30   const Index base_n = 64 * n_block_idx;  in EigenContractionKernelInternal()  local
179       const Index rhs_horiz_0 = base_n + threadIdx.z + 0 * 8;   \  in EigenContractionKernelInternal()
180       const Index rhs_horiz_1 = base_n + threadIdx.z + 1 * 8;   \  in EigenContractionKernelInternal()
181       const Index rhs_horiz_2 = base_n + threadIdx.z + 2 * 8;   \  in EigenContractionKernelInternal()
182       const Index rhs_horiz_3 = base_n + threadIdx.z + 3 * 8;   \  in EigenContractionKernelInternal()
183       const Index rhs_horiz_4 = base_n + threadIdx.z + 4 * 8;   \  in EigenContractionKernelInternal()
184       const Index rhs_horiz_5 = base_n + threadIdx.z + 5 * 8;   \  in EigenContractionKernelInternal()
185       const Index rhs_horiz_6 = base_n + threadIdx.z + 6 * 8;   \  in EigenContractionKernelInternal()
186       const Index rhs_horiz_7 = base_n + threadIdx.z + 7 * 8;   \  in EigenContractionKernelInternal()
465   const int max_j_write = numext::mini((int)((n_size - base_n - threadIdx.z + 7) / 8), 8);  in EigenContractionKernelInternal()
479       output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * 0) = val0;  in EigenContractionKernelInternal()
480       output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * 1) = val1;  in EigenContractionKernelInternal()
481       output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * 2) = val2;  in EigenContractionKernelInternal()
482       output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * 3) = val3;  in EigenContractionKernelInternal()
483       output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * 4) = val4;  in EigenContractionKernelInternal()
484       output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * 5) = val5;  in EigenContractionKernelInternal()
485       output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * 6) = val6;  in EigenContractionKernelInternal()
486       output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * 7) = val7;  in EigenContractionKernelInternal()
491         output(base_m + threadIdx.y + 8 * threadIdx.x, base_n + threadIdx.z + 8 * j) = val;  in EigenContractionKernelInternal()
513   const Index base_n = 64 * n_block_idx;  in EigenContractionKernel()  local
515   if (base_m + 63 < m_size && base_n + 63 < n_size) {  in EigenContractionKernel()
531                        const Index base_m, const Index base_n) {  in EigenFloatContractionKernelInternal16x16()  argument
576     Index rhs_horiz0 = (threadIdx.x>>2)+threadIdx.y*4+base_n;  in EigenFloatContractionKernelInternal16x16()
691   Index horiz_base = threadIdx.y*4+base_n;  in EigenFloatContractionKernelInternal16x16()
768                        const Index base_m, const Index base_n) {  in EigenFloatContractionKernelInternal()  argument
907     Index rhs_horiz0 = threadIdx.y*2+base_n;  in EigenFloatContractionKernelInternal()
908     Index rhs_horiz1 = threadIdx.y*2+1+base_n;  in EigenFloatContractionKernelInternal()
1067   Index horiz_base = (threadIdx.y/4)*8+base_n;  in EigenFloatContractionKernelInternal()
1147   const Index base_n = 64 * n_block_idx;  in EigenFloatContractionKernel()  local
1149   bool check_rhs = (base_n + 63) >= n_size;  in EigenFloatContractionKernel()
1156 …utput, *((LHS_MEM *) lhs_shmem), *((RHS_MEM *) rhs_shmem), m_size, n_size, k_size, base_m, base_n);  in EigenFloatContractionKernel()
1159 …utput, *((LHS_MEM *) lhs_shmem), *((RHS_MEM *) rhs_shmem), m_size, n_size, k_size, base_m, base_n);  in EigenFloatContractionKernel()
1165 …utput, *((LHS_MEM *) lhs_shmem), *((RHS_MEM *) rhs_shmem), m_size, n_size, k_size, base_m, base_n);  in EigenFloatContractionKernel()
1168 …utput, *((LHS_MEM *) lhs_shmem), *((RHS_MEM *) rhs_shmem), m_size, n_size, k_size, base_m, base_n);  in EigenFloatContractionKernel()
1187   const Index base_n = 64 * n_block_idx;  in EigenFloatContractionKernel16x16()  local
1190     if (base_n + 63 < n_size) {  in EigenFloatContractionKernel16x16()
1191 …per, false, false>(lhs, rhs, output, lhs_shmem, rhs_shmem, m_size, n_size, k_size, base_m, base_n);  in EigenFloatContractionKernel16x16()
1193 …pper, false, true>(lhs, rhs, output, lhs_shmem, rhs_shmem, m_size, n_size, k_size, base_m, base_n);  in EigenFloatContractionKernel16x16()
1196     if (base_n + 63 < n_size) {  in EigenFloatContractionKernel16x16()
1197 …pper, true, false>(lhs, rhs, output, lhs_shmem, rhs_shmem, m_size, n_size, k_size, base_m, base_n);  in EigenFloatContractionKernel16x16()
1199 …apper, true, true>(lhs, rhs, output, lhs_shmem, rhs_shmem, m_size, n_size, k_size, base_m, base_n);  in EigenFloatContractionKernel16x16()