Warps as Scheduling Units Block 1 Warps Block 2 Warps Block 3 Warps t0t1t2.t31 0.t1t to t1 t2...t31 一 Each block is divided into 32-thread warps An implementation technique,not part of the CUDA programming model - Warps are scheduling units in SM Threads in a warp execute in Single Instruction Multiple Data (SIMD)manner The number of threads in a warp may vary in future generations 电子科妓女学 O6 Warps as Scheduling Units – Each block is divided into 32-thread warps – An implementation technique, not part of the CUDA programming model – Warps are scheduling units in SM – Threads in a warp execute in Single Instruction Multiple Data (SIMD) manner – The number of threads in a warp may vary in future generations … …t0 t1 t2 … t31 … …t0 t1 t2 … t31 Block 1 Warps Block 2 Warps … …t0 t1 t2 … t31 Block 3 Warps