Any elegant way to calculate lane, warp_id, and warp_num using cooperative_groups?

[202476410arsmart]

int lane = threadIdx.x % 32;  
int warp_id = threadIdx.x / 32;  
int warp_num = blockDim.x / 32;  

Is there a more elegant way to write this, perhaps using cooperative_groups or similar utilities like thread_rank?

[striker159]

I don’t think your code is not elegant. cooperative groups does nothing different, and even looks more verbose.

auto warp = cooperative_groups::tiled_partition<32>(cooperative_groups::this_thread_block());
int lane = warp.thread_rank();
int warp_id = warp.meta_group_rank();
int warp_num = warp.meta_group_size();