Profile additional Qualcomm model components

This commit is contained in:
tzhouam committed 2026-09-24 09:31:53 +08:00
1 parent 5d81c3678b
commit 698f8c13d9
21 files changed
+35175 -41

No files matched your search

@@ -6,6 +6,8 @@ from __future__ import annotations
import argparse import argparse
import hashlib import hashlib
import json import json
import math
from collections import Counter
from pathlib import Path from pathlib import Path
import numpy as np import numpy as np
@@ -61,6 +63,7 @@ def main() -> None:
parser.add_argument("--output-report", type=Path, required=True) parser.add_argument("--output-report", type=Path, required=True)
parser.add_argument("--expected-compile-options", default=COMPILE_OPTIONS) parser.add_argument("--expected-compile-options", default=COMPILE_OPTIONS)
parser.add_argument("--expected-run-options", default="--compute_unit npu") parser.add_argument("--expected-run-options", default="--compute_unit npu")
parser.add_argument("--profile-report", type=Path)
args = parser.parse_args() args = parser.parse_args()
compiled = json.loads(args.compile_report.read_text(encoding="utf-8")) compiled = json.loads(args.compile_report.read_text(encoding="utf-8"))
@@ -143,10 +146,39 @@ def main() -> None:
"limits": [ "limits": [
"The historical source export does not attest an exact checkpoint revision.", "The historical source export does not attest an exact checkpoint revision.",
"The fixture is synthetic and has no token-to-audio quality tolerance.", "The fixture is synthetic and has no token-to-audio quality tolerance.",
"Requested compute unit is not actual placement without a device profile.",
"No predictor, vocoder, persistent cache loop or complete TTS stream was tested.", "No predictor, vocoder, persistent cache loop or complete TTS stream was tested.",
], ],
} }
if args.profile_report is not None:
profiled = json.loads(args.profile_report.read_text(encoding="utf-8"))
if (profiled.get("status") != "SUCCESS"
or profiled.get("model_id") != compiled["target_model_id"]
or profiled.get("device") != inferred["device"]
or profiled.get("options") != args.expected_run_options):
raise ValueError("profile did not execute the same target, device and route")
summary = profiled["profile"]["execution_summary"]
samples = summary["all_inference_times"]
if not samples or any(not isinstance(value, int) or value <= 0 for value in samples):
raise ValueError("profile has no valid inference samples")
units = Counter(row.get("compute_unit") for row in profiled["profile"]["execution_detail"])
if not units:
raise ValueError("profile has no compute-unit detail")
ordered = sorted(samples)
report["profile"] = {
"job_id": profiled["job_id"],
"sample_count": len(samples),
"sample_min_us": ordered[0],
"nearest_rank_p50_us": ordered[math.ceil(0.50 * len(samples)) - 1],
"nearest_rank_p95_us": ordered[math.ceil(0.95 * len(samples)) - 1],
"reported_estimated_inference_time_us": summary["estimated_inference_time"],
"reported_peak_memory_bytes": summary["estimated_inference_peak_memory"],
"reported_first_load_time_us": summary["first_load_time"],
"reported_first_load_peak_memory_bytes": summary["first_load_peak_memory"],
"compute_unit_row_counts": dict(units),
}
report["status"] = "component_inference_numeric_and_placement_measured"
else:
report["limits"].append("Requested compute unit is not actual placement without a device profile.")
args.output_report.write_text(json.dumps(report, indent=2) + "\n", encoding="utf-8") args.output_report.write_text(json.dumps(report, indent=2) + "\n", encoding="utf-8")
print(json.dumps({ print(json.dumps({
"status": report["status"], "status": report["status"],
@@ -29,6 +29,7 @@ def main() -> None:
parser = argparse.ArgumentParser(description=__doc__) parser = argparse.ArgumentParser(description=__doc__)
for name in ("model-dir", "cosmos-dir", "processor-dir", "fixture", "device-output", "audit-report", "output"): for name in ("model-dir", "cosmos-dir", "processor-dir", "fixture", "device-output", "audit-report", "output"):
parser.add_argument(f"--{name}", type=Path, required=True) parser.add_argument(f"--{name}", type=Path, required=True)
parser.add_argument("--device-name", default="Samsung Galaxy S25")
args = parser.parse_args() args = parser.parse_args()
os.environ["INTERNVLA_A1_COSMOS_DIR"] = str(args.cosmos_dir.resolve(strict=True)) os.environ["INTERNVLA_A1_COSMOS_DIR"] = str(args.cosmos_dir.resolve(strict=True))
@@ -48,10 +49,13 @@ def main() -> None:
if sha256(checkpoint) != CHECKPOINT_SHA256 or sha256(args.fixture) != FIXTURE_SHA256: if sha256(checkpoint) != CHECKPOINT_SHA256 or sha256(args.fixture) != FIXTURE_SHA256:
raise ValueError("checkpoint or encoder fixture differs from pinned source") raise ValueError("checkpoint or encoder fixture differs from pinned source")
audit = json.loads(args.audit_report.read_text(encoding="utf-8")) audit = json.loads(args.audit_report.read_text(encoding="utf-8"))
if (audit.get("status") != "component_inference_numeric_measured" if (audit.get("status") not in {
or audit.get("device", {}).get("name") != "Samsung Galaxy S25" "component_inference_numeric_measured",
"component_inference_numeric_and_placement_measured",
}
or audit.get("device", {}).get("name") != args.device_name
or audit.get("files", {}).get("device_output", {}).get("sha256") != sha256(args.device_output)): or audit.get("files", {}).get("device_output", {}).get("sha256") != sha256(args.device_output)):
raise ValueError("hosted latent lacks audited S25 provenance") raise ValueError(f"hosted latent lacks audited {args.device_name} provenance")
with np.load(args.fixture, allow_pickle=False) as fixture: with np.load(args.fixture, allow_pickle=False) as fixture:
source_latent = np.asarray(fixture["pattern_reference"]) source_latent = np.asarray(fixture["pattern_reference"])
with np.load(args.device_output, allow_pickle=False) as output: with np.load(args.device_output, allow_pickle=False) as output:
@@ -130,8 +134,9 @@ def main() -> None:
hosted = hosted_actions.astype(np.float64).ravel() hosted = hosted_actions.astype(np.float64).ravel()
delta = source - hosted delta = source - hosted
report = { report = {
"scope": "S25 hosted Cosmos latent injected into local CPU InternVLA policy with synthetic state/noise; no device-local policy or robot-task result", "scope": f"{args.device_name} hosted Cosmos latent injected into local CPU InternVLA policy with synthetic state/noise; no device-local policy or robot-task result",
"status": "synthetic_action_sensitivity_measured", "status": "synthetic_action_sensitivity_measured",
"hosted_device_name": args.device_name,
"policy_checkpoint_sha256": CHECKPOINT_SHA256, "policy_checkpoint_sha256": CHECKPOINT_SHA256,
"fixture_sha256": FIXTURE_SHA256, "fixture_sha256": FIXTURE_SHA256,
"hosted_output_sha256": sha256(args.device_output), "hosted_output_sha256": sha256(args.device_output),
@@ -18,6 +18,7 @@ from profile_local_text import HOST_ENVIRONMENT, save
async def run(args): async def run(args):
import torch import torch
import vllm_omni
from transformers import AutoTokenizer from transformers import AutoTokenizer
from vllm_omni import AsyncOmni from vllm_omni import AsyncOmni
@@ -35,6 +36,8 @@ async def run(args):
"status": "running", "status": "running",
"argv": sys.argv, "argv": sys.argv,
"python": sys.executable, "python": sys.executable,
"cwd": os.getcwd(),
"loaded_omni_path": str(Path(vllm_omni.__file__).resolve()),
"host_environment": HOST_ENVIRONMENT, "host_environment": HOST_ENVIRONMENT,
"platform": platform.platform(), "platform": platform.platform(),
"runtime": runtime_versions().to_dict(), "runtime": runtime_versions().to_dict(),
@@ -4,10 +4,10 @@ The [Qwen3-TTS vocoder study](evidence/qwen_tts_vocoder_qualcomm/README.md) reus
A local HX370 x86-64 LiteRT CPU control of the identical SA/RB3 TFLite binary matched the original ONNX CPU waveform within 3.45e-6 relative L2. The SA saturation is therefore not universal to the converted artifact; the exact SA CPU control passed numerical and placement checks, while the faulty GPU delegate operation remains unidentified. A local HX370 x86-64 LiteRT CPU control of the identical SA/RB3 TFLite binary matched the original ONNX CPU waveform within 3.45e-6 relative L2. The SA saturation is therefore not universal to the converted artifact; the exact SA CPU control passed numerical and placement checks, while the faulty GPU delegate operation remains unidentified.
The [talker decode-step attempt](evidence/qwen_tts_talker_qualcomm/README.md) pins a separate 28-layer source and synthetic cache fixture. Historical FP16 QNN context-binary inference executed on S24, S25, SA8775P and X Elite, but all four changed the CPU top-1 speech token with 8.57–9.43% logits relative L2. These artifacts are numerically unqualified on the fixture. RB3 rejected the floating-point-input artifact. Explicit X Elite ONNX CPU and SA8775P TFLite CPU fallback compiles passed. X Elite CPU-requested inference preserved CPU top-1 token 80 with logits relative L2 2.07e-5; its placement profile remains pending. SA8775P TFLite CPU-requested talker inference also preserved token 80 with logits relative L2 3.33e-5; its placement profile is pending. No complete talker loop or TTS stream passed. The [talker decode-step attempt](evidence/qwen_tts_talker_qualcomm/README.md) pins a separate 28-layer source and synthetic cache fixture. Historical FP16 QNN context-binary inference executed on S24, S25, SA8775P and X Elite, but all four changed the CPU top-1 speech token with 8.57–9.43% logits relative L2. These artifacts are numerically unqualified on the fixture. RB3 rejected the floating-point-input artifact. Explicit X Elite ONNX CPU and SA8775P TFLite CPU fallback compiles passed. X Elite CPU-requested inference preserved CPU top-1 token 80 with logits relative L2 2.07e-5; its 100-sample decode-step profile measured p50/p95 191.414/309.428 ms, all 1,744 execution-detail rows on CPU. SA8775P TFLite CPU-requested talker inference also preserved token 80 with logits relative L2 3.33e-5; its 100-sample decode-step profile measured p50/p95 52.635/76.925 ms, all 1,800 execution-detail rows on CPU. No complete talker loop or TTS stream passed.
The [MiniCPM-o speech-head study](evidence/minicpmo_tts_qualcomm/README.md) recovered the real 20-layer ONNX source and a matching retained fixture. S24 and S25 FP16 context-binary outputs preserved the ONNX CPU top-1 speech token with 0.346%/0.360% logits relative L2 on one step. Their 100-sample component profiles had p50/p95 8.537/9.006 ms and 7.197/7.615 ms, with all 1,029 execution-detail rows on NPU for each. FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P NPU-requested inference preserved CPU top-1 token 1867 with 0.3568% logits relative L2; its 100-sample profile measured p50/p95 11.901/13.038 ms, all 944 execution-detail rows on NPU. X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2; its 100-sample profile measured p50/p95 9.905/11.037 ms, all 944 execution-detail rows on NPU. RB3 FP16 QNN inference failed at graph load despite compile success; an explicit TFLite CPU fallback same-fixture inference preserved top-1 token 1867 with logits relative L2 1.26e-6 versus ONNX CPU; CPU placement profiling is pending. The complete multimodal model remains unqualified on those devices. The [MiniCPM-o speech-head study](evidence/minicpmo_tts_qualcomm/README.md) recovered the real 20-layer ONNX source and a matching retained fixture. S24 and S25 FP16 context-binary outputs preserved the ONNX CPU top-1 speech token with 0.346%/0.360% logits relative L2 on one step. Their 100-sample component profiles had p50/p95 8.537/9.006 ms and 7.197/7.615 ms, with all 1,029 execution-detail rows on NPU for each. FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P NPU-requested inference preserved CPU top-1 token 1867 with 0.3568% logits relative L2; its 100-sample profile measured p50/p95 11.901/13.038 ms, all 944 execution-detail rows on NPU. X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2; its 100-sample profile measured p50/p95 9.905/11.037 ms, all 944 execution-detail rows on NPU. RB3 FP16 QNN inference failed at graph load despite compile success; an explicit TFLite CPU fallback same-fixture inference preserved top-1 token 1867 with logits relative L2 1.26e-6 versus ONNX CPU; its 100-sample CPU-requested profile measured p50/p95 59.518/71.347 ms with all 1,008 execution-detail rows on CPU. The complete multimodal model remains unqualified on those devices.
The [native Windows Qwen3-TTS sustained analysis](evidence/qwen_tts_sustained_drift/README.md) finds a sharp playback throughput collapse around minute 26–27 in the historical 30-minute RTX 5090 Laptop stream. The last 17 serial requests all had simulated underruns with median RTF 1.331; the separate WSL run did not show that behavior. A revised local profiler now records timestamped device-wide NVML clocks, power, temperature, throttling and utilization to investigate the cause in a repeat run. The historical evidence cannot identify the cause or qualify native Windows sustained playback. The [native Windows Qwen3-TTS sustained analysis](evidence/qwen_tts_sustained_drift/README.md) finds a sharp playback throughput collapse around minute 26–27 in the historical 30-minute RTX 5090 Laptop stream. The last 17 serial requests all had simulated underruns with median RTF 1.331; the separate WSL run did not show that behavior. A revised local profiler now records timestamped device-wide NVML clocks, power, temperature, throttling and utilization to investigate the cause in a repeat run. The historical evidence cannot identify the cause or qualify native Windows sustained playback.
The [InternVLA Cosmos S25 attempt](evidence/internvla_cosmos_qualcomm/README.md) repackages the real policy's image encoder from its parity-tested ONNX export. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one pattern. Injecting it into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. S25 six-frame component placement profile returned 100 samples at p50/p95 49.665/51.382 ms, with all 121 execution-detail rows on NPU. The same encoder compiled to FP16 QNN DLC for X Elite CRD; same-fixture NPU-requested inference returned a finite latent at 0.0608% relative L2 / 64.32 dB SNR versus the source, with placement profiling pending. An SA8775P FP16 QNN DLC compile passed, with same-fixture NPU-requested inference pending. Device-local action-policy support remains unqualified. The checkpoint-named Place_Markpen dataset returned HTTP 404, and the official vLLM-Omni example archive in InternData-A1 returned HTTP 403 gated-access denial under the current authenticated account; [real-observation access evidence](evidence/internvla_real_observation_access/README.md) records the blocker without treating synthetic actions as task validation. The [InternVLA Cosmos S25 attempt](evidence/internvla_cosmos_qualcomm/README.md) repackages the real policy's image encoder from its parity-tested ONNX export. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one pattern. Injecting it into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. S25 six-frame component placement profile returned 100 samples at p50/p95 49.665/51.382 ms, with all 121 execution-detail rows on NPU. The same encoder compiled to FP16 QNN DLC for X Elite CRD; same-fixture NPU-requested inference returned a finite latent at 0.0608% relative L2 / 64.32 dB SNR versus the source, with its 100-sample six-frame profile at p50/p95 82.939/83.837 ms and all 121 execution-detail rows on NPU. An SA8775P FP16 QNN DLC inference returned a finite same-fixture latent at 0.0613% relative L2 / 64.25 dB SNR versus source; placement profiling is pending. Device-local action-policy support remains unqualified. The checkpoint-named Place_Markpen dataset returned HTTP 404, and the official vLLM-Omni example archive in InternData-A1 returned HTTP 403 gated-access denial under the current authenticated account; [real-observation access evidence](evidence/internvla_real_observation_access/README.md) records the blocker without treating synthetic actions as task validation.
@@ -4,10 +4,10 @@ This tests the **real Place_Markpen Cosmos CI8x8 image encoder component**, not
Workbench source model `mn4okj3rq` was [submitted for Galaxy S25 QNN DLC compilation](s25/compile_submission.json) as job `jpxl4d1jp`, with `default_graph_htp_precision=FLOAT16`. The [compile](s25/compile_report.json) succeeded and produced target model `mqkp12rxm`. [Same-fixture pattern inference](s25/inference_report.json) requesting NPU also succeeded. The [audit](s25/audit_report.json) found **0.0524% relative L2 / 65.61 dB SNR** versus the pinned source latent, with finite FP32 `[6,16,32,32]` output. This establishes numerical behavior on one synthetic encoder fixture; the [placement profile](s25/profile_report.json) attributed all 121 execution-detail rows to NPU. Its 100 six-frame component samples had nearest-rank p50/p95 49.665/51.382 ms (minimum 47.438 ms); reported inference peak memory was 158,232,576 bytes. The service reported first/warm load times of 20.996/0.178 s and corresponding peak memories of 679,317,504/279,449,600 bytes for this component job. The combined audit binds placement, numerical output and exact target/device. These exclude the action policy, data handoff, cold-load cost and sustained operation. Workbench source model `mn4okj3rq` was [submitted for Galaxy S25 QNN DLC compilation](s25/compile_submission.json) as job `jpxl4d1jp`, with `default_graph_htp_precision=FLOAT16`. The [compile](s25/compile_report.json) succeeded and produced target model `mqkp12rxm`. [Same-fixture pattern inference](s25/inference_report.json) requesting NPU also succeeded. The [audit](s25/audit_report.json) found **0.0524% relative L2 / 65.61 dB SNR** versus the pinned source latent, with finite FP32 `[6,16,32,32]` output. This establishes numerical behavior on one synthetic encoder fixture; the [placement profile](s25/profile_report.json) attributed all 121 execution-detail rows to NPU. Its 100 six-frame component samples had nearest-rank p50/p95 49.665/51.382 ms (minimum 47.438 ms); reported inference peak memory was 158,232,576 bytes. The service reported first/warm load times of 20.996/0.178 s and corresponding peak memories of 679,317,504/279,449,600 bytes for this component job. The combined audit binds placement, numerical output and exact target/device. These exclude the action policy, data handoff, cold-load cost and sustained operation.
The same source model also [compiled to FP16 QNN DLC on Snapdragon X Elite CRD](xelite/compile_report.json), producing target `mmxj0w8kq` for Windows ARM. [Same-fixture NPU-requested inference](xelite/inference_report.json) on dataset `d26qo8p57` returned a finite FP32 `[6,16,32,32]` latent. Its [audit](xelite/audit_report.json) measured **0.0608% relative L2 / 64.32 dB SNR** versus the pinned source pattern, with cosine 0.999999815. This is one synthetic component inference; the [NPU placement profile](xelite/profile_submission.json) has been submitted and remains pending. Neither inference request nor QNN artifact alone establishes actual NPU node assignment, action quality or a complete device-local policy. The same source model also [compiled to FP16 QNN DLC on Snapdragon X Elite CRD](xelite/compile_report.json), producing target `mmxj0w8kq` for Windows ARM. [Same-fixture NPU-requested inference](xelite/inference_report.json) on dataset `d26qo8p57` returned a finite FP32 `[6,16,32,32]` latent. Its [audit](xelite/audit_report.json) measured **0.0608% relative L2 / 64.32 dB SNR** versus the pinned source pattern, with cosine 0.999999815. The [placement profile](xelite/profile_report.json) returned 100 six-frame component samples at nearest-rank p50/p95 **82.939/83.837 ms**, with all **121 execution-detail rows on NPU**. The 82.446 ms estimated-time field is the sample minimum. Reported inference peak memory was 24,059,904 bytes and first-load peak 692,891,648 bytes. This is component evidence only; neither action quality nor a complete device-local policy was tested.
The same source also [compiled to FP16 QNN DLC on SA8775P ADP](sa8775p/compile_report.json), producing target mn750wz8m for Android. [Same-fixture NPU-requested inference](sa8775p/inference_submission.json) was submitted using dataset d26qo8p57; numerical and placement outcomes remain pending. The same source also [compiled to FP16 QNN DLC on SA8775P ADP](sa8775p/compile_report.json), producing target mn750wz8m for Android. [Same-fixture NPU-requested inference](sa8775p/inference_report.json) using dataset d26qo8p57 returned a finite FP32 [6,16,32,32] latent. Its [audit](sa8775p/audit_report.json) measured **0.0613% relative L2 / 64.25 dB SNR** versus the pinned source pattern, with cosine 0.999999812. Actual NPU node assignment and timing await the [submitted placement profile](sa8775p/profile_submission.json); no complete local policy follows.
The [generic collector](../../../../experiments/collect_aihub_component.py) preserves terminal jobs, and the [Cosmos auditor](../../../../experiments/audit_internvla_cosmos_qualcomm.py) checks source/fixture hashes, target/device/dataset identity, latent shape and finite numerical difference against the pinned source reference. A separate [synthetic action sensitivity check](s25/action_sensitivity.json) injected either the pinned source or S25 latent into the same real Place_Markpen policy on **local CPU**, with zero state/noise. The [reproducer](../../../../experiments/probe_internvla_hosted_latent_actions.py) returned finite `[1,50,32]` actions; hosted-vs-source action relative L2 was **0.742%**, maximum absolute difference **0.0115**, cosine **0.999973**. Repeating the source policy gave bitwise identical actions. This is a cross-device numerical handoff experiment, not a device-local policy or robot-task pass; no real-observation action tolerance is known. The [generic collector](../../../../experiments/collect_aihub_component.py) preserves terminal jobs, and the [Cosmos auditor](../../../../experiments/audit_internvla_cosmos_qualcomm.py) checks source/fixture hashes, target/device/dataset identity, latent shape and finite numerical difference against the pinned source reference. A separate [synthetic action sensitivity check](s25/action_sensitivity.json) injected either the pinned source or S25 latent into the same real Place_Markpen policy on **local CPU**, with zero state/noise. The [reproducer](../../../../experiments/probe_internvla_hosted_latent_actions.py) returned finite `[1,50,32]` actions; hosted-vs-source action relative L2 was **0.742%**, maximum absolute difference **0.0115**, cosine **0.999973**. Repeating the source policy gave bitwise identical actions. This is a cross-device numerical handoff experiment, not a device-local policy or robot-task pass; no real-observation action tolerance is known.
Next: audit X Elite placement and SA8775P same-fixture output, test representative real observations and action tolerances, then build and validate a device-local stateful policy. Full multimodal observation processing, admission, cancellation and sustained profiling remain open. Next: audit SA8775P placement, test representative real observations and action tolerances, then build and validate a device-local stateful policy. Full multimodal observation processing, admission, cancellation and sustained profiling remain open.
@@ -0,0 +1,54 @@
{
"scope": "one six-frame Cosmos image encoder on synthetic pattern; no complete InternVLA policy",
"status": "component_inference_numeric_measured",
"device": {
"name": "SA8775P ADP",
"os": "14",
"attributes": [
"os:android",
"framework:tflite",
"framework:onnx",
"vendor:qualcomm",
"abi:aarch64-android",
"format:auto",
"chipset:qualcomm-sa8775p",
"hexagon:v73",
"soc-model:52",
"htp-supports-fp16:true",
"htp-supports-weight-sharing:true",
"framework:qnn"
]
},
"source_model_id": "mn4okj3rq",
"target_model_id": "mn750wz8m",
"compile_job_id": "j5687wx6g",
"inference_job_id": "jp2rwe2rg",
"input_dataset_id": "d26qo8p57",
"requested_run_options": "--compute_unit npu",
"files": {
"source_onnx": {
"bytes": 125112088,
"sha256": "656a5263f584851a1457ef6dcf60b995f7c67b056b3c90a4f33c814f7caf9ec9"
},
"fixture": {
"bytes": 2081321,
"sha256": "dcbaf1391ca914d4bbccc1cb384daf05666e771d9cdae90c4f069ec8fcf05368"
},
"device_output": {
"bytes": 185475,
"sha256": "b30e4a7e312a0ceaa058fa1f6fe338b3dbc21ef7ee131e9347997822eccd0f40"
}
},
"output_dtype": "float32",
"source_vs_device": {
"relative_l2": 0.0006133529653991887,
"snr_db": 64.24579061372152,
"max_abs_error": 0.009660720825195312,
"cosine": 0.9999998122677838
},
"limits": [
"The fixture is synthetic; no real-observation or action-quality tolerance is established.",
"The action policy, state lifecycle, admission and complete request were not run on this device.",
"Requested NPU is not actual per-node placement without a device profile."
]
}
@@ -0,0 +1,32 @@
{
"timestamp_utc": "2026-09-24T01:18:21.481644+00:00",
"qai_hub_version": "0.55.0",
"facility": "Qualcomm AI Hub Workbench",
"job_id": "jp2rwe2rg",
"url": "https://workbench.aihub.qualcomm.com/jobs/jp2rwe2rg/",
"status": "SUCCESS",
"status_message": "",
"options": "--compute_unit npu",
"hub_version": "aihub-2026.09.11.0",
"device": {
"name": "SA8775P ADP",
"os": "14",
"attributes": [
"os:android",
"framework:tflite",
"framework:onnx",
"vendor:qualcomm",
"abi:aarch64-android",
"format:auto",
"chipset:qualcomm-sa8775p",
"hexagon:v73",
"soc-model:52",
"htp-supports-fp16:true",
"htp-supports-weight-sharing:true",
"framework:qnn"
]
},
"model_id": "mn750wz8m",
"input_dataset_id": "d26qo8p57",
"outputs": "device_output.npz"
}
@@ -0,0 +1,13 @@
{
"timestamp_utc": "2026-09-24T01:18:52.829788+00:00",
"facility": "Qualcomm AI Hub Workbench",
"qai_hub_version": "0.55.0",
"source_compile_job_id": "j5687wx6g",
"target_model_id": "mn750wz8m",
"device": "SA8775P ADP",
"options": "--compute_unit npu",
"kind": "profile",
"status": "SUBMITTED",
"job_id": "jpe7z7ov5",
"url": "https://workbench.aihub.qualcomm.com/jobs/jpe7z7ov5/"
}
@@ -1,6 +1,6 @@
{ {
"scope": "one six-frame Cosmos image encoder on synthetic pattern; no complete InternVLA policy", "scope": "one six-frame Cosmos image encoder on synthetic pattern; no complete InternVLA policy",
"status": "component_inference_numeric_measured", "status": "component_inference_numeric_and_placement_measured",
"device": { "device": {
"name": "Snapdragon X Elite CRD", "name": "Snapdragon X Elite CRD",
"os": "11", "os": "11",
@@ -48,7 +48,22 @@
}, },
"limits": [ "limits": [
"The fixture is synthetic; no real-observation or action-quality tolerance is established.", "The fixture is synthetic; no real-observation or action-quality tolerance is established.",
"The action policy, state lifecycle, admission and complete request were not run on this device.", "The action policy, state lifecycle, admission and complete request were not run on this device."
"Requested NPU is not actual per-node placement without a device profile." ],
] "profile": {
"job_id": "j5wl77r6p",
"sample_count": 100,
"sample_min_us": 82446,
"nearest_rank_p50_us": 82939,
"nearest_rank_p95_us": 83837,
"reported_estimated_inference_time_us": 82446,
"reported_peak_memory_bytes": 24059904,
"reported_first_load_time_us": 50583490,
"reported_warm_load_time_us": 1164543,
"reported_first_load_peak_memory_bytes": 692891648,
"reported_warm_load_peak_memory_bytes": 15200256,
"compute_unit_row_counts": {
"NPU": 121
}
}
} }
@@ -6,6 +6,6 @@ The [S25 historical inference](s25/inference_report.json) used dataset `d7gwxe3y
The historical [S24](s24/compile_report.json) and [S25](s25/compile_report.json) FP16 QNN context-binary compiles and inference jobs succeeded. Against the **same original ONNX CPU fixture**, [S24](s24/audit_report.json) logits relative L2 was **0.3459%** and max K/V relative L2 **1.235%**; [S25](s25/audit_report.json) logits relative L2 was **0.3605%** and max K/V relative L2 **1.311%**. Both preserved the CPU top-1 speech token **1867** on this one synthetic step. The raw [S24](s24/device_output.npz) and [S25](s25/device_output.npz) arrays are retained. The [S24](s24/profile_report.json) and [S25](s25/profile_report.json) profiles each returned 100 fixed-cache 20-layer component samples, with nearest-rank p50/p95 **8.537/9.006 ms** and **7.197/7.615 ms**, respectively. All **1,029 execution-detail rows** in each profile were attributed to NPU. The service estimated-time fields were the sample minima (8.211/7.125 ms), not medians. Reported inference peak memory was 165,433,344/185,991,168 bytes. The [S24](s24/audit_report.json) and [S25](s25/audit_report.json) audits now bind numerical outputs and placement to the exact target/device; generated-audio quality and complete model state remain unverified. The historical [S24](s24/compile_report.json) and [S25](s25/compile_report.json) FP16 QNN context-binary compiles and inference jobs succeeded. Against the **same original ONNX CPU fixture**, [S24](s24/audit_report.json) logits relative L2 was **0.3459%** and max K/V relative L2 **1.235%**; [S25](s25/audit_report.json) logits relative L2 was **0.3605%** and max K/V relative L2 **1.311%**. Both preserved the CPU top-1 speech token **1867** on this one synthetic step. The raw [S24](s24/device_output.npz) and [S25](s25/device_output.npz) arrays are retained. The [S24](s24/profile_report.json) and [S25](s25/profile_report.json) profiles each returned 100 fixed-cache 20-layer component samples, with nearest-rank p50/p95 **8.537/9.006 ms** and **7.197/7.615 ms**, respectively. All **1,029 execution-detail rows** in each profile were attributed to NPU. The service estimated-time fields were the sample minima (8.211/7.125 ms), not medians. Reported inference peak memory was 165,433,344/185,991,168 bytes. The [S24](s24/audit_report.json) and [S25](s25/audit_report.json) audits now bind numerical outputs and placement to the exact target/device; generated-audio quality and complete model state remain unverified.
New FP16 QNN DLC [SA8775P](sa8775p/compile_report.json), [Snapdragon X Elite](xelite/compile_report.json) and [RB3 Gen 2](rb3/compile_report.json) compiles passed from the same source. The SA8775P [NPU-requested inference](sa8775p/inference_report.json) returned all 41 finite tensors on the retained fixture. Its [audit](sa8775p/audit_report.json) found **0.3568% logits relative L2** versus ONNX CPU, preserved CPU top-1 speech token **1867**, and measured maximum K/V relative L2 **1.225%**. The [SA8775P placement profile](sa8775p/profile_report.json) returned 100 component samples with nearest-rank p50/p95 **11.901/13.038 ms**; all **944 execution-detail rows** were attributed to NPU. The 11.627 ms estimated-time field is the sample minimum, not median. Reported inference peak memory was 134,696,960 bytes and first-load peak 1,322,835,968 bytes, neither a whole-model co-residency measurement. The X Elite [NPU-requested inference](xelite/inference_report.json) also returned all 41 finite tensors: its [audit](xelite/audit_report.json) found **0.3170% logits relative L2**, preserved CPU top-1 token **1867**, and measured maximum K/V relative L2 **1.221%**. Its [placement profile](xelite/profile_report.json) returned 100 component samples with nearest-rank p50/p95 **9.905/11.037 ms**; all **944 execution-detail rows** were attributed to NPU. The 9.106 ms estimated-time field is the sample minimum. Reported inference peak memory was 79,208,448 bytes and first-load peak 1,081,573,376 bytes. The exact RB3 [FP16 QNN DLC inference](rb3/inference_report.json) **failed** while loading the graph with QnnModel_composeGraphsFromDlc: MODEL_GRAPH_ERROR, despite a successful compile. RB3 does not advertise FP16 HTP support in the device attributes; this failure constrains the tested artifact, not every RB3 path. An explicit [TFLite CPU fallback compile](rb3/cpu_fallback/compile_report.json) succeeded, producing target mnwv5d7wq; [same-fixture CPU-requested inference](rb3/cpu_fallback/inference_report.json) returned 41 finite tensors, preserved CPU top-1 token **1867**, and had logits relative L2 **1.26e-6** and maximum K/V relative L2 **1.58e-6** versus the original ONNX CPU fixture. This establishes a numerical CPU fallback component, but actual per-node CPU placement and timing await its [submitted profile](rb3/cpu_fallback/profile_submission.json). The [collector](../../../../experiments/collect_aihub_component.py) retrieves terminal jobs, and the [auditor](../../../../experiments/audit_minicpmo_tts_qualcomm.py) pins hashes and checks all 41 output tensors. New FP16 QNN DLC [SA8775P](sa8775p/compile_report.json), [Snapdragon X Elite](xelite/compile_report.json) and [RB3 Gen 2](rb3/compile_report.json) compiles passed from the same source. The SA8775P [NPU-requested inference](sa8775p/inference_report.json) returned all 41 finite tensors on the retained fixture. Its [audit](sa8775p/audit_report.json) found **0.3568% logits relative L2** versus ONNX CPU, preserved CPU top-1 speech token **1867**, and measured maximum K/V relative L2 **1.225%**. The [SA8775P placement profile](sa8775p/profile_report.json) returned 100 component samples with nearest-rank p50/p95 **11.901/13.038 ms**; all **944 execution-detail rows** were attributed to NPU. The 11.627 ms estimated-time field is the sample minimum, not median. Reported inference peak memory was 134,696,960 bytes and first-load peak 1,322,835,968 bytes, neither a whole-model co-residency measurement. The X Elite [NPU-requested inference](xelite/inference_report.json) also returned all 41 finite tensors: its [audit](xelite/audit_report.json) found **0.3170% logits relative L2**, preserved CPU top-1 token **1867**, and measured maximum K/V relative L2 **1.221%**. Its [placement profile](xelite/profile_report.json) returned 100 component samples with nearest-rank p50/p95 **9.905/11.037 ms**; all **944 execution-detail rows** were attributed to NPU. The 9.106 ms estimated-time field is the sample minimum. Reported inference peak memory was 79,208,448 bytes and first-load peak 1,081,573,376 bytes. The exact RB3 [FP16 QNN DLC inference](rb3/inference_report.json) **failed** while loading the graph with QnnModel_composeGraphsFromDlc: MODEL_GRAPH_ERROR, despite a successful compile. RB3 does not advertise FP16 HTP support in the device attributes; this failure constrains the tested artifact, not every RB3 path. An explicit [TFLite CPU fallback compile](rb3/cpu_fallback/compile_report.json) succeeded, producing target mnwv5d7wq; [same-fixture CPU-requested inference](rb3/cpu_fallback/inference_report.json) returned 41 finite tensors, preserved CPU top-1 token **1867**, and had logits relative L2 **1.26e-6** and maximum K/V relative L2 **1.58e-6** versus the original ONNX CPU fixture. The [CPU fallback profile](rb3/cpu_fallback/profile_report.json) returned 100 decode-step component samples at nearest-rank p50/p95 **59.518/71.347 ms**, with all **1,008 execution-detail rows on CPU**. The 59.007 ms estimated-time field is the sample minimum. Reported inference peak memory was 820,981,760 bytes. This is a numerical and placement-passing component fallback, not a full model request. The [collector](../../../../experiments/collect_aihub_component.py) retrieves terminal jobs, and the [auditor](../../../../experiments/audit_minicpmo_tts_qualcomm.py) pins hashes and checks all 41 output tensors.
Next: audit those device outputs and placement, then qualify the entire model with real multimodal inputs, stateful speech generation, task/audio quality, memory admission, cancellation and sustained power/thermal behavior. AI Hub is the test facility, not a deployment dependency. Next: audit those device outputs and placement, then qualify the entire model with real multimodal inputs, stateful speech generation, task/audio quality, memory admission, cancellation and sustained power/thermal behavior. AI Hub is the test facility, not a deployment dependency.
@@ -1,6 +1,6 @@
{ {
"scope": "one fixed-cache 20-layer MiniCPM-o speech-head decode step; no complete multimodal stream", "scope": "one fixed-cache 20-layer MiniCPM-o speech-head decode step; no complete multimodal stream",
"status": "component_inference_numeric_measured", "status": "component_inference_numeric_and_placement_measured",
"device": { "device": {
"name": "Dragonwing RB3 Gen 2 Vision Kit", "name": "Dragonwing RB3 Gen 2 Vision Kit",
"os": "1.6", "os": "1.6",
@@ -298,7 +298,20 @@
"limits": [ "limits": [
"The historical ONNX export does not attest the exact source checkpoint revision.", "The historical ONNX export does not attest the exact source checkpoint revision.",
"One retained decode fixture cannot establish speech-token or generated-audio quality.", "One retained decode fixture cannot establish speech-token or generated-audio quality.",
"Thinker, audio/vision encoders, vocoder, state loop and complete Omni request were not tested.", "Thinker, audio/vision encoders, vocoder, state loop and complete Omni request were not tested."
"Actual CPU/NPU placement requires a device profile; a request or compiled artifact alone is insufficient." ],
] "profile": {
"job_id": "jprl00r0p",
"sample_count": 100,
"sample_min_us": 59007,
"nearest_rank_p50_us": 59518,
"nearest_rank_p95_us": 71347,
"reported_estimated_inference_time_us": 59007,
"reported_peak_memory_bytes": 820981760,
"reported_first_load_time_us": 275127,
"reported_first_load_peak_memory_bytes": 769884160,
"compute_unit_row_counts": {
"CPU": 1008
}
}
} }
@@ -6,7 +6,7 @@ The deterministic synthetic input dataset `d7dw6pgy2` has SHA256 `0bf538e1b64dfc
Historical FP16 QNN context-binary [S24](s24/compile_report.json), [S25](s25/compile_report.json), [SA8775P](sa8775p/compile_report.json), and [Snapdragon X Elite](xelite/compile_report.json) compile jobs succeeded. Same-fixture NPU-requested inference succeeded on all four, but **all changed the CPU reference top-1 logit** on this synthetic step: CPU token 80 versus S24/SA/X Elite 1720 and S25 412. The [S24](s24/audit_report.json), [S25](s25/audit_report.json), [SA8775P](sa8775p/audit_report.json) and [X Elite](xelite/audit_report.json) audits report logits relative L2 **9.33%/8.99%/9.43%/8.57%**, respectively, with maximum K/V relative L2 **11.23%/11.71%/11.54%/11.50%**. These FP16 artifacts are numerically unqualified for this fixture; a successful graph execution is not speech-token parity. The [RB3 compile](rb3/compile_report.json) rejected floating-point input on this device. These are artifact-specific outcomes, not full-model support claims. Historical FP16 QNN context-binary [S24](s24/compile_report.json), [S25](s25/compile_report.json), [SA8775P](sa8775p/compile_report.json), and [Snapdragon X Elite](xelite/compile_report.json) compile jobs succeeded. Same-fixture NPU-requested inference succeeded on all four, but **all changed the CPU reference top-1 logit** on this synthetic step: CPU token 80 versus S24/SA/X Elite 1720 and S25 412. The [S24](s24/audit_report.json), [S25](s25/audit_report.json), [SA8775P](sa8775p/audit_report.json) and [X Elite](xelite/audit_report.json) audits report logits relative L2 **9.33%/8.99%/9.43%/8.57%**, respectively, with maximum K/V relative L2 **11.23%/11.71%/11.54%/11.50%**. These FP16 artifacts are numerically unqualified for this fixture; a successful graph execution is not speech-token parity. The [RB3 compile](rb3/compile_report.json) rejected floating-point input on this device. These are artifact-specific outcomes, not full-model support claims.
Explicit [X Elite ONNX CPU](xelite/cpu_fallback/compile_report.json) and [SA8775P TFLite CPU](sa8775p/cpu_fallback/compile_report.json) fallback compiles succeeded from the same source. The same-fixture [X Elite CPU-requested inference](xelite/cpu_fallback/inference_report.json) succeeded. Its [audit](xelite/cpu_fallback/audit_report.json) found logits relative L2 2.07e-5, maximum new-K/V relative L2 2.18e-5, and preserved ONNX CPU top-1 speech token 80. A [CPU-requested placement profile](xelite/cpu_fallback/profile_submission.json) was submitted and remains pending. The [SA8775P TFLite CPU-requested inference](sa8775p/cpu_fallback/inference_report.json) also succeeded. Its [audit](sa8775p/cpu_fallback/audit_report.json) found logits relative L2 3.33e-5, maximum new-K/V relative L2 3.41e-5, and preserved CPU top-1 speech token 80. The [CPU-requested placement profile](sa8775p/cpu_fallback/profile_submission.json) is pending. Both explicit CPU routes are numerically close on one decode step, not complete talker loops or TTS deployments. Explicit [X Elite ONNX CPU](xelite/cpu_fallback/compile_report.json) and [SA8775P TFLite CPU](sa8775p/cpu_fallback/compile_report.json) fallback compiles succeeded from the same source. The same-fixture [X Elite CPU-requested inference](xelite/cpu_fallback/inference_report.json) succeeded. Its [audit](xelite/cpu_fallback/audit_report.json) found logits relative L2 2.07e-5, maximum new-K/V relative L2 2.18e-5, and preserved ONNX CPU top-1 speech token 80. The [CPU-requested placement profile](xelite/cpu_fallback/profile_report.json) returned 100 decode-step component samples at nearest-rank p50/p95 **191.414/309.428 ms**, with all **1,744 execution-detail rows on CPU**. The 142.921 ms estimated-time field is the sample minimum; reported inference peak memory was 1,928,019,968 bytes. The [SA8775P TFLite CPU-requested inference](sa8775p/cpu_fallback/inference_report.json) also succeeded. Its [audit](sa8775p/cpu_fallback/audit_report.json) found logits relative L2 3.33e-5, maximum new-K/V relative L2 3.41e-5, and preserved CPU top-1 speech token 80. The [CPU-requested placement profile](sa8775p/cpu_fallback/profile_report.json) returned 100 decode-step component samples at nearest-rank p50/p95 **52.635/76.925 ms**, with all **1,800 execution-detail rows on CPU**. The 45.688 ms estimated-time field is the sample minimum; reported inference peak memory was 1,913,794,560 bytes. Both explicit CPU routes are numerically close on one decode step, not complete talker loops or TTS deployments.
The [collector](../../../../experiments/collect_aihub_component.py) retrieves terminal inference outputs without resubmission. The [tensor auditor](../../../../experiments/audit_qwen_tts_talker_qualcomm.py) enforces the source, fixture, compile target, exact device, output shape and finiteness contracts, and reports logits top-1 plus relative error for all 56 new K/V tensors. It does not convert those metrics to a quality pass without an explicit token/audio tolerance. Its 57-output contract was exercised against the pinned CPU reference as a local self-check; that self-check is not device evidence. Requested NPU placement still needs a profile. The [collector](../../../../experiments/collect_aihub_component.py) retrieves terminal inference outputs without resubmission. The [tensor auditor](../../../../experiments/audit_qwen_tts_talker_qualcomm.py) enforces the source, fixture, compile target, exact device, output shape and finiteness contracts, and reports logits top-1 plus relative error for all 56 new K/V tensors. It does not convert those metrics to a quality pass without an explicit token/audio tolerance. Its 57-output contract was exercised against the pinned CPU reference as a local self-check; that self-check is not device evidence. Requested NPU placement still needs a profile.
@@ -1,6 +1,6 @@
{ {
"scope": "one synthetic fixed-cache 28-layer Qwen3-TTS talker decode step; no complete TTS stream", "scope": "one synthetic fixed-cache 28-layer Qwen3-TTS talker decode step; no complete TTS stream",
"status": "component_executed_numeric_measured", "status": "component_inference_numeric_and_placement_measured",
"device": { "device": {
"name": "SA8775P ADP", "name": "SA8775P ADP",
"os": "14", "os": "14",
@@ -394,7 +394,20 @@
"limits": [ "limits": [
"The historical source export does not attest an exact checkpoint revision.", "The historical source export does not attest an exact checkpoint revision.",
"The fixture is synthetic and has no token-to-audio quality tolerance.", "The fixture is synthetic and has no token-to-audio quality tolerance.",
"Requested compute unit is not actual placement without a device profile.",
"No predictor, vocoder, persistent cache loop or complete TTS stream was tested." "No predictor, vocoder, persistent cache loop or complete TTS stream was tested."
] ],
"profile": {
"job_id": "jgdd33x6g",
"sample_count": 100,
"sample_min_us": 45688,
"nearest_rank_p50_us": 52635,
"nearest_rank_p95_us": 76925,
"reported_estimated_inference_time_us": 45688,
"reported_peak_memory_bytes": 1913794560,
"reported_first_load_time_us": 6679118,
"reported_first_load_peak_memory_bytes": 1846460416,
"compute_unit_row_counts": {
"CPU": 1800
}
}
} }
@@ -1,6 +1,6 @@
{ {
"scope": "one synthetic fixed-cache 28-layer Qwen3-TTS talker decode step; no complete TTS stream", "scope": "one synthetic fixed-cache 28-layer Qwen3-TTS talker decode step; no complete TTS stream",
"status": "component_executed_numeric_measured", "status": "component_inference_numeric_and_placement_measured",
"device": { "device": {
"name": "Snapdragon X Elite CRD", "name": "Snapdragon X Elite CRD",
"os": "11", "os": "11",
@@ -394,7 +394,20 @@
"limits": [ "limits": [
"The historical source export does not attest an exact checkpoint revision.", "The historical source export does not attest an exact checkpoint revision.",
"The fixture is synthetic and has no token-to-audio quality tolerance.", "The fixture is synthetic and has no token-to-audio quality tolerance.",
"Requested compute unit is not actual placement without a device profile.",
"No predictor, vocoder, persistent cache loop or complete TTS stream was tested." "No predictor, vocoder, persistent cache loop or complete TTS stream was tested."
] ],
"profile": {
"job_id": "jp2rwed4g",
"sample_count": 100,
"sample_min_us": 142921,
"nearest_rank_p50_us": 191414,
"nearest_rank_p95_us": 309428,
"reported_estimated_inference_time_us": 142921,
"reported_peak_memory_bytes": 1928019968,
"reported_first_load_time_us": 2102540,
"reported_first_load_peak_memory_bytes": 1806598144,
"compute_unit_row_counts": {
"CPU": 1744
}
}
} }
@@ -16,11 +16,11 @@ A [Qwen3-TTS Code2Wav follow-up](../../../e2e_expansion_20260924/evidence/qwen_t
The identical SA/RB3 TFLite binary matched original ONNX CPU within 3.45e-6 relative L2 in a separate local HX370 x86-64 LiteRT CPU control. Together with the exact SA CPU-requested parity result, this narrows the SA saturation to the GPU-requested route; the RB3 embedded CPU route and specific faulty delegate operation remain unverified. The identical SA/RB3 TFLite binary matched original ONNX CPU within 3.45e-6 relative L2 in a separate local HX370 x86-64 LiteRT CPU control. Together with the exact SA CPU-requested parity result, this narrows the SA saturation to the GPU-requested route; the RB3 embedded CPU route and specific faulty delegate operation remain unverified.
A separate [Qwen3-TTS talker decode-step attempt](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) recovered a 28-layer ONNX export and synthetic fixed-cache fixture. Historical FP16 QNN context-binary inference executed on S24, S25, SA8775P and X Elite, but all four changed the CPU reference top-1 speech token with 8.57–9.43% logits relative L2; they are numerically unqualified on this fixture. An RB3 compile rejected floating-point input. Separate CPU-requested talker steps on X Elite ONNX and SA8775P TFLite preserved CPU token 80 with logits relative L2 2.07e-5 and 3.33e-5; both placement profiles are pending. No complete talker loop or TTS stream follows. A separate [Qwen3-TTS talker decode-step attempt](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) recovered a 28-layer ONNX export and synthetic fixed-cache fixture. Historical FP16 QNN context-binary inference executed on S24, S25, SA8775P and X Elite, but all four changed the CPU reference top-1 speech token with 8.57–9.43% logits relative L2; they are numerically unqualified on this fixture. An RB3 compile rejected floating-point input. Separate CPU-requested talker steps on X Elite ONNX and SA8775P TFLite preserved CPU token 80 with logits relative L2 2.07e-5 and 3.33e-5; X Elite had 100 component samples at p50/p95 191.414/309.428 ms with all 1,744 execution-detail rows on CPU; SA had 100 at 52.635/76.925 ms with all 1,800 rows on CPU. No complete talker loop or TTS stream follows.
A [Galaxy S25 InternVLA Cosmos attempt](../../../e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md) packaged the real Place_Markpen image encoder into one ONNX file after two CPU parity fixtures. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one pattern; its 100-sample six-frame component profile had p50/p95 49.665/51.382 ms and all 121 execution-detail rows on NPU. Injecting that latent into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. The same encoder compiled to FP16 QNN DLC for X Elite CRD; same-fixture NPU-requested inference returned a finite latent at 0.0608% relative L2 versus source, with actual placement profiling pending. This advances component and action-sensitivity evidence, not device-local full-policy qualification. The Place_Markpen checkpoint names dataset a2d_real/a2d_pen_holder, which returned HTTP 404 under the current authenticated Hugging Face account. The official vLLM-Omni example names a different Place_Markpen archive in InternData-A1, but access to that gated dataset returned HTTP 403; [the access record](../../../e2e_expansion_20260924/evidence/internvla_real_observation_access/README.md) leaves real-observation reference-action validation open. A [Galaxy S25 InternVLA Cosmos attempt](../../../e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md) packaged the real Place_Markpen image encoder into one ONNX file after two CPU parity fixtures. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one pattern; its 100-sample six-frame component profile had p50/p95 49.665/51.382 ms and all 121 execution-detail rows on NPU. Injecting that latent into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. The same encoder compiled to FP16 QNN DLC for X Elite CRD and SA8775P; same-fixture NPU-requested inferences returned finite latents at 0.0608% and 0.0613% relative L2 versus source, with X Elite profiled on NPU and SA8775P placement pending. This advances component and action-sensitivity evidence, not device-local full-policy qualification. The Place_Markpen checkpoint names dataset a2d_real/a2d_pen_holder, which returned HTTP 404 under the current authenticated Hugging Face account. The official vLLM-Omni example names a different Place_Markpen archive in InternData-A1, but access to that gated dataset returned HTTP 403; [the access record](../../../e2e_expansion_20260924/evidence/internvla_real_observation_access/README.md) leaves real-observation reference-action validation open.
A [MiniCPM-o speech-head study](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md) recovered the historical 20-layer source and matching S24/S25 fixtures. Both FP16 context-binary outputs preserved the ONNX CPU top-1 speech token on this one decode step with logits relative L2 of 0.346%/0.360%. S24/S25 100-sample component profiles attributed all 1,029 execution-detail rows to NPU, with p50/p95 8.537/9.006 ms and 7.197/7.615 ms. New FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P same-fixture inference preserved CPU top-1 token 1867 with 0.3568% logits relative L2; its 100-sample component profile measured p50/p95 11.901/13.038 ms with all 944 execution-detail rows on NPU. X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2; its 100-sample component profile measured p50/p95 9.905/11.037 ms with all 944 execution-detail rows on NPU; RB3 FP16 QNN inference failed at graph load; its explicit TFLite CPU fallback returned 41 finite outputs, preserved CPU top-1 token 1867 and had logits relative L2 1.26e-6 versus ONNX CPU, with placement profiling pending. These are speech-head component results, not multimodal MiniCPM-o requests. A [MiniCPM-o speech-head study](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md) recovered the historical 20-layer source and matching S24/S25 fixtures. Both FP16 context-binary outputs preserved the ONNX CPU top-1 speech token on this one decode step with logits relative L2 of 0.346%/0.360%. S24/S25 100-sample component profiles attributed all 1,029 execution-detail rows to NPU, with p50/p95 8.537/9.006 ms and 7.197/7.615 ms. New FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P same-fixture inference preserved CPU top-1 token 1867 with 0.3568% logits relative L2; its 100-sample component profile measured p50/p95 11.901/13.038 ms with all 944 execution-detail rows on NPU. X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2; its 100-sample component profile measured p50/p95 9.905/11.037 ms with all 944 execution-detail rows on NPU; RB3 FP16 QNN inference failed at graph load; its explicit TFLite CPU fallback returned 41 finite outputs, preserved CPU top-1 token 1867 and had logits relative L2 1.26e-6 versus ONNX CPU, with 100 CPU-requested component samples at p50/p95 59.518/71.347 ms and all 1,008 execution-detail rows on CPU. These are speech-head component results, not multimodal MiniCPM-o requests.
A [raw sustained-window audit](../../../e2e_expansion_20260924/evidence/qwen_tts_sustained_drift/README.md) found that native Windows RTX Qwen3-TTS playback slows sharply after about 26 minutes: all 17 serial medium requests in minutes 27–30 underrun in the simulated first-audio playout, with median RTF 1.331. The separate WSL run stayed below real time without underruns. The cause is not established; a timestamped GPU telemetry repeat is underway. A [raw sustained-window audit](../../../e2e_expansion_20260924/evidence/qwen_tts_sustained_drift/README.md) found that native Windows RTX Qwen3-TTS playback slows sharply after about 26 minutes: all 17 serial medium requests in minutes 27–30 underrun in the simulated first-audio playout, with median RTF 1.331. The separate WSL run stayed below real time without underruns. The cause is not established; a timestamped GPU telemetry repeat is underway.
@@ -43,11 +43,11 @@ NOT E2E means no complete device/model workload was verified. “Scoped E2E” i
| PC HX370 + Radeon 890M / Windows worker | scoped E2E: Omni Spark text | scoped E2E: Omni Radeon talker/codec + CPU predictor | NOT E2E | scoped E2E: Omni Radeon+CPU image+audio → text+speech (native-storage 20-run profile) | synthetic policy pass: Omni CPU+Radeon Cosmos (WSL/native Windows) | | PC HX370 + Radeon 890M / Windows worker | scoped E2E: Omni Spark text | scoped E2E: Omni Radeon talker/codec + CPU predictor | NOT E2E | scoped E2E: Omni Radeon+CPU image+audio → text+speech (native-storage 20-run profile) | synthetic policy pass: Omni CPU+Radeon Cosmos (WSL/native Windows) |
| PC HX370 + AMD NPU / Windows worker | NOT E2E (output-head component passed) | NOT E2E | NOT E2E | NOT E2E | NOT E2E | | PC HX370 + AMD NPU / Windows worker | NOT E2E (output-head component passed) | NOT E2E | NOT E2E | NOT E2E | NOT E2E |
| PC CPU+iGPU+NPU +/- discrete GPU, joint model execution | NOT E2E | NOT E2E | NOT E2E | NOT E2E | NOT E2E | | PC CPU+iGPU+NPU +/- discrete GPU, joint model execution | NOT E2E | NOT E2E | NOT E2E | NOT E2E | NOT E2E |
| PC Snapdragon X Elite CRD / AI Hub | NOT E2E (one Spark attention layer on NPU) | NOT E2E (CPU talker/vocoder steps; QNN vocoder numeric) | NOT E2E | NOT E2E (NPU speech-head step profiled) | NOT E2E (Cosmos encoder numeric; placement pending) | | PC Snapdragon X Elite CRD / AI Hub | NOT E2E (one Spark attention layer on NPU) | NOT E2E (CPU talker/vocoder steps; QNN vocoder numeric) | NOT E2E | NOT E2E (NPU speech-head step profiled) | NOT E2E (Cosmos NPU encoder profiled) |
| Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | NOT E2E (one Spark attention layer on NPU) | NOT E2E (FP16 talker token mismatch) | NOT E2E | NOT E2E (NPU speech head step profiled) | NOT E2E (Cosmos NPU encoder profiled) | | Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | NOT E2E (one Spark attention layer on NPU) | NOT E2E (FP16 talker token mismatch) | NOT E2E | NOT E2E (NPU speech head step profiled) | NOT E2E (Cosmos NPU encoder profiled) |
| Mobile Galaxy S24 / Snapdragon 8 Gen 3 | NOT E2E (one Spark attention layer on NPU) | NOT E2E (FP16 talker token mismatch) | NOT E2E | NOT E2E (NPU speech head step profiled) | NOT E2E | | Mobile Galaxy S24 / Snapdragon 8 Gen 3 | NOT E2E (one Spark attention layer on NPU) | NOT E2E (FP16 talker token mismatch) | NOT E2E | NOT E2E (NPU speech head step profiled) | NOT E2E |
| Embedded SA8775P ADP / AI Hub | NOT E2E (one Spark attention layer on NPU) | NOT E2E (CPU talker/vocoder steps; NPU vocoder failed) | NOT E2E | NOT E2E (NPU speech-head step profiled) | NOT E2E (Cosmos compile; inference pending) | | Embedded SA8775P ADP / AI Hub | NOT E2E (one Spark attention layer on NPU) | NOT E2E (CPU talker/vocoder steps; NPU vocoder failed) | NOT E2E | NOT E2E (NPU speech-head step profiled) | NOT E2E (Cosmos encoder numeric; placement pending) |
| Embedded RB3 Gen 2 / QCS6490 / AI Hub | NOT E2E (Spark CPU layer passed; W8A8 NPU quality open) | NOT E2E (mixed vocoder numeric unqualified) | NOT E2E | NOT E2E (speech-head CPU numeric; placement pending) | NOT E2E | | Embedded RB3 Gen 2 / QCS6490 / AI Hub | NOT E2E (Spark CPU layer passed; W8A8 NPU quality open) | NOT E2E (mixed vocoder numeric unqualified) | NOT E2E | NOT E2E (CPU speech-head step profiled) | NOT E2E |
## Per-cell reasons and next steps ## Per-cell reasons and next steps
@@ -89,10 +89,10 @@ NOT E2E means no complete device/model workload was verified. “Scoped E2E” i
| PC CPU+iGPU+NPU +/- discrete GPU, joint model execution | MiniCPM-o 4.5 | NOT_E2E_PROFILEABLE | No qualifying same-model multi-accelerator deployment/artifact plan or overlap/recovery evidence. Individual routes are not a joint pipeline. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. | | PC CPU+iGPU+NPU +/- discrete GPU, joint model execution | MiniCPM-o 4.5 | NOT_E2E_PROFILEABLE | No qualifying same-model multi-accelerator deployment/artifact plan or overlap/recovery evidence. Individual routes are not a joint pipeline. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
| PC CPU+iGPU+NPU +/- discrete GPU, joint model execution | InternVLA-A1 | NOT_E2E_PROFILEABLE | No qualifying same-model multi-accelerator deployment/artifact plan or overlap/recovery evidence. Individual routes are not a joint pipeline. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. | | PC CPU+iGPU+NPU +/- discrete GPU, joint model execution | InternVLA-A1 | NOT_E2E_PROFILEABLE | No qualifying same-model multi-accelerator deployment/artifact plan or overlap/recovery evidence. Individual routes are not a joint pipeline. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
| PC Snapdragon X Elite CRD / AI Hub | Spark-X2.5 | COMPONENT_NPU_INFERENCE_PROFILED; NOT_E2E_PROFILEABLE | A real Spark-X2.5-1.7B one-layer W8A16 QDQ attention artifact, compiled as QNN DLC, ran on the exact hosted Snapdragon X Elite CRD (Windows 11 ARM64). Inference returned finite hidden/new-K/new-V tensors bitwise identical to the prior S24/S25 NPU outputs on the same fixture; the profile attributed all 73 execution-detail rows to NPU. Across 100 component samples, nearest-rank p50/p95 was 1.433/1.606 ms; the 1.346 ms estimated-time field is the minimum. Hidden/K/V relative L2 versus original FP32 ONNX CPU was 1.351%/0.377%/0.650%, without a next-token quality tolerance. There is no complete 28-layer prefill/decode/KV-ring/sampling loop, resident Omni controller or complete-request timing. | Qualify a complete X Elite resident artifact/controller with state, token quality, memory admission/cancel and sustained power; [X Elite NPU raw evidence](../../../e2e_expansion_20260923/evidence/spark_xelite_full_attention/README.md). | | PC Snapdragon X Elite CRD / AI Hub | Spark-X2.5 | COMPONENT_NPU_INFERENCE_PROFILED; NOT_E2E_PROFILEABLE | A real Spark-X2.5-1.7B one-layer W8A16 QDQ attention artifact, compiled as QNN DLC, ran on the exact hosted Snapdragon X Elite CRD (Windows 11 ARM64). Inference returned finite hidden/new-K/new-V tensors bitwise identical to the prior S24/S25 NPU outputs on the same fixture; the profile attributed all 73 execution-detail rows to NPU. Across 100 component samples, nearest-rank p50/p95 was 1.433/1.606 ms; the 1.346 ms estimated-time field is the minimum. Hidden/K/V relative L2 versus original FP32 ONNX CPU was 1.351%/0.377%/0.650%, without a next-token quality tolerance. There is no complete 28-layer prefill/decode/KV-ring/sampling loop, resident Omni controller or complete-request timing. | Qualify a complete X Elite resident artifact/controller with state, token quality, memory admission/cancel and sustained power; [X Elite NPU raw evidence](../../../e2e_expansion_20260923/evidence/spark_xelite_full_attention/README.md). |
| PC Snapdragon X Elite CRD / AI Hub | Qwen3-TTS 0.6B CustomVoice | COMPONENT_ONNX_CPU_VOCODER_PARITY_PROFILED; DIRECTML_RUNTIME_FAILED; TALKER_FP16_NUMERIC_FAIL; QNN_NPU_VOCODER_NUMERIC_MEASURED_PROFILE_TIMEOUT; TALKER_ONNX_CPU_NUMERIC_MEASURED; TALKER_CPU_PLACEMENT_PENDING; NOT_E2E_PROFILEABLE | The fresh ONNX vocoder inference on the exact Windows 11 ARM64 X Elite CRD requested CPU and returned an unsaturated waveform on the pinned fixture, matching local ONNX Runtime CPU with 8.09e-7 relative L2 and 121.84 dB SNR. Its 94-sample component profile had nearest-rank p50/p95 6.384/6.644 s and all 634 execution-detail rows on CPU; 5.917 s estimated time is the sample minimum. This is one component, without a listening-quality gate. Corrected DirectML inference and profile both failed at provider initialization with Windows error 80070057. A separate FP16 QNN DLC vocoder inference returned an unsaturated waveform at 1.138% relative L2 / 38.88 dB SNR versus local ONNX CPU on the same fixture; the NPU node-placement profile timed out during graph preparation with no timing or per-node samples, and no listening-quality tolerance is set. The 28-layer FP16 QNN talker step executed, yet changed CPU top-1 token 80 to 1720 with 8.57% logits relative L2 on a synthetic cache fixture; a separate ONNX CPU-requested talker step preserved CPU token 80 with logits relative L2 2.07e-5 and maximum K/V relative L2 2.18e-5; its placement profile is pending. The older TFLite vocoder target was rejected on Windows ARM; no complete stream exists. | Reduce or partition the QNN vocoder profile graph and audit CPU talker placement; investigate DirectML failure, then qualify complete local streaming, memory and sustained behavior; [vocoder](../../../e2e_expansion_20260924/evidence/qwen_tts_vocoder_qualcomm/README.md) and [talker](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) evidence. | | PC Snapdragon X Elite CRD / AI Hub | Qwen3-TTS 0.6B CustomVoice | COMPONENT_ONNX_CPU_VOCODER_PARITY_PROFILED; DIRECTML_RUNTIME_FAILED; TALKER_FP16_NUMERIC_FAIL; QNN_NPU_VOCODER_NUMERIC_MEASURED_PROFILE_TIMEOUT; TALKER_ONNX_CPU_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The fresh ONNX vocoder inference on the exact Windows 11 ARM64 X Elite CRD requested CPU and returned an unsaturated waveform on the pinned fixture, matching local ONNX Runtime CPU with 8.09e-7 relative L2 and 121.84 dB SNR. Its 94-sample component profile had nearest-rank p50/p95 6.384/6.644 s and all 634 execution-detail rows on CPU; 5.917 s estimated time is the sample minimum. This is one component, without a listening-quality gate. Corrected DirectML inference and profile both failed at provider initialization with Windows error 80070057. A separate FP16 QNN DLC vocoder inference returned an unsaturated waveform at 1.138% relative L2 / 38.88 dB SNR versus local ONNX CPU on the same fixture; the NPU node-placement profile timed out during graph preparation with no timing or per-node samples, and no listening-quality tolerance is set. The 28-layer FP16 QNN talker step executed, yet changed CPU top-1 token 80 to 1720 with 8.57% logits relative L2 on a synthetic cache fixture; a separate ONNX CPU-requested talker step preserved CPU token 80 with logits relative L2 2.07e-5 and maximum K/V relative L2 2.18e-5. Its 100-sample component profile had p50/p95 191.414/309.428 ms and all 1,744 execution-detail rows on CPU. The older TFLite vocoder target was rejected on Windows ARM; no complete stream exists. | Reduce or partition the QNN vocoder profile graph and audit CPU talker placement; investigate DirectML failure, then qualify complete local streaming, memory and sustained behavior; [vocoder](../../../e2e_expansion_20260924/evidence/qwen_tts_vocoder_qualcomm/README.md) and [talker](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) evidence. |
| PC Snapdragon X Elite CRD / AI Hub | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. | | PC Snapdragon X Elite CRD / AI Hub | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
| PC Snapdragon X Elite CRD / AI Hub | MiniCPM-o 4.5 | COMPONENT_SPEECH_HEAD_NPU_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The real 20-layer MiniCPM-o speech-head FP16 QNN DLC ran on the exact X Elite CRD. Same-fixture inference returned 41 finite tensors, preserved ONNX CPU top-1 speech token 1867, and had 0.3170% logits relative L2 and 1.221% maximum K/V relative L2. Its 100-sample component profile had nearest-rank p50/p95 9.905/11.037 ms, with all 944 execution-detail rows on NPU. Reported inference peak memory was 79,208,448 bytes and first-load peak 1,081,573,376 bytes. Thinker/encoders/vocoder and a device-local stateful Omni backend remain unqualified. | Validate the complete multimodal pipeline, state, quality, co-resident memory and sustained behavior; [X Elite raw evidence](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md). | | PC Snapdragon X Elite CRD / AI Hub | MiniCPM-o 4.5 | COMPONENT_SPEECH_HEAD_NPU_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The real 20-layer MiniCPM-o speech-head FP16 QNN DLC ran on the exact X Elite CRD. Same-fixture inference returned 41 finite tensors, preserved ONNX CPU top-1 speech token 1867, and had 0.3170% logits relative L2 and 1.221% maximum K/V relative L2. Its 100-sample component profile had nearest-rank p50/p95 9.905/11.037 ms, with all 944 execution-detail rows on NPU. Reported inference peak memory was 79,208,448 bytes and first-load peak 1,081,573,376 bytes. Thinker/encoders/vocoder and a device-local stateful Omni backend remain unqualified. | Validate the complete multimodal pipeline, state, quality, co-resident memory and sustained behavior; [X Elite raw evidence](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md). |
| PC Snapdragon X Elite CRD / AI Hub | InternVLA-A1 | COMPONENT_COSMOS_NUMERIC_MEASURED; PLACEMENT_PENDING; NOT_E2E_PROFILEABLE | The real Place_Markpen Cosmos FP16 QNN image encoder returned a finite FP32 latent for the six-frame synthetic pattern on exact X Elite CRD. Against the pinned source, relative L2 was 0.0608%, SNR 64.32 dB and cosine 0.999999815. The NPU-requested route awaits actual node-placement/timing profile. No real-observation action tolerance, complete policy or local stateful Omni backend exists. | Audit the [X Elite Cosmos placement profile](../../../e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md), then validate real observations, action units/timing, state, admission, cancellation and sustained behavior. | | PC Snapdragon X Elite CRD / AI Hub | InternVLA-A1 | COMPONENT_COSMOS_NPU_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The real Place_Markpen Cosmos FP16 QNN image encoder returned a finite FP32 latent for the six-frame synthetic pattern on exact X Elite CRD. Against the pinned source, relative L2 was 0.0608%, SNR 64.32 dB and cosine 0.999999815. The 100-sample six-frame encoder profile had p50/p95 82.939/83.837 ms, with all 121 execution-detail rows on NPU. Reported component inference peak memory was 24,059,904 bytes and first-load peak 692,891,648 bytes. No real-observation action tolerance, complete policy or local stateful Omni backend exists. | Validate real observations, action units/timing, state, admission, cancellation and sustained behavior; [X Elite Cosmos raw evidence](../../../e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md). |
| Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | Spark-X2.5 | COMPONENT_NPU_INFERENCE_PROFILED; NOT_E2E_PROFILEABLE | The real 1.7B checkpoint one-layer W8A16 full-attention decoder artifact executed for a fixed 1,024-token cache bucket. Same-fixture outputs matched S24; the hosted S25 profile had 100 component samples at nearest-rank p50/p95 0.957/1.028 ms with all 73 execution-detail rows on NPU. FP32-source hidden-state relative L2 was 1.351% without a downstream token-quality gate. No full 28-layer prefill/decode, KV/ring, sampling or resident controller ran on S25. | Implement and validate local 128-token generation, 512/1024 transitions, cancellation, resident memory and sustained power; [S25/S24 component evidence](../../../e2e_expansion_20260923/evidence/spark_s24_full_attention/README.md). | | Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | Spark-X2.5 | COMPONENT_NPU_INFERENCE_PROFILED; NOT_E2E_PROFILEABLE | The real 1.7B checkpoint one-layer W8A16 full-attention decoder artifact executed for a fixed 1,024-token cache bucket. Same-fixture outputs matched S24; the hosted S25 profile had 100 component samples at nearest-rank p50/p95 0.957/1.028 ms with all 73 execution-detail rows on NPU. FP32-source hidden-state relative L2 was 1.351% without a downstream token-quality gate. No full 28-layer prefill/decode, KV/ring, sampling or resident controller ran on S25. | Implement and validate local 128-token generation, 512/1024 transitions, cancellation, resident memory and sustained power; [S25/S24 component evidence](../../../e2e_expansion_20260923/evidence/spark_s24_full_attention/README.md). |
| Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | Qwen3-TTS 0.6B CustomVoice | TALKER_FP16_NUMERIC_FAIL; NOT_E2E_PROFILEABLE | Fresh FP16 predictor inference passed; its one output exactly matches the same compiled artifact/dataset reference. The real 28-layer FP16 QNN talker step executed on the synthetic cache fixture, but changed CPU top-1 speech token 80 to 412 with 8.99% logits relative L2 and maximum new-K/V relative L2 11.71%. This tested talker artifact is numerically unqualified; no complete co-resident NPU/GPU stream exists. | Improve and revalidate the [talker artifact](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md), then qualify the device-local predictor/vocoder pipeline, output quality, recovery and 30-minute thermal behavior. | | Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | Qwen3-TTS 0.6B CustomVoice | TALKER_FP16_NUMERIC_FAIL; NOT_E2E_PROFILEABLE | Fresh FP16 predictor inference passed; its one output exactly matches the same compiled artifact/dataset reference. The real 28-layer FP16 QNN talker step executed on the synthetic cache fixture, but changed CPU top-1 speech token 80 to 412 with 8.99% logits relative L2 and maximum new-K/V relative L2 11.71%. This tested talker artifact is numerically unqualified; no complete co-resident NPU/GPU stream exists. | Improve and revalidate the [talker artifact](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md), then qualify the device-local predictor/vocoder pipeline, output quality, recovery and 30-minute thermal behavior. |
| Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. | | Mobile Galaxy S25 / Snapdragon 8 Elite for Galaxy | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
@@ -104,14 +104,14 @@ NOT E2E means no complete device/model workload was verified. “Scoped E2E” i
| Mobile Galaxy S24 / Snapdragon 8 Gen 3 | MiniCPM-o 4.5 | COMPONENT_FP16_SPEECH_HEAD_NPU_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The retained FP16 speech-head job used a dataset with identical tensors to the recovered S25 fixture. Against original ONNX CPU, logits relative L2 was 0.3459%, maximum new-K/V relative L2 1.235%, and top-1 speech token 1867 matched. Earlier parity reported 49.22 dB SNR/top-1 match; tested W8A16 top-1 failed. The 100-sample component profile had p50/p95 8.537/9.006 ms, with all 1,029 execution-detail rows on NPU and reported inference peak memory 165,433,344 bytes. Generated-audio quality and the rest of the model remain unqualified. | Qualify thinker/encoders/vocoder and stateful local integration; [speech-head evidence](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md). | | Mobile Galaxy S24 / Snapdragon 8 Gen 3 | MiniCPM-o 4.5 | COMPONENT_FP16_SPEECH_HEAD_NPU_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The retained FP16 speech-head job used a dataset with identical tensors to the recovered S25 fixture. Against original ONNX CPU, logits relative L2 was 0.3459%, maximum new-K/V relative L2 1.235%, and top-1 speech token 1867 matched. Earlier parity reported 49.22 dB SNR/top-1 match; tested W8A16 top-1 failed. The 100-sample component profile had p50/p95 8.537/9.006 ms, with all 1,029 execution-detail rows on NPU and reported inference peak memory 165,433,344 bytes. Generated-audio quality and the rest of the model remain unqualified. | Qualify thinker/encoders/vocoder and stateful local integration; [speech-head evidence](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md). |
| Mobile Galaxy S24 / Snapdragon 8 Gen 3 | InternVLA-A1 | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. | | Mobile Galaxy S24 / Snapdragon 8 Gen 3 | InternVLA-A1 | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
| Embedded SA8775P ADP / AI Hub | Spark-X2.5 | COMPONENT_NPU_INFERENCE_PROFILED; NOT_E2E_PROFILEABLE | A real Spark-X2.5-1.7B one-layer W8A16 QDQ attention artifact, compiled as QNN DLC, ran on the exact hosted SA8775P ADP (Android 14). Inference returned finite hidden/new-K/new-V tensors bitwise identical to the prior S24/S25 NPU outputs on the same fixture; the profile attributed all 73 execution-detail rows to NPU. Across 100 component samples, nearest-rank p50/p95 was 1.837/2.079 ms; the 1.749 ms estimated-time field is the minimum. Hidden/K/V relative L2 versus original FP32 ONNX CPU was 1.351%/0.377%/0.650%, without a next-token quality tolerance. There is no complete 28-layer prefill/decode/KV-ring/sampling loop, resident Omni controller or complete-request timing. | Qualify a complete embedded resident artifact/controller with state, token quality, memory admission/cancel and sustained power; [SA8775P NPU raw evidence](../../../e2e_expansion_20260923/evidence/spark_sa8775_full_attention/README.md). | | Embedded SA8775P ADP / AI Hub | Spark-X2.5 | COMPONENT_NPU_INFERENCE_PROFILED; NOT_E2E_PROFILEABLE | A real Spark-X2.5-1.7B one-layer W8A16 QDQ attention artifact, compiled as QNN DLC, ran on the exact hosted SA8775P ADP (Android 14). Inference returned finite hidden/new-K/new-V tensors bitwise identical to the prior S24/S25 NPU outputs on the same fixture; the profile attributed all 73 execution-detail rows to NPU. Across 100 component samples, nearest-rank p50/p95 was 1.837/2.079 ms; the 1.749 ms estimated-time field is the minimum. Hidden/K/V relative L2 versus original FP32 ONNX CPU was 1.351%/0.377%/0.650%, without a next-token quality tolerance. There is no complete 28-layer prefill/decode/KV-ring/sampling loop, resident Omni controller or complete-request timing. | Qualify a complete embedded resident artifact/controller with state, token quality, memory admission/cancel and sustained power; [SA8775P NPU raw evidence](../../../e2e_expansion_20260923/evidence/spark_sa8775_full_attention/README.md). |
| Embedded SA8775P ADP / AI Hub | Qwen3-TTS 0.6B CustomVoice | COMPONENT_CPU_VOCODER_PARITY_AND_PLACEMENT_PASS; TALKER_TFLITE_CPU_NUMERIC_MEASURED; QNN_NPU_VOCODER_INFERENCE_FAILED; COMPONENT_GPU_NUMERIC_FAIL; TALKER_FP16_NUMERIC_FAIL; NOT_E2E_PROFILEABLE | On the exact SA8775P, the same TFLite binary's CPU-requested Code2Wav inference matched original ONNX CPU within 4.13e-6 relative L2 / 107.67 dB SNR, without saturation. Its 100-sample CPU-requested component profile had p50/p95 5.384/5.534 s with all 738 execution-detail rows on CPU; the service estimated-time field 5.250 s is the sample minimum. The GPU-requested output saturated 96.63% of samples with 11.832 relative L2; its fresh 100-sample mixed CPU/GPU profile p50/p95 was 1.776/1.880 s, timing failed audio. Even with FP32-preserving GPU options, 84.89% of samples saturated with 11.389 relative L2. The 28-layer FP16 QNN talker step executed but changed CPU top-1 token 80 to 1720 with 9.43% logits relative L2. A separate TFLite CPU-requested talker step preserved CPU top-1 token 80 with logits relative L2 3.33e-5 and maximum K/V relative L2 3.41e-5; its placement profile is pending. A separate FP16 QNN DLC vocoder compile passed, but same-fixture NPU-requested inference failed after compilation without a waveform or placement samples. No complete pipeline exists. | Audit talker CPU placement and isolate the FP16 QNN vocoder runtime failure, GPU delegate failure and FP16 talker numerical drift, then establish listening quality, co-resident streaming, memory and sustained behavior; [vocoder](../../../e2e_expansion_20260924/evidence/qwen_tts_vocoder_qualcomm/README.md) and [talker](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) evidence. | | Embedded SA8775P ADP / AI Hub | Qwen3-TTS 0.6B CustomVoice | COMPONENT_CPU_VOCODER_PARITY_AND_PLACEMENT_PASS; TALKER_TFLITE_CPU_NUMERIC_PROFILED; QNN_NPU_VOCODER_INFERENCE_FAILED; COMPONENT_GPU_NUMERIC_FAIL; TALKER_FP16_NUMERIC_FAIL; NOT_E2E_PROFILEABLE | On the exact SA8775P, the same TFLite binary's CPU-requested Code2Wav inference matched original ONNX CPU within 4.13e-6 relative L2 / 107.67 dB SNR, without saturation. Its 100-sample CPU-requested component profile had p50/p95 5.384/5.534 s with all 738 execution-detail rows on CPU; the service estimated-time field 5.250 s is the sample minimum. The GPU-requested output saturated 96.63% of samples with 11.832 relative L2; its fresh 100-sample mixed CPU/GPU profile p50/p95 was 1.776/1.880 s, timing failed audio. Even with FP32-preserving GPU options, 84.89% of samples saturated with 11.389 relative L2. The 28-layer FP16 QNN talker step executed but changed CPU top-1 token 80 to 1720 with 9.43% logits relative L2. A separate TFLite CPU-requested talker step preserved CPU top-1 token 80 with logits relative L2 3.33e-5 and maximum K/V relative L2 3.41e-5. Its 100-sample component profile had p50/p95 52.635/76.925 ms with all 1,800 execution-detail rows on CPU. A separate FP16 QNN DLC vocoder compile passed, but same-fixture NPU-requested inference failed after compilation without a waveform or placement samples. No complete pipeline exists. | Isolate the FP16 QNN vocoder runtime failure, GPU delegate failure and FP16 talker numerical drift, then establish listening quality, co-resident streaming, memory and sustained behavior; [vocoder](../../../e2e_expansion_20260924/evidence/qwen_tts_vocoder_qualcomm/README.md) and [talker](../../../e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) evidence. |
| Embedded SA8775P ADP / AI Hub | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. | | Embedded SA8775P ADP / AI Hub | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
| Embedded SA8775P ADP / AI Hub | MiniCPM-o 4.5 | COMPONENT_SPEECH_HEAD_NPU_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The real 20-layer MiniCPM-o speech-head FP16 QNN DLC ran on the exact SA8775P ADP. Same-fixture inference returned 41 finite tensors, preserved ONNX CPU top-1 token 1867, and had 0.3568% logits relative L2 and 1.225% maximum K/V relative L2. Its 100-sample component profile had nearest-rank p50/p95 11.901/13.038 ms, with all 944 execution-detail rows on NPU. Reported inference peak memory was 134,696,960 bytes and first-load peak 1,322,835,968 bytes; no complete thinker/encoders/vocoder bundle or device-local stateful controller exists. | Validate complete multimodal generation, state, speech quality, co-resident memory, cancellation and sustained power; [SA8775P raw evidence](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md). | | Embedded SA8775P ADP / AI Hub | MiniCPM-o 4.5 | COMPONENT_SPEECH_HEAD_NPU_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The real 20-layer MiniCPM-o speech-head FP16 QNN DLC ran on the exact SA8775P ADP. Same-fixture inference returned 41 finite tensors, preserved ONNX CPU top-1 token 1867, and had 0.3568% logits relative L2 and 1.225% maximum K/V relative L2. Its 100-sample component profile had nearest-rank p50/p95 11.901/13.038 ms, with all 944 execution-detail rows on NPU. Reported inference peak memory was 134,696,960 bytes and first-load peak 1,322,835,968 bytes; no complete thinker/encoders/vocoder bundle or device-local stateful controller exists. | Validate complete multimodal generation, state, speech quality, co-resident memory, cancellation and sustained power; [SA8775P raw evidence](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md). |
| Embedded SA8775P ADP / AI Hub | InternVLA-A1 | COMPONENT_COSMOS_FP16_QNN_COMPILE_PASS_INFERENCE_PENDING; NOT_E2E_PROFILEABLE | The real Place_Markpen Cosmos image encoder source compiled to FP16 QNN DLC on exact SA8775P ADP. Same-fixture NPU-requested inference is pending; no numerical output, actual placement, full action policy or local stateful Omni backend exists yet. | Audit the [SA8775P Cosmos inference](../../../e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md), then obtain placement, real-observation action validation and a complete local policy. | | Embedded SA8775P ADP / AI Hub | InternVLA-A1 | COMPONENT_COSMOS_NUMERIC_MEASURED; PLACEMENT_PENDING; NOT_E2E_PROFILEABLE | The real Place_Markpen Cosmos FP16 QNN image encoder returned a finite FP32 latent for the six-frame synthetic pattern on exact SA8775P ADP. Against the pinned source, relative L2 was 0.0613%, SNR 64.25 dB and cosine 0.999999812. The NPU-requested route awaits actual node-placement/timing profile. No real-observation action tolerance, complete policy or local stateful Omni backend exists. | Audit the [SA8775P Cosmos placement profile](../../../e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md), then validate real observations, action units/timing, state, admission, cancellation and sustained behavior. |
| Embedded RB3 Gen 2 / QCS6490 / AI Hub | Spark-X2.5 | COMPONENT_CPU_PARITY_PASS; COMPONENT_NPU_NUMERIC_UNQUALIFIED; NOT_E2E_PROFILEABLE | On the exact hosted Dragonwing RB3 Gen 2 Vision Kit (QCS6490, QC Linux 1.6), the reused S24 W8A16 QNN DLC and a separately RB3-compiled W8A16 DLC each failed inference/profile at QnnModel_composeGraphsFromDlc: MODEL_GRAPH_ERROR. An FP32 ONNX one-layer CPU route passed with all 56 profile rows on CPU, source-relative hidden/K/V L2 of 8.73e-7/2.98e-7/3.37e-7 and 100-sample component p50/p95 19.304/21.113 ms. A calibrated W8A8 QNN DLC ran with all 61 rows on NPU and p50/p95 5.610/5.761 ms, but hidden/K/V relative L2 was 20.914%/1.702%/4.413% versus FP32 CPU, without a next-token quality gate. These are different precisions, not a qualified speedup; no complete 28-layer prefill/decode/KV-ring/sampling, local Omni controller or complete-request timing exists. | Improve representative W8A8 calibration or find another compatible quality-passing artifact; validate token parity, then full device-resident generation, admission/cancel and sustained power; [RB3 raw evidence](../../../e2e_expansion_20260923/evidence/spark_rb3_full_attention/README.md). | | Embedded RB3 Gen 2 / QCS6490 / AI Hub | Spark-X2.5 | COMPONENT_CPU_PARITY_PASS; COMPONENT_NPU_NUMERIC_UNQUALIFIED; NOT_E2E_PROFILEABLE | On the exact hosted Dragonwing RB3 Gen 2 Vision Kit (QCS6490, QC Linux 1.6), the reused S24 W8A16 QNN DLC and a separately RB3-compiled W8A16 DLC each failed inference/profile at QnnModel_composeGraphsFromDlc: MODEL_GRAPH_ERROR. An FP32 ONNX one-layer CPU route passed with all 56 profile rows on CPU, source-relative hidden/K/V L2 of 8.73e-7/2.98e-7/3.37e-7 and 100-sample component p50/p95 19.304/21.113 ms. A calibrated W8A8 QNN DLC ran with all 61 rows on NPU and p50/p95 5.610/5.761 ms, but hidden/K/V relative L2 was 20.914%/1.702%/4.413% versus FP32 CPU, without a next-token quality gate. These are different precisions, not a qualified speedup; no complete 28-layer prefill/decode/KV-ring/sampling, local Omni controller or complete-request timing exists. | Improve representative W8A8 calibration or find another compatible quality-passing artifact; validate token parity, then full device-resident generation, admission/cancel and sustained power; [RB3 raw evidence](../../../e2e_expansion_20260923/evidence/spark_rb3_full_attention/README.md). |
| Embedded RB3 Gen 2 / QCS6490 / AI Hub | Qwen3-TTS 0.6B CustomVoice | COMPONENT_MIXED_VOCODER_PROFILED_NUMERIC_UNQUALIFIED; NOT_E2E_PROFILEABLE | Fresh same-fixture TFLite GPU-requested vocoder inference returned a finite unsaturated waveform, but relative L2 versus original ONNX CPU was 15.73% (16.06 dB SNR), without a listening-quality tolerance; it is not a qualified audio path. The fresh 100-sample component profile had nearest-rank p50/p95 3.099/3.133 s and 403 CPU plus 335 GPU execution-detail rows. Its 3.070 s estimated-time field is the sample minimum, not median. The historical NPU talker build was rejected for floating-point inputs; no full model pipeline exists. | Establish waveform quality and a compatible integer talker artifact, then validate complete local execution, memory and sustained behavior; [vocoder evidence](../../../e2e_expansion_20260924/evidence/qwen_tts_vocoder_qualcomm/README.md). | | Embedded RB3 Gen 2 / QCS6490 / AI Hub | Qwen3-TTS 0.6B CustomVoice | COMPONENT_MIXED_VOCODER_PROFILED_NUMERIC_UNQUALIFIED; NOT_E2E_PROFILEABLE | Fresh same-fixture TFLite GPU-requested vocoder inference returned a finite unsaturated waveform, but relative L2 versus original ONNX CPU was 15.73% (16.06 dB SNR), without a listening-quality tolerance; it is not a qualified audio path. The fresh 100-sample component profile had nearest-rank p50/p95 3.099/3.133 s and 403 CPU plus 335 GPU execution-detail rows. Its 3.070 s estimated-time field is the sample minimum, not median. The historical NPU talker build was rejected for floating-point inputs; no full model pipeline exists. | Establish waveform quality and a compatible integer talker artifact, then validate complete local execution, memory and sustained behavior; [vocoder evidence](../../../e2e_expansion_20260924/evidence/qwen_tts_vocoder_qualcomm/README.md). |
| Embedded RB3 Gen 2 / QCS6490 / AI Hub | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. | | Embedded RB3 Gen 2 / QCS6490 / AI Hub | Qwen3.8-27B | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
| Embedded RB3 Gen 2 / QCS6490 / AI Hub | MiniCPM-o 4.5 | COMPONENT_SPEECH_HEAD_FP16_QNN_LOAD_FAILED; CPU_FALLBACK_NUMERIC_MEASURED_PLACEMENT_PENDING; NOT_E2E_PROFILEABLE | The real 20-layer MiniCPM-o speech-head FP16 QNN DLC compiled for exact RB3 Gen 2, but same-fixture NPU-requested inference failed at graph load with QnnModel_composeGraphsFromDlc: MODEL_GRAPH_ERROR. RB3 device attributes do not advertise FP16 HTP support. An explicit TFLite CPU-requested fallback returned 41 finite tensors, preserved ONNX CPU top-1 token 1867, and had logits relative L2 1.26e-6 and maximum K/V relative L2 1.58e-6. Actual CPU placement and timing profile is pending; no complete model bundle or local stateful controller exists. | Audit the [RB3 speech-head CPU placement profile](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md), then qualify complete multimodal generation and sustained behavior. | | Embedded RB3 Gen 2 / QCS6490 / AI Hub | MiniCPM-o 4.5 | COMPONENT_SPEECH_HEAD_FP16_QNN_LOAD_FAILED; CPU_FALLBACK_NUMERIC_PROFILED; NOT_E2E_PROFILEABLE | The real 20-layer MiniCPM-o speech-head FP16 QNN DLC compiled for exact RB3 Gen 2, but same-fixture NPU-requested inference failed at graph load with QnnModel_composeGraphsFromDlc: MODEL_GRAPH_ERROR. RB3 device attributes do not advertise FP16 HTP support. An explicit TFLite CPU-requested fallback returned 41 finite tensors, preserved ONNX CPU top-1 token 1867, and had logits relative L2 1.26e-6 and maximum K/V relative L2 1.58e-6. Its 100-sample component profile had p50/p95 59.518/71.347 ms with all 1,008 execution-detail rows on CPU; reported inference peak memory was 820,981,760 bytes. No complete model bundle or local stateful controller exists. | Qualify complete multimodal generation and sustained behavior; [RB3 speech-head CPU evidence](../../../e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md). |
| Embedded RB3 Gen 2 / QCS6490 / AI Hub | InternVLA-A1 | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. | | Embedded RB3 Gen 2 / QCS6490 / AI Hub | InternVLA-A1 | NOT_E2E_PROFILEABLE | No complete qualified target artifact and device-local stateful Omni backend. Hub jobs provide component execution, not an on-device generation/streaming controller; ADB has no attached device. | Supply and validate the missing matching artifact/backend, then run the complete specified workload. |
## Measured request timing ## Measured request timing
+3 -3
View File
@@ -19,13 +19,13 @@ The [Qwen3-TTS Code2Wav Qualcomm follow-up](../../benchmarks/edge_harness/result
The identical SA/RB3 TFLite vocoder binary also matched original ONNX CPU within 3.45e-6 relative L2 on a separate local HX370 x86-64 LiteRT 2.2.0 CPU control. Exact-SA CPU-requested parity narrows the failure to the GPU-requested route. The FP32-preserving GPU option did not fix it, so per-node placement and delegate-operation isolation remain open. The identical SA/RB3 TFLite vocoder binary also matched original ONNX CPU within 3.45e-6 relative L2 on a separate local HX370 x86-64 LiteRT 2.2.0 CPU control. Exact-SA CPU-requested parity narrows the failure to the GPU-requested route. The FP32-preserving GPU option did not fix it, so per-node placement and delegate-operation isolation remain open.
A separate [28-layer talker decode-step attempt](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) retains a synthetic fixed-cache fixture and CPU reference. Historical FP16 QNN context-binary inference executed on S24, S25, SA8775P and X Elite, but all four changed the CPU top-1 speech token with 8.57–9.43% logits relative L2. RB3 rejected that floating-point-input artifact. Explicit X Elite ONNX CPU and SA8775P TFLite CPU fallback compiles passed. X Elite CPU-requested inference preserved CPU top-1 token 80 with logits relative L2 2.07e-5; its placement profile remains pending. SA8775P TFLite CPU-requested talker inference also preserved token 80 with logits relative L2 3.33e-5; its placement profile is pending. None advances M2 to numerical talker or full-stream qualification yet. A separate [28-layer talker decode-step attempt](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/qwen_tts_talker_qualcomm/README.md) retains a synthetic fixed-cache fixture and CPU reference. Historical FP16 QNN context-binary inference executed on S24, S25, SA8775P and X Elite, but all four changed the CPU top-1 speech token with 8.57–9.43% logits relative L2. RB3 rejected that floating-point-input artifact. Explicit X Elite ONNX CPU and SA8775P TFLite CPU fallback compiles passed. X Elite CPU-requested inference preserved CPU top-1 token 80 with logits relative L2 2.07e-5; its 100-sample decode-step profile measured p50/p95 191.414/309.428 ms with all 1,744 execution-detail rows on CPU. SA8775P TFLite CPU-requested talker inference also preserved token 80 with logits relative L2 3.33e-5; its 100-sample decode-step profile measured p50/p95 52.635/76.925 ms with all 1,800 execution-detail rows on CPU. None advances M2 to numerical talker or full-stream qualification yet.
For M4, a [MiniCPM-o 4.5 speech-head study](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md) recovered the historical 20-layer ONNX source and matching S24/S25 fixtures. Both FP16 context-binary outputs preserved the ONNX CPU top-1 speech token on one decode step, with logits relative L2 0.346%/0.360%. Their 100-sample component profiles had p50/p95 8.537/9.006 ms and 7.197/7.615 ms, with all 1,029 execution-detail rows on NPU for each. New FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P same-fixture inference preserved CPU top-1 speech token 1867 with 0.3568% logits relative L2 and 1.225% maximum K/V relative L2; its 100-sample component profile measured p50/p95 11.901/13.038 ms with all 944 execution-detail rows on NPU. X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2; its 100-sample component profile measured p50/p95 9.905/11.037 ms with all 944 execution-detail rows on NPU. RB3 FP16 QNN inference failed at graph load despite compile success; a TFLite CPU fallback same-fixture inference preserved top-1 token 1867 with logits relative L2 1.26e-6 versus ONNX CPU; CPU placement profiling is pending. This is component C numerical evidence only, not thinker/encoder/vocoder or complete multimodal support. For M4, a [MiniCPM-o 4.5 speech-head study](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/minicpmo_tts_qualcomm/README.md) recovered the historical 20-layer ONNX source and matching S24/S25 fixtures. Both FP16 context-binary outputs preserved the ONNX CPU top-1 speech token on one decode step, with logits relative L2 0.346%/0.360%. Their 100-sample component profiles had p50/p95 8.537/9.006 ms and 7.197/7.615 ms, with all 1,029 execution-detail rows on NPU for each. New FP16 QNN DLC compiles passed for SA8775P, X Elite and RB3. SA8775P same-fixture inference preserved CPU top-1 speech token 1867 with 0.3568% logits relative L2 and 1.225% maximum K/V relative L2; its 100-sample component profile measured p50/p95 11.901/13.038 ms with all 944 execution-detail rows on NPU. X Elite same-fixture inference also preserved token 1867 with 0.3170% logits relative L2; its 100-sample component profile measured p50/p95 9.905/11.037 ms with all 944 execution-detail rows on NPU. RB3 FP16 QNN inference failed at graph load despite compile success; a TFLite CPU fallback same-fixture inference preserved top-1 token 1867 with logits relative L2 1.26e-6 versus ONNX CPU; its 100-sample CPU-requested profile measured p50/p95 59.518/71.347 ms with all 1,008 execution-detail rows on CPU. This is component C numerical evidence only, not thinker/encoder/vocoder or complete multimodal support.
The [native Windows Qwen3-TTS sustained audit](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/qwen_tts_sustained_drift/README.md) identifies an additional M2 desktop blocker. The historical RTX 5090 Laptop complete-model stream changed regime around minute 26–27; all 17 requests in minutes 27–30 had simulated playback underruns and median RTF 1.331. The separate WSL run had no sustained underruns. Existing logs do not identify the cause; timestamped GPU clocks, power, temperature and throttling are now available in the local profiler for a controlled repeat. Native Windows sustained playback remains unqualified. The [native Windows Qwen3-TTS sustained audit](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/qwen_tts_sustained_drift/README.md) identifies an additional M2 desktop blocker. The historical RTX 5090 Laptop complete-model stream changed regime around minute 26–27; all 17 requests in minutes 27–30 had simulated playback underruns and median RTF 1.331. The separate WSL run had no sustained underruns. Existing logs do not identify the cause; timestamped GPU clocks, power, temperature and throttling are now available in the local profiler for a controlled repeat. Native Windows sustained playback remains unqualified.
For M4, a [Galaxy S25 Cosmos encoder attempt](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md) repackages the real Place_Markpen FP32 ONNX export after retaining two CPU parity fixtures. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one synthetic pattern. Injecting that output into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. S25 six-frame component profile returned 100 samples at p50/p95 49.665/51.382 ms, with all 121 execution-detail rows on NPU; the action policy did not run on S25. The same encoder compiled to FP16 QNN DLC for X Elite CRD; same-fixture NPU-requested inference returned a finite latent at 0.0608% relative L2 versus source, with placement profiling pending. An SA8775P FP16 QNN DLC compile passed, with same-fixture NPU-requested inference pending. This is component/action-sensitivity evidence, not device-local M4 qualification. The checkpoint train config names a2d_real/a2d_pen_holder; an [authenticated access check](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/internvla_real_observation_access/README.md) returned HTTP 404. The official vLLM-Omni example names an InternData-A1 Place_Markpen archive, but that dataset returned HTTP 403 gated-access denial under the current account, so real-observation reference-action quality is still unmeasured. For M4, a [Galaxy S25 Cosmos encoder attempt](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/internvla_cosmos_qualcomm/README.md) repackages the real Place_Markpen FP32 ONNX export after retaining two CPU parity fixtures. FLOAT16 QNN DLC inference returned a latent within 0.0524% relative L2 of the source on one synthetic pattern. Injecting that output into the real policy on local CPU changed synthetic actions by 0.742% relative L2, without a task tolerance. S25 six-frame component profile returned 100 samples at p50/p95 49.665/51.382 ms, with all 121 execution-detail rows on NPU; the action policy did not run on S25. The same encoder compiled to FP16 QNN DLC for X Elite CRD; same-fixture NPU-requested inference returned a finite latent at 0.0608% relative L2 versus source, with its 100-sample six-frame profile at p50/p95 82.939/83.837 ms and all 121 execution-detail rows on NPU. An SA8775P FP16 QNN DLC inference returned a finite same-fixture latent at 0.0613% relative L2 versus source; placement profiling is pending. This is component/action-sensitivity evidence, not device-local M4 qualification. The checkpoint train config names a2d_real/a2d_pen_holder; an [authenticated access check](../../benchmarks/edge_harness/results/e2e_expansion_20260924/evidence/internvla_real_observation_access/README.md) returned HTTP 404. The official vLLM-Omni example names an InternData-A1 Place_Markpen archive, but that dataset returned HTTP 403 gated-access denial under the current account, so real-observation reference-action quality is still unmeasured.
The [WSL CPU expansion](../../benchmarks/edge_harness/results/e2e_expansion_20260923/README.md) The [WSL CPU expansion](../../benchmarks/edge_harness/results/e2e_expansion_20260923/README.md)
adds constrained three-stage MiniCPM-o text-to-speech requests and a CPU-only adds constrained three-stage MiniCPM-o text-to-speech requests and a CPU-only