CUDA implementation of Dijkstra's algorithm
Budget: ₹12,500 – ₹37,500 INR
* A Dijkstra's algorithm can be used find the shortest path from source node to destination node.
* Current Stack depth of CUDA architecture is 24. i.e., the number of nested threads that can be invoked by a single thread is limited to 24 threads.
* In this version of code (CUDA) for Dijkstra algorithm, when the stack depth is reached, the parent thread has to wait for for one of its child threads to complete. so that, it can continue doing its operations to find the shortest path.
* New version of code is needed to somehow overcome this wait time. i.e., a new way of allocating the threads/data is needed.
* I will provide the version of code I have (if needed)
* Current Stack depth of CUDA architecture is 24. i.e., the number of nested threads that can be invoked by a single thread is limited to 24 threads.
* In this version of code (CUDA) for Dijkstra algorithm, when the stack depth is reached, the parent thread has to wait for for one of its child threads to complete. so that, it can continue doing its operations to find the shortest path.
* New version of code is needed to somehow overcome this wait time. i.e., a new way of allocating the threads/data is needed.
* I will provide the version of code I have (if needed)