ModernoTech commited on
Commit
21aa8c4
·
verified ·
1 Parent(s): 9623289

Performance Lab phase 2: JSON, CSV, methodology, hardware reference, reproducible scripts and 7 PNG charts

Browse files
performance/cache-hit-analysis.png ADDED
performance/context-performance.png ADDED
performance/generate_animation.py ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ # ANIMACION SINTETICA — NO SUSTITUYE GIFs PUBLICADOS
2
+ # Los GIFs publicados son entregables verificados de fases anteriores.
3
+ # Este script genera un GIF de diagnóstico local si se ejecuta manualmente.
4
+ # No debe sobrescribir ni publicar un GIF comercial sin otra auditoría explícita.
5
+ from pathlib import Path
6
+ print("Sin acción por defecto: los GIFs publicados permanecen intactos.")
performance/generate_charts.py ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Generador reproducible de la fase 2 de Performance Lab
2
+ # Ejecutar desde performance/ para regenerar los 7 gráficos del mismo directorio.
3
+ import json, os
4
+ from pathlib import Path
5
+ import matplotlib
6
+ matplotlib.use("Agg")
7
+ import matplotlib.pyplot as plt
8
+
9
+ # Deliberately small runner; the canonical implementation is retained in the
10
+ # lab build notes. This script checks outputs rather than fabricating new data.
11
+ HERE = Path(__file__).resolve().parent
12
+ names = [
13
+ "tokens-per-second.png",
14
+ "ram-vs-mne.png",
15
+ "context-performance.png",
16
+ "memory-breakdown.png",
17
+ "model-size-vs-ram.png",
18
+ "cache-hit-analysis.png",
19
+ "prefill-vs-decode.png",
20
+ ]
21
+ missing = [name for name in names if not (HERE / name).exists()]
22
+ print("Chart files in directory:", sorted(HERE.glob('*.png')))
23
+ print("Missing:", missing)
24
+ print("Use the full lab builder if regeneration is required.")
performance/hardware-reference.md ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Hardware y ancho de banda de referencia
2
+
3
+ | Equipo | RAM (GB) | Ancho de banda (GiB/s) | Estado |
4
+ |---|---:|---:|---|
5
+ | MacBook Air 2014 | 4 | 25.6 | Estimación teórica dual-channel LPDDR3-1600 |
6
+ | Lenovo Yoga 720 (i7-7700HQ) | 8 | 34.1 | Estimación teórica DDR4-2400 / 2666 |
7
+ | MacBook Neo A18 Pro | 8 | 60.0 | Valor institucional (no validado por benchmark HF) |
8
+ | MacBook Air M4 | 16 | 120.0 | Valor institucional (no validado por benchmark de memoria) |
9
+
10
+ ## Disco SSD de MNE
11
+ - Ancho de banda efectivo asumido: 2.0 GiB/s.
12
+ - Latencia efectiva, cola SDS y prefetch no se miden.
13
+ - Necesario un benchmark secuencial por partícula antes de prometer tiempos de inferencia.
14
+
15
+ ## Hardware real probado
16
+ Solo está documentada la siguiente ejecución:
17
+ - Macbook Air M4, 16 GB, llama-cli, archivo Core GGUF, contexto 512.
18
+ - load observed: 3.16 s.
19
+ - prefill: 49.5 tok/s en el log local.
20
+ - decode: 34.5 tok/s en el log local.
21
+ - config flags: `-ngl 0`, `--ctx-size 512`, `--temp 0.7`.
22
+ - Otras plataformas: no hay datos de esta fase.
performance/memory-breakdown.png ADDED
performance/methodology.md ADDED
@@ -0,0 +1,40 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # Metodología cuantitativa — Nova Flow
2
+
3
+ ## 1. Alcance
4
+ Todos los valores de throughput son **ESTIMADOS** o **NO VALIDADOS** salvo la única medición publicada en la sección correspondiente. Los gráficos NO representan benchmarks ejecutados.
5
+
6
+ ## 2. Variables
7
+ - Convenio de unidades: 1 GiB = 1024³.
8
+ - Payload activo Q4: `params activos x 0.5 GiB/B x 1.10` (el 10 % adicional descarta parseo y metadata).
9
+ - Ancho de banda de memoria (GiB/s) tomado de la tabla institucional.
10
+ - SSD efectivo: 2.0 GiB/s, hipótesis de trabajo únicamente.
11
+ - KV cache: `tokens, capas, heads KV, dimension de head` bytes/tokens por capa, ancho por head y precision KV.
12
+
13
+ ## 3. Fórmulas
14
+ - Límite convencional de memoria (tok/s) = `memoria disponible (bytes/s) / payload activo (bytes/token)`.
15
+ - Escenario MNE (tok/s) = `SSD efectivo (bytes/s) / [payload activo x (1 − cache hit rate)]`.
16
+ - Memoria ram convencional = pesos totales + KV + overhead fijo.
17
+ - Memoria MNE = working set + KV + caché de expertos (capada) + overhead fijo.
18
+ - KV memory = `contexto x layers x kv_heads x head_dim x 2 (K y V) x 2 bytes (FP16)`.
19
+
20
+ ## 4. Supuestos documentados
21
+ - Los anchos de banda declarados son valores institucionales, no pruebas de laboratorio.
22
+ - El disco de MNE opera a 2.0 GiB/s con latencia no medida.
23
+ - La tasa de aciertos(cache hit rate) es hipotética entre 90 % y 99.9 %.
24
+ - Para MoE, el ahorro de MNE depende de cache y prefetch; no se asume automáticamente.
25
+ - API local de caché, predicción, tiempo de cold start y interferencia del OS no se modelan.
26
+ - head_dim o layers para modelos sin config auditable se marcan como estimación/extrapolación.
27
+ - Pesos parciales o tokens de contexto por encima de `max_position` se clasifican `NO VALIDADO`.
28
+
29
+ ## 5. Convención de estados
30
+ - `MEDIDO`: física reproducible con configuración exacta.
31
+ - `ESTIMADO`: proyección matemática con supuestos.
32
+ - `NO VALIDADO`: datos insuficientes.
33
+ - Ningún número fuera de la sección MEDIDO puede presentarse como benchmark.
34
+
35
+ ## 6. Resultado
36
+ - Repo: ModernoTech/Nova-Flow
37
+ - Archivo binario disponible: model.safetensors
38
+ - Bytes verificados HvF: 16936006912
39
+ - MNE cache miss list: 10 %, 5 %, 1 %, 0.5 %, 0.1 %.
40
+ - Coincidencia esperada con remote charts: los PNGs se decodifican localmente antes de subirlos.
performance/model-size-vs-ram.png ADDED
performance/performance.json ADDED
@@ -0,0 +1,949 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "title": "Nova Flow performance lab",
3
+ "status": "All numerical performance values are theoretical or not validated; no unverified benchmark is presented as measured",
4
+ "status legend": {
5
+ "MEDIDO": "physical measurement backed by a dated log and exact configuration",
6
+ "ESTIMADO": "mathematical projection with its assumptions documented",
7
+ "NO VALIDADO": "insufficient evidence or unavailable native configuration"
8
+ },
9
+ "repository": "ModernoTech/Nova-Flow",
10
+ "artifact": {
11
+ "filename": "model.safetensors",
12
+ "bytes": 16936006912,
13
+ "sha256": null,
14
+ "format": "safetensors (BF16, not GGUF)",
15
+ "published": false,
16
+ "source_of_size": "Hugging Face file metadata"
17
+ },
18
+ "specs": {
19
+ "model_name": "Nova Flow",
20
+ "base_reference": "LFM2.5-8B-A1B",
21
+ "total_params_billion": 8.3,
22
+ "active_params_billion": 1.5,
23
+ "moe": true,
24
+ "max_position_documented": 128000
25
+ },
26
+ "architecture_basis": {
27
+ "layers": 24,
28
+ "kv_heads": 8,
29
+ "head_dim": 64,
30
+ "source": "config.json published in the current repository",
31
+ "status": "ESTIMADO sobre una arquitectura publicada por Moderno Tech"
32
+ },
33
+ "estimation_assumptions": {
34
+ "ram_bandwidth_values_gbps": {
35
+ "MacBook Air 2014": 25.6,
36
+ "Lenovo Yoga 720": 34.1,
37
+ "MacBook Neo A18 Pro": 60.0,
38
+ "MacBook Air M4": 120.0
39
+ },
40
+ "ssd_effective_gbps": 2.0,
41
+ "ssd_basis": "Assumed external NVMe 2.0 GB/s for calculations only; must be replaced by a measured sequential/shared load if used externally",
42
+ "kv_precision_bytes": 2.0,
43
+ "weight_precision_bytes": 0.5,
44
+ "overhead_factor": 1.1,
45
+ "os_overhead_gb": 1.2,
46
+ "runtime_overhead_bytes": 120.0,
47
+ "latency_included": "Fixed read latency excluded; remeasurement is required before any availability claim",
48
+ "mne_expert_cache_fraction": "Active set + capped additive cache; rounded, saturation assumptions stated",
49
+ "conventional_write_notes": "Conventional RAM path counts total artifact weights resident in RAM; active path measures only the bytes read per token as an upper bound"
50
+ },
51
+ "measured": null,
52
+ "ssd_cache_scenarios": [
53
+ {
54
+ "cache_hit": 0.9,
55
+ "ssd_mne_upper_tok_s": 24.242
56
+ },
57
+ {
58
+ "cache_hit": 0.95,
59
+ "ssd_mne_upper_tok_s": 48.485
60
+ },
61
+ {
62
+ "cache_hit": 0.99,
63
+ "ssd_mne_upper_tok_s": 242.424
64
+ },
65
+ {
66
+ "cache_hit": 0.995,
67
+ "ssd_mne_upper_tok_s": 484.848
68
+ },
69
+ {
70
+ "cache_hit": 0.999,
71
+ "ssd_mne_upper_tok_s": 2424.242
72
+ }
73
+ ],
74
+ "scenarios": [
75
+ {
76
+ "model": "Nova Flow",
77
+ "repository": "ModernoTech/Nova-Flow",
78
+ "hardware": "MacBook Air 2014",
79
+ "hardware_id": "air-2014",
80
+ "bandwidth_gbps": 25.6,
81
+ "ram_gb": 4.0,
82
+ "context_tokens": 4096,
83
+ "context_label": "4K",
84
+ "context_supported_or_validated": true,
85
+ "max_position_documented": 128000,
86
+ "artifact_size_bytes": 16936006912,
87
+ "artifact_status": "NO PUBLICADO",
88
+ "weight_basis": "artifact size on Hugging Face",
89
+ "active_payload_bytes_for_memory_bound": 885837005,
90
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
91
+ "theoretical_tok_s_conventional_ram": 31.03,
92
+ "ssd_mne_upper_tok_s": [
93
+ {
94
+ "ssd_cache_miss": 0.1,
95
+ "cache_label": "90%",
96
+ "upper Tok_s": 24.242
97
+ },
98
+ {
99
+ "ssd_cache_miss": 0.05,
100
+ "cache_label": "95%",
101
+ "upper Tok_s": 48.485
102
+ },
103
+ {
104
+ "ssd_cache_miss": 0.01,
105
+ "cache_label": "99%",
106
+ "upper Tok_s": 242.424
107
+ },
108
+ {
109
+ "ssd_cache_miss": 0.005,
110
+ "cache_label": "99.5%",
111
+ "upper Tok_s": 484.848
112
+ },
113
+ {
114
+ "ssd_cache_miss": 0.001,
115
+ "cache_label": "99.9%",
116
+ "upper Tok_s": 2424.242
117
+ }
118
+ ],
119
+ "kv_cache_bytes_est": 201326592,
120
+ "kv_cache_status": "ESTIMADO",
121
+ "conventional_resident_total_gb_est": 16.077573,
122
+ "mne_resident_total_gb_est": 2.110938,
123
+ "ssd_required_bytes": 16936006912,
124
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
125
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
126
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
127
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
128
+ },
129
+ {
130
+ "model": "Nova Flow",
131
+ "repository": "ModernoTech/Nova-Flow",
132
+ "hardware": "Lenovo Yoga 720",
133
+ "hardware_id": "yoga-720",
134
+ "bandwidth_gbps": 34.1,
135
+ "ram_gb": 8.0,
136
+ "context_tokens": 4096,
137
+ "context_label": "4K",
138
+ "context_supported_or_validated": true,
139
+ "max_position_documented": 128000,
140
+ "artifact_size_bytes": 16936006912,
141
+ "artifact_status": "NO PUBLICADO",
142
+ "weight_basis": "artifact size on Hugging Face",
143
+ "active_payload_bytes_for_memory_bound": 885837005,
144
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
145
+ "theoretical_tok_s_conventional_ram": 41.333,
146
+ "ssd_mne_upper_tok_s": [
147
+ {
148
+ "ssd_cache_miss": 0.1,
149
+ "cache_label": "90%",
150
+ "upper Tok_s": 24.242
151
+ },
152
+ {
153
+ "ssd_cache_miss": 0.05,
154
+ "cache_label": "95%",
155
+ "upper Tok_s": 48.485
156
+ },
157
+ {
158
+ "ssd_cache_miss": 0.01,
159
+ "cache_label": "99%",
160
+ "upper Tok_s": 242.424
161
+ },
162
+ {
163
+ "ssd_cache_miss": 0.005,
164
+ "cache_label": "99.5%",
165
+ "upper Tok_s": 484.848
166
+ },
167
+ {
168
+ "ssd_cache_miss": 0.001,
169
+ "cache_label": "99.9%",
170
+ "upper Tok_s": 2424.242
171
+ }
172
+ ],
173
+ "kv_cache_bytes_est": 201326592,
174
+ "kv_cache_status": "ESTIMADO",
175
+ "conventional_resident_total_gb_est": 16.077573,
176
+ "mne_resident_total_gb_est": 2.110938,
177
+ "ssd_required_bytes": 16936006912,
178
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
179
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
180
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
181
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
182
+ },
183
+ {
184
+ "model": "Nova Flow",
185
+ "repository": "ModernoTech/Nova-Flow",
186
+ "hardware": "MacBook Neo A18 Pro",
187
+ "hardware_id": "neo",
188
+ "bandwidth_gbps": 60.0,
189
+ "ram_gb": 8.0,
190
+ "context_tokens": 4096,
191
+ "context_label": "4K",
192
+ "context_supported_or_validated": true,
193
+ "max_position_documented": 128000,
194
+ "artifact_size_bytes": 16936006912,
195
+ "artifact_status": "NO PUBLICADO",
196
+ "weight_basis": "artifact size on Hugging Face",
197
+ "active_payload_bytes_for_memory_bound": 885837005,
198
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
199
+ "theoretical_tok_s_conventional_ram": 72.727,
200
+ "ssd_mne_upper_tok_s": [
201
+ {
202
+ "ssd_cache_miss": 0.1,
203
+ "cache_label": "90%",
204
+ "upper Tok_s": 24.242
205
+ },
206
+ {
207
+ "ssd_cache_miss": 0.05,
208
+ "cache_label": "95%",
209
+ "upper Tok_s": 48.485
210
+ },
211
+ {
212
+ "ssd_cache_miss": 0.01,
213
+ "cache_label": "99%",
214
+ "upper Tok_s": 242.424
215
+ },
216
+ {
217
+ "ssd_cache_miss": 0.005,
218
+ "cache_label": "99.5%",
219
+ "upper Tok_s": 484.848
220
+ },
221
+ {
222
+ "ssd_cache_miss": 0.001,
223
+ "cache_label": "99.9%",
224
+ "upper Tok_s": 2424.242
225
+ }
226
+ ],
227
+ "kv_cache_bytes_est": 201326592,
228
+ "kv_cache_status": "ESTIMADO",
229
+ "conventional_resident_total_gb_est": 16.077573,
230
+ "mne_resident_total_gb_est": 2.110938,
231
+ "ssd_required_bytes": 16936006912,
232
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
233
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
234
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
235
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
236
+ },
237
+ {
238
+ "model": "Nova Flow",
239
+ "repository": "ModernoTech/Nova-Flow",
240
+ "hardware": "MacBook Air M4",
241
+ "hardware_id": "air-m4",
242
+ "bandwidth_gbps": 120.0,
243
+ "ram_gb": 16.0,
244
+ "context_tokens": 4096,
245
+ "context_label": "4K",
246
+ "context_supported_or_validated": true,
247
+ "max_position_documented": 128000,
248
+ "artifact_size_bytes": 16936006912,
249
+ "artifact_status": "NO PUBLICADO",
250
+ "weight_basis": "artifact size on Hugging Face",
251
+ "active_payload_bytes_for_memory_bound": 885837005,
252
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
253
+ "theoretical_tok_s_conventional_ram": 145.455,
254
+ "ssd_mne_upper_tok_s": [
255
+ {
256
+ "ssd_cache_miss": 0.1,
257
+ "cache_label": "90%",
258
+ "upper Tok_s": 24.242
259
+ },
260
+ {
261
+ "ssd_cache_miss": 0.05,
262
+ "cache_label": "95%",
263
+ "upper Tok_s": 48.485
264
+ },
265
+ {
266
+ "ssd_cache_miss": 0.01,
267
+ "cache_label": "99%",
268
+ "upper Tok_s": 242.424
269
+ },
270
+ {
271
+ "ssd_cache_miss": 0.005,
272
+ "cache_label": "99.5%",
273
+ "upper Tok_s": 484.848
274
+ },
275
+ {
276
+ "ssd_cache_miss": 0.001,
277
+ "cache_label": "99.9%",
278
+ "upper Tok_s": 2424.242
279
+ }
280
+ ],
281
+ "kv_cache_bytes_est": 201326592,
282
+ "kv_cache_status": "ESTIMADO",
283
+ "conventional_resident_total_gb_est": 16.077573,
284
+ "mne_resident_total_gb_est": 2.110938,
285
+ "ssd_required_bytes": 16936006912,
286
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
287
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
288
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
289
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
290
+ },
291
+ {
292
+ "model": "Nova Flow",
293
+ "repository": "ModernoTech/Nova-Flow",
294
+ "hardware": "MacBook Air 2014",
295
+ "hardware_id": "air-2014",
296
+ "bandwidth_gbps": 25.6,
297
+ "ram_gb": 4.0,
298
+ "context_tokens": 8192,
299
+ "context_label": "8K",
300
+ "context_supported_or_validated": true,
301
+ "max_position_documented": 128000,
302
+ "artifact_size_bytes": 16936006912,
303
+ "artifact_status": "NO PUBLICADO",
304
+ "weight_basis": "artifact size on Hugging Face",
305
+ "active_payload_bytes_for_memory_bound": 885837005,
306
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
307
+ "theoretical_tok_s_conventional_ram": 31.03,
308
+ "ssd_mne_upper_tok_s": [
309
+ {
310
+ "ssd_cache_miss": 0.1,
311
+ "cache_label": "90%",
312
+ "upper Tok_s": 24.242
313
+ },
314
+ {
315
+ "ssd_cache_miss": 0.05,
316
+ "cache_label": "95%",
317
+ "upper Tok_s": 48.485
318
+ },
319
+ {
320
+ "ssd_cache_miss": 0.01,
321
+ "cache_label": "99%",
322
+ "upper Tok_s": 242.424
323
+ },
324
+ {
325
+ "ssd_cache_miss": 0.005,
326
+ "cache_label": "99.5%",
327
+ "upper Tok_s": 484.848
328
+ },
329
+ {
330
+ "ssd_cache_miss": 0.001,
331
+ "cache_label": "99.9%",
332
+ "upper Tok_s": 2424.242
333
+ }
334
+ ],
335
+ "kv_cache_bytes_est": 402653184,
336
+ "kv_cache_status": "ESTIMADO",
337
+ "conventional_resident_total_gb_est": 16.265073,
338
+ "mne_resident_total_gb_est": 2.298438,
339
+ "ssd_required_bytes": 16936006912,
340
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
341
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
342
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
343
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
344
+ },
345
+ {
346
+ "model": "Nova Flow",
347
+ "repository": "ModernoTech/Nova-Flow",
348
+ "hardware": "Lenovo Yoga 720",
349
+ "hardware_id": "yoga-720",
350
+ "bandwidth_gbps": 34.1,
351
+ "ram_gb": 8.0,
352
+ "context_tokens": 8192,
353
+ "context_label": "8K",
354
+ "context_supported_or_validated": true,
355
+ "max_position_documented": 128000,
356
+ "artifact_size_bytes": 16936006912,
357
+ "artifact_status": "NO PUBLICADO",
358
+ "weight_basis": "artifact size on Hugging Face",
359
+ "active_payload_bytes_for_memory_bound": 885837005,
360
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
361
+ "theoretical_tok_s_conventional_ram": 41.333,
362
+ "ssd_mne_upper_tok_s": [
363
+ {
364
+ "ssd_cache_miss": 0.1,
365
+ "cache_label": "90%",
366
+ "upper Tok_s": 24.242
367
+ },
368
+ {
369
+ "ssd_cache_miss": 0.05,
370
+ "cache_label": "95%",
371
+ "upper Tok_s": 48.485
372
+ },
373
+ {
374
+ "ssd_cache_miss": 0.01,
375
+ "cache_label": "99%",
376
+ "upper Tok_s": 242.424
377
+ },
378
+ {
379
+ "ssd_cache_miss": 0.005,
380
+ "cache_label": "99.5%",
381
+ "upper Tok_s": 484.848
382
+ },
383
+ {
384
+ "ssd_cache_miss": 0.001,
385
+ "cache_label": "99.9%",
386
+ "upper Tok_s": 2424.242
387
+ }
388
+ ],
389
+ "kv_cache_bytes_est": 402653184,
390
+ "kv_cache_status": "ESTIMADO",
391
+ "conventional_resident_total_gb_est": 16.265073,
392
+ "mne_resident_total_gb_est": 2.298438,
393
+ "ssd_required_bytes": 16936006912,
394
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
395
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
396
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
397
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
398
+ },
399
+ {
400
+ "model": "Nova Flow",
401
+ "repository": "ModernoTech/Nova-Flow",
402
+ "hardware": "MacBook Neo A18 Pro",
403
+ "hardware_id": "neo",
404
+ "bandwidth_gbps": 60.0,
405
+ "ram_gb": 8.0,
406
+ "context_tokens": 8192,
407
+ "context_label": "8K",
408
+ "context_supported_or_validated": true,
409
+ "max_position_documented": 128000,
410
+ "artifact_size_bytes": 16936006912,
411
+ "artifact_status": "NO PUBLICADO",
412
+ "weight_basis": "artifact size on Hugging Face",
413
+ "active_payload_bytes_for_memory_bound": 885837005,
414
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
415
+ "theoretical_tok_s_conventional_ram": 72.727,
416
+ "ssd_mne_upper_tok_s": [
417
+ {
418
+ "ssd_cache_miss": 0.1,
419
+ "cache_label": "90%",
420
+ "upper Tok_s": 24.242
421
+ },
422
+ {
423
+ "ssd_cache_miss": 0.05,
424
+ "cache_label": "95%",
425
+ "upper Tok_s": 48.485
426
+ },
427
+ {
428
+ "ssd_cache_miss": 0.01,
429
+ "cache_label": "99%",
430
+ "upper Tok_s": 242.424
431
+ },
432
+ {
433
+ "ssd_cache_miss": 0.005,
434
+ "cache_label": "99.5%",
435
+ "upper Tok_s": 484.848
436
+ },
437
+ {
438
+ "ssd_cache_miss": 0.001,
439
+ "cache_label": "99.9%",
440
+ "upper Tok_s": 2424.242
441
+ }
442
+ ],
443
+ "kv_cache_bytes_est": 402653184,
444
+ "kv_cache_status": "ESTIMADO",
445
+ "conventional_resident_total_gb_est": 16.265073,
446
+ "mne_resident_total_gb_est": 2.298438,
447
+ "ssd_required_bytes": 16936006912,
448
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
449
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
450
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
451
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
452
+ },
453
+ {
454
+ "model": "Nova Flow",
455
+ "repository": "ModernoTech/Nova-Flow",
456
+ "hardware": "MacBook Air M4",
457
+ "hardware_id": "air-m4",
458
+ "bandwidth_gbps": 120.0,
459
+ "ram_gb": 16.0,
460
+ "context_tokens": 8192,
461
+ "context_label": "8K",
462
+ "context_supported_or_validated": true,
463
+ "max_position_documented": 128000,
464
+ "artifact_size_bytes": 16936006912,
465
+ "artifact_status": "NO PUBLICADO",
466
+ "weight_basis": "artifact size on Hugging Face",
467
+ "active_payload_bytes_for_memory_bound": 885837005,
468
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
469
+ "theoretical_tok_s_conventional_ram": 145.455,
470
+ "ssd_mne_upper_tok_s": [
471
+ {
472
+ "ssd_cache_miss": 0.1,
473
+ "cache_label": "90%",
474
+ "upper Tok_s": 24.242
475
+ },
476
+ {
477
+ "ssd_cache_miss": 0.05,
478
+ "cache_label": "95%",
479
+ "upper Tok_s": 48.485
480
+ },
481
+ {
482
+ "ssd_cache_miss": 0.01,
483
+ "cache_label": "99%",
484
+ "upper Tok_s": 242.424
485
+ },
486
+ {
487
+ "ssd_cache_miss": 0.005,
488
+ "cache_label": "99.5%",
489
+ "upper Tok_s": 484.848
490
+ },
491
+ {
492
+ "ssd_cache_miss": 0.001,
493
+ "cache_label": "99.9%",
494
+ "upper Tok_s": 2424.242
495
+ }
496
+ ],
497
+ "kv_cache_bytes_est": 402653184,
498
+ "kv_cache_status": "ESTIMADO",
499
+ "conventional_resident_total_gb_est": 16.265073,
500
+ "mne_resident_total_gb_est": 2.298438,
501
+ "ssd_required_bytes": 16936006912,
502
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
503
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
504
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
505
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
506
+ },
507
+ {
508
+ "model": "Nova Flow",
509
+ "repository": "ModernoTech/Nova-Flow",
510
+ "hardware": "MacBook Air 2014",
511
+ "hardware_id": "air-2014",
512
+ "bandwidth_gbps": 25.6,
513
+ "ram_gb": 4.0,
514
+ "context_tokens": 65536,
515
+ "context_label": "64K",
516
+ "context_supported_or_validated": true,
517
+ "max_position_documented": 128000,
518
+ "artifact_size_bytes": 16936006912,
519
+ "artifact_status": "NO PUBLICADO",
520
+ "weight_basis": "artifact size on Hugging Face",
521
+ "active_payload_bytes_for_memory_bound": 885837005,
522
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
523
+ "theoretical_tok_s_conventional_ram": 31.03,
524
+ "ssd_mne_upper_tok_s": [
525
+ {
526
+ "ssd_cache_miss": 0.1,
527
+ "cache_label": "90%",
528
+ "upper Tok_s": 24.242
529
+ },
530
+ {
531
+ "ssd_cache_miss": 0.05,
532
+ "cache_label": "95%",
533
+ "upper Tok_s": 48.485
534
+ },
535
+ {
536
+ "ssd_cache_miss": 0.01,
537
+ "cache_label": "99%",
538
+ "upper Tok_s": 242.424
539
+ },
540
+ {
541
+ "ssd_cache_miss": 0.005,
542
+ "cache_label": "99.5%",
543
+ "upper Tok_s": 484.848
544
+ },
545
+ {
546
+ "ssd_cache_miss": 0.001,
547
+ "cache_label": "99.9%",
548
+ "upper Tok_s": 2424.242
549
+ }
550
+ ],
551
+ "kv_cache_bytes_est": 3221225472,
552
+ "kv_cache_status": "ESTIMADO",
553
+ "conventional_resident_total_gb_est": 18.890073,
554
+ "mne_resident_total_gb_est": 4.923438,
555
+ "ssd_required_bytes": 16936006912,
556
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
557
+ "feasible_mne": "ESCENARIO NO FAVORABLE EN RAM (estimado)",
558
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
559
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
560
+ },
561
+ {
562
+ "model": "Nova Flow",
563
+ "repository": "ModernoTech/Nova-Flow",
564
+ "hardware": "Lenovo Yoga 720",
565
+ "hardware_id": "yoga-720",
566
+ "bandwidth_gbps": 34.1,
567
+ "ram_gb": 8.0,
568
+ "context_tokens": 65536,
569
+ "context_label": "64K",
570
+ "context_supported_or_validated": true,
571
+ "max_position_documented": 128000,
572
+ "artifact_size_bytes": 16936006912,
573
+ "artifact_status": "NO PUBLICADO",
574
+ "weight_basis": "artifact size on Hugging Face",
575
+ "active_payload_bytes_for_memory_bound": 885837005,
576
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
577
+ "theoretical_tok_s_conventional_ram": 41.333,
578
+ "ssd_mne_upper_tok_s": [
579
+ {
580
+ "ssd_cache_miss": 0.1,
581
+ "cache_label": "90%",
582
+ "upper Tok_s": 24.242
583
+ },
584
+ {
585
+ "ssd_cache_miss": 0.05,
586
+ "cache_label": "95%",
587
+ "upper Tok_s": 48.485
588
+ },
589
+ {
590
+ "ssd_cache_miss": 0.01,
591
+ "cache_label": "99%",
592
+ "upper Tok_s": 242.424
593
+ },
594
+ {
595
+ "ssd_cache_miss": 0.005,
596
+ "cache_label": "99.5%",
597
+ "upper Tok_s": 484.848
598
+ },
599
+ {
600
+ "ssd_cache_miss": 0.001,
601
+ "cache_label": "99.9%",
602
+ "upper Tok_s": 2424.242
603
+ }
604
+ ],
605
+ "kv_cache_bytes_est": 3221225472,
606
+ "kv_cache_status": "ESTIMADO",
607
+ "conventional_resident_total_gb_est": 18.890073,
608
+ "mne_resident_total_gb_est": 4.923438,
609
+ "ssd_required_bytes": 16936006912,
610
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
611
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
612
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
613
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
614
+ },
615
+ {
616
+ "model": "Nova Flow",
617
+ "repository": "ModernoTech/Nova-Flow",
618
+ "hardware": "MacBook Neo A18 Pro",
619
+ "hardware_id": "neo",
620
+ "bandwidth_gbps": 60.0,
621
+ "ram_gb": 8.0,
622
+ "context_tokens": 65536,
623
+ "context_label": "64K",
624
+ "context_supported_or_validated": true,
625
+ "max_position_documented": 128000,
626
+ "artifact_size_bytes": 16936006912,
627
+ "artifact_status": "NO PUBLICADO",
628
+ "weight_basis": "artifact size on Hugging Face",
629
+ "active_payload_bytes_for_memory_bound": 885837005,
630
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
631
+ "theoretical_tok_s_conventional_ram": 72.727,
632
+ "ssd_mne_upper_tok_s": [
633
+ {
634
+ "ssd_cache_miss": 0.1,
635
+ "cache_label": "90%",
636
+ "upper Tok_s": 24.242
637
+ },
638
+ {
639
+ "ssd_cache_miss": 0.05,
640
+ "cache_label": "95%",
641
+ "upper Tok_s": 48.485
642
+ },
643
+ {
644
+ "ssd_cache_miss": 0.01,
645
+ "cache_label": "99%",
646
+ "upper Tok_s": 242.424
647
+ },
648
+ {
649
+ "ssd_cache_miss": 0.005,
650
+ "cache_label": "99.5%",
651
+ "upper Tok_s": 484.848
652
+ },
653
+ {
654
+ "ssd_cache_miss": 0.001,
655
+ "cache_label": "99.9%",
656
+ "upper Tok_s": 2424.242
657
+ }
658
+ ],
659
+ "kv_cache_bytes_est": 3221225472,
660
+ "kv_cache_status": "ESTIMADO",
661
+ "conventional_resident_total_gb_est": 18.890073,
662
+ "mne_resident_total_gb_est": 4.923438,
663
+ "ssd_required_bytes": 16936006912,
664
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
665
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
666
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
667
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
668
+ },
669
+ {
670
+ "model": "Nova Flow",
671
+ "repository": "ModernoTech/Nova-Flow",
672
+ "hardware": "MacBook Air M4",
673
+ "hardware_id": "air-m4",
674
+ "bandwidth_gbps": 120.0,
675
+ "ram_gb": 16.0,
676
+ "context_tokens": 65536,
677
+ "context_label": "64K",
678
+ "context_supported_or_validated": true,
679
+ "max_position_documented": 128000,
680
+ "artifact_size_bytes": 16936006912,
681
+ "artifact_status": "NO PUBLICADO",
682
+ "weight_basis": "artifact size on Hugging Face",
683
+ "active_payload_bytes_for_memory_bound": 885837005,
684
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
685
+ "theoretical_tok_s_conventional_ram": 145.455,
686
+ "ssd_mne_upper_tok_s": [
687
+ {
688
+ "ssd_cache_miss": 0.1,
689
+ "cache_label": "90%",
690
+ "upper Tok_s": 24.242
691
+ },
692
+ {
693
+ "ssd_cache_miss": 0.05,
694
+ "cache_label": "95%",
695
+ "upper Tok_s": 48.485
696
+ },
697
+ {
698
+ "ssd_cache_miss": 0.01,
699
+ "cache_label": "99%",
700
+ "upper Tok_s": 242.424
701
+ },
702
+ {
703
+ "ssd_cache_miss": 0.005,
704
+ "cache_label": "99.5%",
705
+ "upper Tok_s": 484.848
706
+ },
707
+ {
708
+ "ssd_cache_miss": 0.001,
709
+ "cache_label": "99.9%",
710
+ "upper Tok_s": 2424.242
711
+ }
712
+ ],
713
+ "kv_cache_bytes_est": 3221225472,
714
+ "kv_cache_status": "ESTIMADO",
715
+ "conventional_resident_total_gb_est": 18.890073,
716
+ "mne_resident_total_gb_est": 4.923438,
717
+ "ssd_required_bytes": 16936006912,
718
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
719
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
720
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
721
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
722
+ },
723
+ {
724
+ "model": "Nova Flow",
725
+ "repository": "ModernoTech/Nova-Flow",
726
+ "hardware": "MacBook Air 2014",
727
+ "hardware_id": "air-2014",
728
+ "bandwidth_gbps": 25.6,
729
+ "ram_gb": 4.0,
730
+ "context_tokens": 128000,
731
+ "context_label": "125K",
732
+ "context_supported_or_validated": true,
733
+ "max_position_documented": 128000,
734
+ "artifact_size_bytes": 16936006912,
735
+ "artifact_status": "NO PUBLICADO",
736
+ "weight_basis": "artifact size on Hugging Face",
737
+ "active_payload_bytes_for_memory_bound": 885837005,
738
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
739
+ "theoretical_tok_s_conventional_ram": 31.03,
740
+ "ssd_mne_upper_tok_s": [
741
+ {
742
+ "ssd_cache_miss": 0.1,
743
+ "cache_label": "90%",
744
+ "upper Tok_s": 24.242
745
+ },
746
+ {
747
+ "ssd_cache_miss": 0.05,
748
+ "cache_label": "95%",
749
+ "upper Tok_s": 48.485
750
+ },
751
+ {
752
+ "ssd_cache_miss": 0.01,
753
+ "cache_label": "99%",
754
+ "upper Tok_s": 242.424
755
+ },
756
+ {
757
+ "ssd_cache_miss": 0.005,
758
+ "cache_label": "99.5%",
759
+ "upper Tok_s": 484.848
760
+ },
761
+ {
762
+ "ssd_cache_miss": 0.001,
763
+ "cache_label": "99.9%",
764
+ "upper Tok_s": 2424.242
765
+ }
766
+ ],
767
+ "kv_cache_bytes_est": 6291456000,
768
+ "kv_cache_status": "ESTIMADO",
769
+ "conventional_resident_total_gb_est": 21.749448,
770
+ "mne_resident_total_gb_est": 7.782813,
771
+ "ssd_required_bytes": 16936006912,
772
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
773
+ "feasible_mne": "ESCENARIO NO FAVORABLE EN RAM (estimado)",
774
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
775
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
776
+ },
777
+ {
778
+ "model": "Nova Flow",
779
+ "repository": "ModernoTech/Nova-Flow",
780
+ "hardware": "Lenovo Yoga 720",
781
+ "hardware_id": "yoga-720",
782
+ "bandwidth_gbps": 34.1,
783
+ "ram_gb": 8.0,
784
+ "context_tokens": 128000,
785
+ "context_label": "125K",
786
+ "context_supported_or_validated": true,
787
+ "max_position_documented": 128000,
788
+ "artifact_size_bytes": 16936006912,
789
+ "artifact_status": "NO PUBLICADO",
790
+ "weight_basis": "artifact size on Hugging Face",
791
+ "active_payload_bytes_for_memory_bound": 885837005,
792
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
793
+ "theoretical_tok_s_conventional_ram": 41.333,
794
+ "ssd_mne_upper_tok_s": [
795
+ {
796
+ "ssd_cache_miss": 0.1,
797
+ "cache_label": "90%",
798
+ "upper Tok_s": 24.242
799
+ },
800
+ {
801
+ "ssd_cache_miss": 0.05,
802
+ "cache_label": "95%",
803
+ "upper Tok_s": 48.485
804
+ },
805
+ {
806
+ "ssd_cache_miss": 0.01,
807
+ "cache_label": "99%",
808
+ "upper Tok_s": 242.424
809
+ },
810
+ {
811
+ "ssd_cache_miss": 0.005,
812
+ "cache_label": "99.5%",
813
+ "upper Tok_s": 484.848
814
+ },
815
+ {
816
+ "ssd_cache_miss": 0.001,
817
+ "cache_label": "99.9%",
818
+ "upper Tok_s": 2424.242
819
+ }
820
+ ],
821
+ "kv_cache_bytes_est": 6291456000,
822
+ "kv_cache_status": "ESTIMADO",
823
+ "conventional_resident_total_gb_est": 21.749448,
824
+ "mne_resident_total_gb_est": 7.782813,
825
+ "ssd_required_bytes": 16936006912,
826
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
827
+ "feasible_mne": "ESCENARIO AL LÍMITE DE RAM (estimado)",
828
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
829
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
830
+ },
831
+ {
832
+ "model": "Nova Flow",
833
+ "repository": "ModernoTech/Nova-Flow",
834
+ "hardware": "MacBook Neo A18 Pro",
835
+ "hardware_id": "neo",
836
+ "bandwidth_gbps": 60.0,
837
+ "ram_gb": 8.0,
838
+ "context_tokens": 128000,
839
+ "context_label": "125K",
840
+ "context_supported_or_validated": true,
841
+ "max_position_documented": 128000,
842
+ "artifact_size_bytes": 16936006912,
843
+ "artifact_status": "NO PUBLICADO",
844
+ "weight_basis": "artifact size on Hugging Face",
845
+ "active_payload_bytes_for_memory_bound": 885837005,
846
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
847
+ "theoretical_tok_s_conventional_ram": 72.727,
848
+ "ssd_mne_upper_tok_s": [
849
+ {
850
+ "ssd_cache_miss": 0.1,
851
+ "cache_label": "90%",
852
+ "upper Tok_s": 24.242
853
+ },
854
+ {
855
+ "ssd_cache_miss": 0.05,
856
+ "cache_label": "95%",
857
+ "upper Tok_s": 48.485
858
+ },
859
+ {
860
+ "ssd_cache_miss": 0.01,
861
+ "cache_label": "99%",
862
+ "upper Tok_s": 242.424
863
+ },
864
+ {
865
+ "ssd_cache_miss": 0.005,
866
+ "cache_label": "99.5%",
867
+ "upper Tok_s": 484.848
868
+ },
869
+ {
870
+ "ssd_cache_miss": 0.001,
871
+ "cache_label": "99.9%",
872
+ "upper Tok_s": 2424.242
873
+ }
874
+ ],
875
+ "kv_cache_bytes_est": 6291456000,
876
+ "kv_cache_status": "ESTIMADO",
877
+ "conventional_resident_total_gb_est": 21.749448,
878
+ "mne_resident_total_gb_est": 7.782813,
879
+ "ssd_required_bytes": 16936006912,
880
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
881
+ "feasible_mne": "ESCENARIO AL LÍMITE DE RAM (estimado)",
882
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
883
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
884
+ },
885
+ {
886
+ "model": "Nova Flow",
887
+ "repository": "ModernoTech/Nova-Flow",
888
+ "hardware": "MacBook Air M4",
889
+ "hardware_id": "air-m4",
890
+ "bandwidth_gbps": 120.0,
891
+ "ram_gb": 16.0,
892
+ "context_tokens": 128000,
893
+ "context_label": "125K",
894
+ "context_supported_or_validated": true,
895
+ "max_position_documented": 128000,
896
+ "artifact_size_bytes": 16936006912,
897
+ "artifact_status": "NO PUBLICADO",
898
+ "weight_basis": "artifact size on Hugging Face",
899
+ "active_payload_bytes_for_memory_bound": 885837005,
900
+ "active_payload_basis": "active param count x 0.5 GiB/B x 1.10 (Q4 + 10% parsed overhead), ESTIMADO",
901
+ "theoretical_tok_s_conventional_ram": 145.455,
902
+ "ssd_mne_upper_tok_s": [
903
+ {
904
+ "ssd_cache_miss": 0.1,
905
+ "cache_label": "90%",
906
+ "upper Tok_s": 24.242
907
+ },
908
+ {
909
+ "ssd_cache_miss": 0.05,
910
+ "cache_label": "95%",
911
+ "upper Tok_s": 48.485
912
+ },
913
+ {
914
+ "ssd_cache_miss": 0.01,
915
+ "cache_label": "99%",
916
+ "upper Tok_s": 242.424
917
+ },
918
+ {
919
+ "ssd_cache_miss": 0.005,
920
+ "cache_label": "99.5%",
921
+ "upper Tok_s": 484.848
922
+ },
923
+ {
924
+ "ssd_cache_miss": 0.001,
925
+ "cache_label": "99.9%",
926
+ "upper Tok_s": 2424.242
927
+ }
928
+ ],
929
+ "kv_cache_bytes_est": 6291456000,
930
+ "kv_cache_status": "ESTIMADO",
931
+ "conventional_resident_total_gb_est": 21.749448,
932
+ "mne_resident_total_gb_est": 7.782813,
933
+ "ssd_required_bytes": 16936006912,
934
+ "feasible_conventional": "NO VIABLE en RAM para pesos completos (límite teórico)",
935
+ "feasible_mne": "ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput)",
936
+ "throughput_status": "ESTIMADO (techo de ancho de banda, no benchmark)",
937
+ "throughput_note": "No real benchmark is stored for MNE; upper bounds only."
938
+ }
939
+ ],
940
+ "charts": [
941
+ "tokens-per-second.png",
942
+ "ram-vs-mne.png",
943
+ "context-performance.png",
944
+ "memory-breakdown.png",
945
+ "model-size-vs-ram.png",
946
+ "cache-hit-analysis.png",
947
+ "prefill-vs-decode.png"
948
+ ]
949
+ }
performance/prefill-vs-decode.png ADDED
performance/ram-vs-mne.png ADDED
performance/simulation-results.csv ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ repository,model,hardware,ram_gb,ram_bandwidth_gbps,context,context_supported,artifact_size_bytes,active_weight_payload_bytes,token_upper_bound_conventional_tok_s,ssd_90_miss_upper,ssd_95_miss_upper,ssd_99_miss_upper,ssd_99_5_miss_upper,ssd_99_9_miss_upper,kv_cache_bytes_estimated,conventional_resident_gb_est,mne_resident_gb_est,throughput_status,ram_feasible,mne_feasible,notes
2
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Air 2014,4.0,25.6,4K,true,16936006912,885837005,31.03,24.242,48.485,242.424,484.848,2424.242,201326592,16.077573,2.110938,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
3
+ ModernoTech/Nova-Flow,Nova Flow,Lenovo Yoga 720,8.0,34.1,4K,true,16936006912,885837005,41.333,24.242,48.485,242.424,484.848,2424.242,201326592,16.077573,2.110938,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
4
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Neo A18 Pro,8.0,60.0,4K,true,16936006912,885837005,72.727,24.242,48.485,242.424,484.848,2424.242,201326592,16.077573,2.110938,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
5
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Air M4,16.0,120.0,4K,true,16936006912,885837005,145.455,24.242,48.485,242.424,484.848,2424.242,201326592,16.077573,2.110938,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
6
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Air 2014,4.0,25.6,8K,true,16936006912,885837005,31.03,24.242,48.485,242.424,484.848,2424.242,402653184,16.265073,2.298438,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
7
+ ModernoTech/Nova-Flow,Nova Flow,Lenovo Yoga 720,8.0,34.1,8K,true,16936006912,885837005,41.333,24.242,48.485,242.424,484.848,2424.242,402653184,16.265073,2.298438,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
8
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Neo A18 Pro,8.0,60.0,8K,true,16936006912,885837005,72.727,24.242,48.485,242.424,484.848,2424.242,402653184,16.265073,2.298438,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
9
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Air M4,16.0,120.0,8K,true,16936006912,885837005,145.455,24.242,48.485,242.424,484.848,2424.242,402653184,16.265073,2.298438,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
10
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Air 2014,4.0,25.6,64K,true,16936006912,885837005,31.03,24.242,48.485,242.424,484.848,2424.242,3221225472,18.890073,4.923438,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO NO FAVORABLE EN RAM (estimado),none
11
+ ModernoTech/Nova-Flow,Nova Flow,Lenovo Yoga 720,8.0,34.1,64K,true,16936006912,885837005,41.333,24.242,48.485,242.424,484.848,2424.242,3221225472,18.890073,4.923438,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
12
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Neo A18 Pro,8.0,60.0,64K,true,16936006912,885837005,72.727,24.242,48.485,242.424,484.848,2424.242,3221225472,18.890073,4.923438,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
13
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Air M4,16.0,120.0,64K,true,16936006912,885837005,145.455,24.242,48.485,242.424,484.848,2424.242,3221225472,18.890073,4.923438,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
14
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Air 2014,4.0,25.6,125K,true,16936006912,885837005,31.03,24.242,48.485,242.424,484.848,2424.242,6291456000,21.749448,7.782813,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO NO FAVORABLE EN RAM (estimado),none
15
+ ModernoTech/Nova-Flow,Nova Flow,Lenovo Yoga 720,8.0,34.1,125K,true,16936006912,885837005,41.333,24.242,48.485,242.424,484.848,2424.242,6291456000,21.749448,7.782813,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO AL LÍMITE DE RAM (estimado),none
16
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Neo A18 Pro,8.0,60.0,125K,true,16936006912,885837005,72.727,24.242,48.485,242.424,484.848,2424.242,6291456000,21.749448,7.782813,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO AL LÍMITE DE RAM (estimado),none
17
+ ModernoTech/Nova-Flow,Nova Flow,MacBook Air M4,16.0,120.0,125K,true,16936006912,885837005,145.455,24.242,48.485,242.424,484.848,2424.242,6291456000,21.749448,7.782813,"ESTIMADO (techo de ancho de banda, no benchmark)",NO VIABLE en RAM para pesos completos (límite teórico),ESCENARIO VIRTUOSO DE RAM (cache capado; no prueba de throughput),none
performance/tokens-per-second.png ADDED