|
| | InteractionTensorBatch () |
| |
| double & | operator() (int t, int u, int v, int b) |
| |
| double | operator() (int t, int u, int v, int b) const |
| |
| double * | slot (int t, int u, int v) |
| | Pointer to the BatchSize-wide slot for a given (t,u,v).
|
| |
| const double * | slot (int t, int u, int v) const |
| |
template<int MaxL, int BatchSize = simd_batch_size>
struct occ::mults::InteractionTensorBatch< MaxL, BatchSize >
SOA interaction tensor batch: BatchSize tensors stored interleaved.
Layout: data[hermite_index(t,u,v) * BatchSize + batch_idx] This enables vectorized load/store across the batch dimension.
- Template Parameters
-
| MaxL | Maximum total rank |
| BatchSize | Number of tensors in the batch |