improve performance of trsm

1 files changed, 18 insertions, 14 deletions

diff --git a/Eigen/src/Core/products/TriangularSolverMatrix.h b/Eigen/src/Core/products/TriangularSolverMatrix.h
index a6ad9c322..4f5f85cbc 100644
--- a/Eigen/src/Core/products/TriangularSolverMatrix.h
+++ b/Eigen/src/Core/products/TriangularSolverMatrix.h
@@ -75,6 +75,7 @@ struct triangular_solve_matrix<Scalar,Index,OnTheLeft,Mode,Conjugate,TriStorageO
     ei_declare_aligned_stack_constructed_variable(Scalar, blockA, kc*mc, 0);
     ei_declare_aligned_stack_constructed_variable(Scalar, allocatedBlockB, sizeB, 0);
     Scalar* blockB = allocatedBlockB + sizeW;
+    Scalar* blockW = allocatedBlockB;
 
     conj_if<Conjugate> conj;
     gebp_kernel<Scalar, Scalar, Index, Traits::mr, Traits::nr, Conjugate, false> gebp_kernel;
@@ -92,16 +93,19 @@ struct triangular_solve_matrix<Scalar,Index,OnTheLeft,Mode,Conjugate,TriStorageO
       // A11 (the triangular part) and A21 the remaining rectangular part.
       // Then the high level algorithm is:
       //  - B = R1                    => general block copy (done during the next step)
-      //  - R1 = L1^-1 B              => tricky part
+      //  - R1 = A11^-1 B             => tricky part
       //  - update B from the new R1  => actually this has to be performed continuously during the above step
-      //  - R2 = L2 * B               => GEPP
-
-      // The tricky part: compute R1 = L1^-1 B while updating B from R1
-      // The idea is to split L1 into multiple small vertical panels.
-      // Each panel can be split into a small triangular part A1 which is processed without optimization,
-      // and the remaining small part A2 which is processed using gebp with appropriate block strides
+      //  - R2 -= A21 * B             => GEPP
+
+      // The tricky part: compute R1 = A11^-1 B while updating B from R1
+      // The idea is to split A11 into multiple small vertical panels.
+      // Each panel can be split into a small triangular part T1k which is processed without optimization,
+      // and the remaining small part T2k which is processed using gebp with appropriate block strides
+      Index subcols = (kc/Traits::nr)*Traits::nr; // TODO kc might not be an ideal choice here
+      for(Index j2=0; j2<cols; j2+=subcols)
       {
-        // for each small vertical panels of lhs
+        Index actual_cols = std::min(cols-j2,subcols);
+        // for each small vertical panels [T1k^T, T2k^T]^T of lhs
         for (Index k1=0; k1<actual_kc; k1+=SmallPanelWidth)
         {
           Index actualPanelWidth = std::min<Index>(actual_kc-k1, SmallPanelWidth);
@@ -114,7 +118,7 @@ struct triangular_solve_matrix<Scalar,Index,OnTheLeft,Mode,Conjugate,TriStorageO
             Index rs = actualPanelWidth - k - 1; // remaining size
 
             Scalar a = (Mode & UnitDiag) ? Scalar(1) : Scalar(1)/conj(tri(i,i));
-            for (Index j=0; j<cols; ++j)
+            for (Index j=j2; j<j2+actual_cols; ++j)
             {
               if (TriStorageOrder==RowMajor)
               {
@@ -143,7 +147,7 @@ struct triangular_solve_matrix<Scalar,Index,OnTheLeft,Mode,Conjugate,TriStorageO
           Index blockBOffset = IsLower ? k1 : lengthTarget;
 
           // update the respective rows of B from other
-          pack_rhs(blockB, _other+startBlock, otherStride, actualPanelWidth, cols, actual_kc, blockBOffset);
+          pack_rhs(blockB+actual_kc*j2, &other(startBlock,j2), otherStride, actualPanelWidth, actual_cols, actual_kc, blockBOffset);
 
           // GEBP
           if (lengthTarget>0)
@@ -152,13 +156,13 @@ struct triangular_solve_matrix<Scalar,Index,OnTheLeft,Mode,Conjugate,TriStorageO
 
             pack_lhs(blockA, &tri(startTarget,startBlock), triStride, actualPanelWidth, lengthTarget);
 
-            gebp_kernel(_other+startTarget, otherStride, blockA, blockB, lengthTarget, actualPanelWidth, cols, Scalar(-1),
-                        actualPanelWidth, actual_kc, 0, blockBOffset);
+            gebp_kernel(&other(startTarget,j2), otherStride, blockA, blockB+actual_kc*j2, lengthTarget, actualPanelWidth, actual_cols, Scalar(-1),
+                        actualPanelWidth, actual_kc, 0, blockBOffset, blockW);
           }
         }
       }
-
-      // R2 = A2 * B => GEPP
+      
+      // R2 -= A21 * B => GEPP
       {
         Index start = IsLower ? k2+kc : 0;
         Index end   = IsLower ? size : k2-kc;