Cuda Study Notes (v)

Finally, the content of the thread is parsed: In SIMD terms, each of the 32 threads is called a line Cheng, which executes the same instruction, and each thread uses a private register to make this operation request.

Suddenly feel, do Cuda program like to go to Beijing to work: write MPI, but also to see Pthread, and then switch to English class to write a pile of homework, and also see jquery sometimes write a page unavoidably use, and go to Beijing, the morning bus rushed to the station, stop waiting for bus to rest, get off to the subway and back. This is Cuda. Cuda is to have the ability to toss, a while academic, a moment of technology.

Cuda Study Notes (v)