Home
last modified time | relevance | path

Searched defs:nshared (Results 1 – 1 of 1) sorted by relevance

/external/pytorch/aten/src/ATen/native/cuda/
Dlayer_norm_kernel.cu750 int nshared = threads.y > 1 ? threads.y * 3/2 *sizeof(T_ACC) : 0; in launch_vectorized_layer_norm_kernel() local
1168 int nshared = in LayerNormBackwardKernelImplInternal() local
1183 int nshared = (num_threads()/warp_size) * sizeof(T_ACC); in LayerNormBackwardKernelImplInternal() local