19 …d_device(const T *sT, T rU[DIM_U][NUM_COMP][rU_SIZE], T rV[DIM_V][NUM_COMP][rV_SIZE], const int tx,  in magma_interp_3d_device()  argument
34     if (tx < (P * P)) {  in magma_interp_3d_device()
35       const int batchid = tx;  in magma_interp_3d_device()
49     if (tx < (P * Q)) {  in magma_interp_3d_device()
50       const int batchid = tx / Q;  in magma_interp_3d_device()
51       const int tx_     = tx % Q;  in magma_interp_3d_device()
64     if (tx < (P * Q)) {  in magma_interp_3d_device()
65       const int batchid = tx / Q;  in magma_interp_3d_device()
66       const int tx_     = tx % Q;  in magma_interp_3d_device()
76     if (tx < (Q * Q)) {  in magma_interp_3d_device()
84           rTmp[0] += sTmp(tx, i, sld) * sT(i, j);  in magma_interp_3d_device()
99   const int tx      = threadIdx.x;  in __launch_bounds__()
120     read_T_notrans_gm2sm<BASIS_P, BASIS_Q>(tx, dT, sT);  in __launch_bounds__()
124   read_U_3d<CeedScalar, BASIS_P, 1, BASIS_NUM_COMP, BASIS_P, 0>(dU, cstrdU, rU, sTmp, tx);  in __launch_bounds__()
127 …<CeedScalar, 1, 1, BASIS_NUM_COMP, BASIS_P, BASIS_Q, BASIS_P, BASIS_Q>(sT, rU, rV, tx, rTmp, sTmp);  in __launch_bounds__()
131   write_V_3d<CeedScalar, BASIS_Q, 1, BASIS_NUM_COMP, BASIS_Q, 0>(dV, cstrdV, rV, tx);  in __launch_bounds__()
140   const int tx      = threadIdx.x;  in __launch_bounds__()
161     read_T_trans_gm2sm<BASIS_Q, BASIS_P>(tx, dT, sT);  in __launch_bounds__()
165   read_U_3d<CeedScalar, BASIS_Q, 1, BASIS_NUM_COMP, BASIS_Q, 0>(dU, cstrdU, rU, sTmp, tx);  in __launch_bounds__()
168 …<CeedScalar, 1, 1, BASIS_NUM_COMP, BASIS_Q, BASIS_P, BASIS_Q, BASIS_P>(sT, rU, rV, tx, rTmp, sTmp);  in __launch_bounds__()
172   write_V_3d<CeedScalar, BASIS_P, 1, BASIS_NUM_COMP, BASIS_P, 0>(dV, cstrdV, rV, tx);  in __launch_bounds__()
181   const int tx      = threadIdx.x;  in __launch_bounds__()
202     read_T_trans_gm2sm<BASIS_Q, BASIS_P>(tx, dT, sT);  in __launch_bounds__()
206   read_U_3d<CeedScalar, BASIS_Q, 1, BASIS_NUM_COMP, BASIS_Q, 0>(dU, cstrdU, rU, sTmp, tx);  in __launch_bounds__()
209 …<CeedScalar, 1, 1, BASIS_NUM_COMP, BASIS_Q, BASIS_P, BASIS_Q, BASIS_P>(sT, rU, rV, tx, rTmp, sTmp);  in __launch_bounds__()
213   sum_V_3d<CeedScalar, BASIS_P, 1, BASIS_NUM_COMP, BASIS_P, 0>(dV, cstrdV, rV, tx);  in __launch_bounds__()