trueno 0.16.4

High-performance SIMD compute library with GPU support for matrix operations
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
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284

#![allow(missing_docs)]
//! Builder for `TunerFeatures` with automatic normalization.

use crate::hardware::HardwareCapability;

use crate::tuner::types::{KernelType, QuantType};

use super::TunerFeatures;

// ============================================================================
// TunerFeaturesBuilder
// ============================================================================

/// Builder for TunerFeatures with automatic normalization.
#[derive(Default)]
pub struct TunerFeaturesBuilder {
    model_params_b: Option<f32>,
    hidden_dim: Option<u32>,
    num_layers: Option<u32>,
    num_heads: Option<u32>,
    head_dim: Option<u32>,
    vocab_size: Option<u32>,
    batch_size: Option<u32>,
    seq_len: Option<u32>,
    cuda_graphs: bool,
    kv_caches: Option<u32>,
    is_prefill: bool,
    quant_type: Option<QuantType>,
    kernel_type: Option<KernelType>,
    gpu_mem_bw_gbs: Option<f32>,
    gpu_compute_tflops: Option<f32>,
    gpu_sm_count: Option<u32>,
    gpu_l2_cache_mb: Option<f32>, // v1.1.0
    is_zero_copy: bool,           // v1.1.0
    measured_tps: Option<f32>,
}

impl TunerFeaturesBuilder {
    /// Set model size in billions of parameters
    pub fn model_params_b(mut self, params: f32) -> Self {
        self.model_params_b = Some(params);
        self
    }

    /// Set hidden dimension
    pub fn hidden_dim(mut self, dim: u32) -> Self {
        self.hidden_dim = Some(dim);
        self
    }

    /// Set number of layers
    pub fn num_layers(mut self, layers: u32) -> Self {
        self.num_layers = Some(layers);
        self
    }

    /// Set number of attention heads
    pub fn num_heads(mut self, heads: u32) -> Self {
        self.num_heads = Some(heads);
        self
    }

    /// Set head dimension
    pub fn head_dim(mut self, dim: u32) -> Self {
        self.head_dim = Some(dim);
        self
    }

    /// Set vocabulary size
    pub fn vocab_size(mut self, size: u32) -> Self {
        self.vocab_size = Some(size);
        self
    }

    /// Set batch size (M)
    pub fn batch_size(mut self, m: u32) -> Self {
        self.batch_size = Some(m);
        self
    }

    /// Set sequence length
    pub fn seq_len(mut self, len: u32) -> Self {
        self.seq_len = Some(len);
        self
    }

    /// Enable CUDA graphs
    pub fn cuda_graphs(mut self, enabled: bool) -> Self {
        self.cuda_graphs = enabled;
        self
    }

    /// Set number of KV caches
    pub fn kv_caches(mut self, count: u32) -> Self {
        self.kv_caches = Some(count);
        self
    }

    /// Set prefill mode
    pub fn is_prefill(mut self, prefill: bool) -> Self {
        self.is_prefill = prefill;
        self
    }

    /// Set quantization type
    pub fn quant_type(mut self, qt: QuantType) -> Self {
        self.quant_type = Some(qt);
        self
    }

    /// Set kernel type
    pub fn kernel_type(mut self, kt: KernelType) -> Self {
        self.kernel_type = Some(kt);
        self
    }

    /// Set GPU memory bandwidth in GB/s
    pub fn gpu_mem_bw_gbs(mut self, bw: f32) -> Self {
        self.gpu_mem_bw_gbs = Some(bw);
        self
    }

    /// Set GPU compute in TFLOPS
    pub fn gpu_compute_tflops(mut self, tflops: f32) -> Self {
        self.gpu_compute_tflops = Some(tflops);
        self
    }

    /// Set GPU SM count
    pub fn gpu_sm_count(mut self, count: u32) -> Self {
        self.gpu_sm_count = Some(count);
        self
    }

    /// Set measured throughput (for training data)
    pub fn measured_tps(mut self, tps: f32) -> Self {
        self.measured_tps = Some(tps);
        self
    }

    /// Set L2 cache size in MB (v1.1.0)
    pub fn gpu_l2_cache_mb(mut self, l2_mb: f32) -> Self {
        self.gpu_l2_cache_mb = Some(l2_mb);
        self
    }

    /// Set zero-copy memory path enabled (v1.1.0)
    pub fn is_zero_copy(mut self, enabled: bool) -> Self {
        self.is_zero_copy = enabled;
        self
    }

    /// Set hardware capability (auto-fills GPU features)
    #[allow(clippy::cast_precision_loss, clippy::cast_possible_truncation)]
    // SAFETY: GPU bandwidth/TFLOPS values fit in f32 (practical max ~10K); f64→f32 truncation negligible.
    pub fn hardware(mut self, hw: &HardwareCapability) -> Self {
        if let Some(gpu) = &hw.gpu {
            self.gpu_mem_bw_gbs = Some(gpu.memory_bw_gbps as f32);
            self.gpu_compute_tflops = Some(gpu.peak_tflops_fp32 as f32);
            // SM count not directly available; estimate from compute capability
            self.gpu_sm_count = None;
        }
        self
    }

    /// Build the feature vector with normalization
    #[allow(clippy::cast_precision_loss)]
    // SAFETY: All u32 values are model hyperparams (hidden_dim ≤ 16384, layers ≤ 128,
    //         heads ≤ 128, vocab ≤ 1M, batch ≤ 64, seq_len ≤ 32768) — well within f32 mantissa.
    /// Build features, panicking if raw inputs violate physical constraints.
    ///
    /// Use `try_build()` for fallible construction.
    pub fn build(self) -> TunerFeatures {
        self.try_build().expect("TunerFeatures: invalid raw input (see try_build for details)")
    }

    /// Build features, returning Err if raw inputs violate physical constraints.
    ///
    /// F024: model_params_b, gpu_mem_bw, gpu_compute must be non-negative.
    /// F024: Pre-normalization validation — physical quantities must be non-negative.
    fn validate_non_negative(&self) -> Result<(), crate::tuner::TunerError> {
        if let Some(p) = self.model_params_b {
            if p < 0.0 {
                return Err(crate::tuner::TunerError::InvalidFeature(format!(
                    "model_params_b must be non-negative, got {p}"
                )));
            }
        }
        if let Some(bw) = self.gpu_mem_bw_gbs {
            if bw < 0.0 {
                return Err(crate::tuner::TunerError::InvalidFeature(format!(
                    "gpu_mem_bw_gbs must be non-negative, got {bw}"
                )));
            }
        }
        if let Some(tf) = self.gpu_compute_tflops {
            if tf < 0.0 {
                return Err(crate::tuner::TunerError::InvalidFeature(format!(
                    "gpu_compute_tflops must be non-negative, got {tf}"
                )));
            }
        }
        Ok(())
    }

    pub fn try_build(self) -> Result<TunerFeatures, crate::tuner::TunerError> {
        self.validate_non_negative()?;

        let batch_size = self.batch_size.unwrap_or(1);
        let kv_caches = self.kv_caches.unwrap_or(batch_size);

        // Create one-hot encodings
        let mut quant_onehot = [0.0f32; 8];
        if let Some(qt) = self.quant_type {
            quant_onehot[qt.to_index()] = 1.0;
        }

        let mut kernel_onehot = [0.0f32; 16];
        if let Some(kt) = self.kernel_type {
            kernel_onehot[kt.to_index()] = 1.0;
        }

        // Calculate derived features
        // C-14 (Meyer DbC): 0 = unknown, no architecture-specific magic number.
        let hidden_dim = self.hidden_dim.unwrap_or(0) as f32;
        let batch_size_f = batch_size as f32;
        let quant_bytes = self.quant_type.map(|q| q.bytes_per_param()).unwrap_or(0.5625);

        // Arithmetic intensity for GEMV: 2*N*K FLOPs / (N*K*bytes + K + N) bytes
        // Simplified: ~2 / bytes_per_param for memory-bound inference
        let arithmetic_intensity = (2.0 / quant_bytes).min(10.0) / 10.0;

        // Theoretical efficiency starts at 0 (unknown until measured)
        let theoretical_efficiency = 0.0;

        Ok(TunerFeatures {
            // Normalized static features
            model_params_b: self
                .model_params_b
                .map(|p| (p.max(f32::EPSILON).log10() + 1.0) / 3.0) // log10(0.1)=-1, log10(100)=2 → [0, 1]
                .unwrap_or(0.0)
                .clamp(0.0, 1.0),
            hidden_dim_norm: (hidden_dim / 16384.0).clamp(0.0, 1.0),
            // N-07 (Meyer DbC): 0 when unknown — no model-specific defaults.
            num_layers_norm: (self.num_layers.unwrap_or(0) as f32 / 128.0).clamp(0.0, 1.0),
            num_heads_norm: (self.num_heads.unwrap_or(0) as f32 / 128.0).clamp(0.0, 1.0),
            head_dim_norm: (self.head_dim.unwrap_or(0) as f32 / 256.0).clamp(0.0, 1.0),
            vocab_size_log: self
                .vocab_size
                .map(|v| (v.max(1) as f32).log10() / 6.0) // log10(1M)=6
                .unwrap_or(0.0)
                .clamp(0.0, 1.0),
            batch_size_norm: (batch_size_f / 64.0).clamp(0.0, 1.0),
            seq_len_log: self
                .seq_len
                .map(|s| (s.max(1) as f32).log2() / 15.0) // log2(32K)≈15
                .unwrap_or(0.0)
                .clamp(0.0, 1.0),
            cuda_graphs: if self.cuda_graphs { 1.0 } else { 0.0 },
            kv_cache_ratio: (kv_caches as f32 / batch_size_f).clamp(0.0, 1.0),
            is_prefill: if self.is_prefill { 1.0 } else { 0.0 },

            // One-hot encodings
            quant_type_onehot: quant_onehot,
            kernel_type_onehot: kernel_onehot,

            // Hardware features (5) [v1.1.0]
            gpu_mem_bw_norm: (self.gpu_mem_bw_gbs.unwrap_or(1000.0) / 3000.0).clamp(0.0, 1.0),
            gpu_compute_norm: (self.gpu_compute_tflops.unwrap_or(100.0) / 500.0).clamp(0.0, 1.0),
            gpu_sm_norm: (self.gpu_sm_count.unwrap_or(128) as f32 / 200.0).clamp(0.0, 1.0),
            gpu_l2_cache_norm: (self.gpu_l2_cache_mb.unwrap_or(48.0) / 128.0).clamp(0.0, 1.0), // v1.1.0
            is_zero_copy: if self.is_zero_copy { 1.0 } else { 0.0 }, // v1.1.0

            // Derived features
            arithmetic_intensity,
            theoretical_efficiency,

            // Labels
            measured_tps: self.measured_tps,
            best_kernel_id: None,
            bottleneck_class: None,
        })
    }
}