bank.cu中一处疑似笔误

bank.cu中transpose代码 ``` int nx2 = bx + threadIdx.y; int ny2 = by + threadIdx.x; if (nx2 < N && ny2 < N) { B[nx2 * N + ny2] = S[threadIdx.x][threadIdx.y]; } ``` 中转置后的B矩阵的坐标 (nx2, ny2) 中计算应为 `nx2 = by + threadIdx.y` 和 `ny2 = bx + threadIdx.x`？这样和A矩阵的坐标计算正好是转置的关系