Answer the following questions related to GPUs and CUDA.

a$.[7]$ Suppose we have two kernels: kernelA and kernel in the same program. The

program launches kernelA. When kernelA finishes, it generates data to be used by kernel.

What is the best way for kernelA to send data to kernelB? Justify.

b$.[5]$ Can we make two kernels, from the same program, execute on the same GPU at the

same time? If yes, explain how. If not, explain why not.

c$.[9]$ If the warp concept did not exist, what would be the implications on GPU

programming? State three implications and, for each one, explain in $1-2$ lines.