Verified on dual RTX 5080: 355/355 detections, both GPUs in round-robin,
zero CUDA errors. Requires TensorRT >= 10.16.1 in SharedCore (10.15.1 has
a Myelin l2cm crash when creating a context on a second Blackwell device).
1. blobFromGpuMats: thread_local GpuMat caches (tl_blob/tl_floatImg) were
device-blind — a pool thread warming GPU[1] then GPU[0] reused device-1
memory with a device-0 stream, causing an async illegal memory access
that poisoned every CUDA context (sticky 700). Track owning device and
rebuild the cache on device switch.
2. runInference: pool round-robin can dispatch to a slot on a different
GPU than the one the caller prepared inputs on. Detect foreign input
GpuMats via cudaPointerGetAttributes and migrate them through a pinned
host staging buffer (two-hop D2H+H2D; direct cross-device cudaMemcpy2D
is unsupported and GeForce has no P2P).
3. warmUp: now pins its own CUDA device before uploading dummy inputs and
returns success/failure; loadNetwork rejects engines whose warmup fails
so broken slots can never join the pool silently.
4. Blackwell builder workaround: setMaxAuxStreams(0) on SM 12.x plus an
.aux0 engine-cache suffix so pre-workaround caches rebuild.
Also: ANSFR-UnitTest gains --gpu N flag (sets CUDA_VISIBLE_DEVICES) for
single-GPU bisection, and rich diagnostics on input-copy failures.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
1. core/ANSLicensingSystem/ANSLicense.h:268 — added default initializers to ModelConfig POD members (inpWidth = 0, inpHeight = 0, etc.). This was the actual bug: CreateANSODHandle declares a local ModelConfig and only sets a few fields, leaving inpWidth/inpHeight as indeterminate stack memory. On your machine that memory consistently held ~80, slipping past the <=0 → 640 guard and driving the ORT session at 80×80 — which crashes YOLOv11's /model.12/Concat because 2·ceil(80/32)=6 ≠ ceil(80/16)=5.
2. modules/ANSODEngine/ANSONNXYOLO.h:75 — ONNXYOLO::setInputShape() now snaps width/height up to the next multiple of 32 (with a 64-pixel floor). Defense in depth: even if a future caller hands in a non-stride-32 size, the FPN concat alignment is preserved.
3. modules/ANSODEngine/ANSONNXYOLO.cpp — two diagnostic additions:
WarmUpEngine bails with a clear log line if the warm-up dims aren't a multiple of 32, instead of letting the model.12 error fire and mask the real cause.
transform() logs the input shape, canvas shape, source size, and dynamic flag once per session in DebugView — so any future shape-mismatch crash points at the offending HxW directly.
Thread-local staging Mat (video_player.cpp:1400-1407) — single biggest win. Eliminates the 12 MB per-call malloc/free cycle.
Contiguous get_buffer2 allocator (video_decoder.cpp:35-102) — keeps the 3 bulk memcpys cache-friendly. Would also enable FAST/zero-copy for resolutions where visible_h % 64 == 0.
SW-decoder thread config (video_decoder.cpp:528-540) — thread_count=0, thread_type=FRAME|SLICE. FRAME is downgraded to SLICE-only by AV_CODEC_FLAG_LOW_DELAY, but decode throughput is sufficient for your input rate.
SetTargetFPS(100) delivery throttle (already there) — caps onVideoFrame post-decode work at 10 FPS. Keeps the caller path warm-cached.
Instrumentation — [MEDIA_DecInit] / [MEDIA_Convert] / [MEDIA_SWDec] / [MEDIA_Timing] / [MEDIA_JpegTiming] — always-on regression detector, zero cost when ANSCORE_DEBUGVIEW=OFF.
When the decoder hasn't produced a frame in 5s, skip the call to
_playerClient->getImage() entirely and return the cached frame with
unchanged _pts. LabVIEW sees STALE PTS one poll earlier and can
trigger reconnect sooner.
Threshold matches the existing checks on the duplicate-PTS branch and
in areImagesIdentical() so all three stale paths agree. Near-zero cost:
one getLastFrameAgeMs() call before the main path.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>