diff options
| -rw-r--r-- | doc/I02_HiPerformance.dox | 28 |
1 files changed, 12 insertions, 16 deletions
diff --git a/doc/I02_HiPerformance.dox b/doc/I02_HiPerformance.dox index 8f23b5d19..2c77f87b3 100644 --- a/doc/I02_HiPerformance.dox +++ b/doc/I02_HiPerformance.dox @@ -63,6 +63,14 @@ handled by a single GEMM-like call are correctly detected. Make sure the matrix product is the top most expression.</td> </tr> <tr> +<td>\code m1 += s1 * (m2 * m3).lazy(); \endcode</td> +<td>\code m1 += s1 * m2 * m3; // using a naive product \endcode</td> +<td>\code m1 += (s1 * m2 * m3).lazy(); \endcode</td> +<td>Even though this expression is evaluated without temporary, it is actually even + worse than the previous case because here the .lazy() enforces Eigen to use a + naive (and slow) evaluation of the product.</td> +</tr> +<tr> <td>\code m1 = m1 + m2 * m3; \endcode</td> <td>\code temp = (m2 * m3).lazy(); m1 = m1 + temp; \endcode</td> <td>\code m1 += (m2 * m3).lazy(); \endcode</td> @@ -70,24 +78,12 @@ handled by a single GEMM-like call are correctly detected. and so the matrix product will be immediately evaluated.</td> </tr> <tr> -<td>\code m1 += ((s1 * m2).transpose() * m3).lazy(); \endcode</td> -<td>\code temp = (s1*m2).transpose(); m1 = (temp * m3).lazy(); \endcode</td> -<td>\code m1 += (s1 * m2.transpose() * m3).lazy(); \endcode</td> -<td>This is because our expression analyzer stops at the first expression which cannot - be converted to a scalar multiple of a conjugate and therefore the nested scalar - multiple cannot be properly extracted.</td> -</tr> -<tr> -<td>\code m1 += (m2.conjugate().transpose() * m3).lazy(); \endcode</td> -<td>\code temp = m2.conjugate().transpose(); m1 += (temp * m3).lazy(); \endcode</td> -<td>\code m1 += (m2.adjoint() * m3).lazy(); \endcode</td> -<td>Same reason. Use .adjoint() or .transpose().conjugate()</td> -</tr> -<tr> <td>\code m1 += ((s1*m2).block(....) * m3).lazy(); \endcode</td> <td>\code temp = (s1*m2).block(....); m1 += (temp * m3).lazy(); \endcode</td> <td>\code m1 += (s1 * m2.block(....) * m3).lazy(); \endcode</td> -<td>Same reason.</td> +<td>This is because our expression analyzer is currently not able to extract trivial + expressions nested in a Block expression. Therefore the nested scalar + multiple cannot be properly extracted.</td> </tr> </table> @@ -129,7 +125,7 @@ Of course all these remarks hold for all other kind of products that we will des </tr> <tr> <td>SYR</td> -<td>m.seductive<LowerTriangular>().rankUpdate(v,s)</td> +<td>m.sefadjointView<LowerTriangular>().rankUpdate(v,s)</td> <td></td> <td>Computes m += s * v * v.adjoint()</td> </tr> |