;

use super::arch::compute_dispatch;
use super::arch::native_calls;

#[derive(Debug, Clone, Copy, Default)]
pub struct SimdFeatures {
    pub avx512f: bool,
    pub avx2: bool,
    pub avx: bool,
    pub fma: bool,
    pub sse4_2: bool,
    pub sse2: bool,
    pub neon: bool,
}

#[derive(Debug, Clone)]
pub struct CpuProfile {
    pub logical_cores: u8,
    pub l2_cache_kb: u32,
    pub has_ht: bool,
    pub simd_features: SimdFeatures,
    pub vector_width_bits: u32,
}

impl CpuProfile {
    pub fn detect() -> Self {
        let cfg = compute_dispatch::default_cpu_config();
        let simd = detect_simd();
        let vendor_scale = match cfg.vendor {
            compute_dispatch::Vendor::Amd => 1usize,
            compute_dispatch::Vendor::Intel => 1usize,
            compute_dispatch::Vendor::Apple => 1usize,
            compute_dispatch::Vendor::Unknown => 1usize,
        };
        let power_scale = 1usize;
        let budget_scale = ((cfg.frame_budget_us / 8_333).max(1)) as usize;
        let requested = available_parallelism()
            .map(|n| n.get())
            .unwrap_or(1)
            .max(cfg.worker_hint.max(1))
            .saturating_mul(vendor_scale)
            .saturating_mul(power_scale)
            .saturating_mul(budget_scale);
        let logical_cores =
            compute_dispatch::clamp_cpu_workers(requested.min(cfg.render_workers.max(1))) as u8;
        let l2_cache_kb = detect_l2_cache_kb();
        let physical = detect_physical_cores();
        let has_ht = (logical_cores as usize) > physical;
        let vector_width_bits = if simd.avx512f {
            512
        } else if simd.avx2 || simd.avx {
            256
        } else if simd.sse4_2 || simd.sse2 || simd.neon {
            128
        } else {
            64
        };
        Self {
            logical_cores,
            l2_cache_kb,
            has_ht,
            simd_features: simd,
            vector_width_bits,
        }
    }

    pub fn optimal_tile_width(&self) -> usize {
        if self.simd_features.avx512f {
            16
        } else if self.simd_features.avx2 || self.simd_features.avx {
            8
        } else if self.simd_features.sse4_2 || self.simd_features.sse2 || self.simd_features.neon {
            4
        } else {
            2
        }
    }

    pub fn log_summary(&self) {
        crate::runtime_log!(
            "cpu: cores={} l2={}KB ht={} vec={}bit",
            self.logical_cores,
            self.l2_cache_kb,
            self.has_ht,
            self.vector_width_bits,
        );
    }

    pub fn schedule(&self, work_items: usize) -> compute_dispatch::Schedule {
        compute_dispatch::build_cpu_schedule(work_items)
    }
}

fn detect_simd() -> SimdFeatures {
    let f = native_calls::host_detect_simd_features();
    SimdFeatures {
        avx512f: f.avx512f,
        avx2: f.avx2,
        avx: f.avx,
        fma: f.fma,
        sse4_2: f.sse4_2,
        sse2: f.sse2,
        neon: f.neon,
    }
}

fn detect_l2_cache_kb() -> u32 {
    native_calls::host_detect_l2_cache_kb()
}

fn detect_physical_cores() -> usize {
    native_calls::host_detect_physical_cores()
}

#[derive(Debug, Clone)]
pub struct CoreSnapshot {
    pub core_id: u32,
    pub frequency_hz: u64,
}

pub fn detect_core_frequencies() -> Vec<CoreSnapshot> {
    native_calls::host_detect_core_frequencies()
        .into_iter()
        .map(|f| CoreSnapshot {
            core_id: f.core_id,
            frequency_hz: f.frequency_hz,
        })
        .collect()
}

pub fn thread_affinity_mask() -> usize {
    native_calls::host_thread_affinity_mask()
}

pub fn pin_thread_to_core(core_id: usize) {
    native_calls::host_pin_thread_to_core(core_id);
}