魔方问题是算法设计中一个难点，搜索算法的复杂度很高，在这个例子中将采用并行处理的算法对其中的限界剪枝算法进行搜索。-algorithm for solving the problem is a difficult design, search algorithm is highly complex, In this example will use parallel processing algorithms which limits pruning algorithm for the sear

并行化思路实现dijkstra算法。假设有p个处理器，N个顶点。给每个处理器分配N/p个顶点，求出局部的最小值，复杂度为O (N/p)。然后后一半的处理器将自己的最小值发送给第前p/2个处理器。前一半处理器接收到传来的值后，与局部的最小值比较，作为新值。继续循环，直到剩下一个处理器为止。-Parallelized dijkstra algorithm. Suppose there is p a processor, N vertex. Give each processor distributi

用CUDA实现BFS算法源码。CUDA是NV公司基于GPU的统一计算架构，BFS复杂度由CPU上的O(V+E)降为O(diameter)，直径diameter即图中root到leaf的最长距离。-This is the algorithm of BFS(breadth first search) on CUDA, and its time complexity is down to O(diameter) which diameter means the longest distance bet

Parallel Complexity Theory，并行复杂性理论，讲述并行算法的设计和复杂度分析-Parallel Complexity Theory