Fix multi-GPU inference on Blackwell (dual RTX 5080): four stacked fixes
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>
This commit is contained in:
@@ -240,7 +240,9 @@ public:
|
||||
void clearGpuBuffers();
|
||||
int getBindingIndexByName(const std::string& name);
|
||||
void UpdateOptions(const ANSCENTER::Options& options) { m_options = options; }
|
||||
void warmUp(int iterations = 5);
|
||||
// Returns false if any warmup inference failed — callers must treat the
|
||||
// engine as unusable (a slot that fails warmup must NOT join the pool).
|
||||
bool warmUp(int iterations = 5);
|
||||
void setVerbose(bool v) { m_verbose = v; }
|
||||
void setDisableGraphs(bool v) { m_disableGraphs = v; }
|
||||
|
||||
|
||||
Reference in New Issue
Block a user