i have one doubt … i used to do my programs using single kernel function… but i want know how to use multiple kernel in one program … thing is i dont have graphics card… i am running my programs in EmDebug mode… is it support for multi kernel functions , if yes then can any one clarify
There is no need to put cudaThreadSynchronize() between two kernel calls. (unless you are benchmarking them individually) The driver will run them sequentially for you.