tf::cudaFlowSequentialOptimizer tf::cudaFlowOptimizerBase taskflow/cuda/cuda_optimizer.hpp class friend class cudaFlowCapturer cudaFlowCapturer cudaFlowCapturer tf::cudaFlowSequentialOptimizer::cudaFlowSequentialOptimizer ()=default cudaFlowSequentialOptimizer constructs a sequential optimizer cudaGraph_t cudaGraph_t tf::cudaFlowSequentialOptimizer::_optimize (cudaFlowGraph &graph) _optimize cudaFlowGraph & graph class to capture a CUDA graph using a sequential stream A sequential capturing algorithm finds a topological order of the described graph and captures dependent GPU tasks using a single stream. All GPU tasks run sequentially without breaking inter dependencies. tf::cudaFlowSequentialOptimizer_levelize tf::cudaFlowSequentialOptimizer_optimize tf::cudaFlowSequentialOptimizer_toposort tf::cudaFlowSequentialOptimizercudaFlowCapturer tf::cudaFlowSequentialOptimizercudaFlowSequentialOptimizer