What is the meaning of each member variable of the CUDA_BATCH_MEM_OP_NODE_PARAMS structure?
typedef struct CUDA_BATCH_MEM_OP_NODE_PARAMS_st { CUcontext ctx; unsigned int count; CUstreamBatchMemOpParams *paramArray; unsigned int flags; } CUDA_BATCH_MEM_OP_NODE_PARAMS; I want to validate the CUDA_BATCH_MEM_OP_NODE_PARAMS parameters, but I don’t know what their member variables mean, especially count and flags. What range should their variable values be in? It seems that flags are not currently used, it must be […]
How can I get the parameters of a conditional CUDA node?
The CUDA graphs API has many sets of functions related to the same node type, e.g.:
With graph edges now qualified with properties – can a pair of nodes have multiple edges?
With CUDA 12.3, CUDA graph (template) edges have properties, such as source and destination ports and edge type – is it now possible to have multiple different edges between the same pair of graph nodes? Specifically, edges with different properties?