类似地,所有对字符串和数组的访问在运行时被检查来确保他们没有越界,消除越界导致的内存重写和数据损坏的问题。
Similarly, all accesses to arrays and strings are checked at run-time to ensure that they are in bounds, eliminating the possibility of overwriting memory and corrupting data.
如果数据未在正确的一侧,则另一侧访问数据的需求则会持续发起越界调用。
If data resides on the wrong side, constant transitions will be triggered by the need of the other side to reach for that data.
算法通过运用线性存储器纹理解决了越界问题,并借助CUDA的共享内存以减少对全局内存的访问次数。
By using CUDA's texture from linear memory, over-bound accessing is avoided. For reducing access of global memory, the algorithm employs CUDA's Shared memory.
应用推荐