Document Windows MiniCPM-o image and audio end-to-end runs

docs/design/omni_edge_e2e_roadmap.md

@@ -33,7 +33,10 @@ previous 0.58 budget failed the 2,048-token KV admission gate; the successful
 run still force-killed stage 0 during shutdown and reported a shared-memory
 cleanup warning. A separate 2 s, 440 Hz synthetic tone now traverses audio
 input to text+speech and yielded a plausible whistling description, also with
-a shared-memory cleanup warning. Broader image/audio quality, video and clean
+a shared-memory cleanup warning. Native Windows RTX also completed these two
+separate synthetic image and tone inputs through all three stages, returning
+the same respective text and thinker token IDs as WSL with nonzero WAVs.
+Broader image/audio quality, video, repeated multimodal latency and clean
 lifecycle remain open.
 
 ## Architecture and prerequisites
@@ -91,7 +94,7 @@ capability and lifecycle implementation; a state-handle type alone is insufficie
 | M1 | S25 Spark | Device-local Omni-compatible controller; embedding, prefill, continuous decode, ring/full KV, head and sampler; resident state | Reference checks, at least 128 output tokens, 512-window/1024-bucket transitions, cancellation, resident memory and sustained profile |
 | M2 | Qwen3-TTS | Resolve desktop playback and shutdown findings and CPU compatibility; then mobile talker/predictor plus GPU vocoder | Complete PCM and tail, history/chunk correctness, interruption/recovery, quality checks, no post-startup underruns and RTF below 1 in the declared workload |
 | M3 | AMD stages | Connect compatible real encoders or other useful coarse stages through current Omni workers; start from the existing 890M vision evidence | Actual node/device placement, numerical and task quality, complete downstream output, shared-memory and handoff costs; reject unhelpful splits |
-| M4a | MiniCPM-o 4.5 | Three-stage desktop text+WAV produced coherent text and nonzero WAV on RTX 5090 Laptop under WSL and native Windows, and WSL CPU, using separate constrained plans. One WSL RTX synthetic image passed through vision encoding to correct text and nonzero WAV after increasing the thinker GPU budget from 0.58 to 0.61; one synthetic tone passed through audio input to plausible text and nonzero WAV. Shutdown cleanup remains open. The CPU plan needed vLLM 0.28/0.29 processor compatibility and used swap; 20 serial CPU requests measured p50 21.86 s/p95 22.97 s. Native Windows needed longer startup limits; 20 serial requests measured p50 18.38 s/p95 18.73 s with host RAM/pagefile pressure. Add video encoder and persistent-state reference before mobile artifacts | Speech intelligibility/alignment, loading-peak admission, concurrency, broad image/audio-understanding suites, video, shutdown lifecycle, then combined streaming and interruption |
+| M4a | MiniCPM-o 4.5 | Three-stage desktop text+WAV produced coherent text and nonzero WAV on RTX 5090 Laptop under WSL and native Windows, and WSL CPU, using separate constrained plans. One synthetic red-square image and one synthetic tone each passed through the three-stage WSL RTX and native Windows RTX paths to text and nonzero WAV; matching inputs produced identical text and thinker token IDs across the two OS environments. The WSL image plan needed a thinker GPU budget increase from 0.58 to 0.61 for KV admission. WSL multimodal shutdown cleanup remains open. The CPU plan needed vLLM 0.28/0.29 processor compatibility and used swap; 20 serial CPU requests measured p50 21.86 s/p95 22.97 s. Native Windows needed longer startup limits; 20 serial text requests measured p50 18.38 s/p95 18.73 s with host RAM/pagefile pressure. Add video encoder and persistent-state reference before mobile artifacts | Speech intelligibility/alignment, loading-peak admission, concurrency, broad image/audio-understanding suites, video, shutdown lifecycle, then combined streaming and interruption |
 | M4b | InternVLA-A1 | Real weights, observation preprocessing, shared prefix state, iterative action head and versioned action metadata; strict-load synthetic policy forwards now pass on WSL and native Windows, each on CPU and RTX 5090 Laptop | Reference action agreement on real observations, units/order/timestamps, observation age, stale-input handling and application-specific deadline reporting |
 | Independent | Qwen3.8-27B | Public Omni pipeline binding; text, then image/video and longer contexts; same-model CPU/iGPU artifact and explicit offload evaluation | Modality-specific quality, actual loading/runtime memory and measured performance; each precision/context receives its own qualification |