[BUG]: Single-thread block reduce compilation issues

[gevtushenko]

Is this a duplicate?

• I confirmed there appear to be no duplicate issues for this bug and that I agree to the Code of Conduct

Type of Bug

Compile-time Error

Component

CUB

Describe the bug

Recent fixes of single-thread warp reduction broke single-thread block reduction.

How to Reproduce

#include <cub/block/block_reduce.cuh>

__global__ void kernel() {
  typename cub::BlockReduce<double, 1, cub::BLOCK_REDUCE_WARP_REDUCTIONS>::TempStorage storage;
}

int main() {}

Expected behavior

Successful compilation

Reproduction link

No response

Operating System

No response

nvidia-smi output

No response

NVCC version

No response