GPU/CUDA
vwenyu-L
这个作者很懒,什么都没留下…
展开
-
gpu/cuda-01-grid/block/thread
dim3 gridSize(m,n,z); dim3 blockSize(8,8,1); kernel>>(); threadIdx.x .y .z blockDim.x .y ,z blockIdx.x .y .z gridIdx.x .y .z原创 2016-08-28 15:58:08 · 330 阅读 · 0 评论 -
gpu/cuda-02-communication pattern
map gather scatter原创 2016-08-28 22:17:56 · 330 阅读 · 0 评论 -
gpu/cuda-03-cuda memory
local mem - thread shared mem - threads of one blobk global mem - SMs cpu: host mem原创 2016-08-31 17:44:36 · 301 阅读 · 0 评论