NVIDIA provides a quite detailed but concise tutorial on CUDA programming. The example used there is like the convolution example used in this class.
"Threads in a thread block actually do run concurrently". To my understanding, the threads are divided into warps, and in one core, it can only run limited amount of warps at the same time, so how can it run concurrently?
NVIDIA provides a quite detailed but concise tutorial on CUDA programming. The example used there is like the convolution example used in this class.
"Threads in a thread block actually do run concurrently". To my understanding, the threads are divided into warps, and in one core, it can only run limited amount of warps at the same time, so how can it run concurrently?