trying optimize 3mm

polybench-c-4.2.1-beta/linear-algebra/kernels/3mm/3mm_optimized.c

@@ -96,11 +96,11 @@ void kernel_3mm(int ni, int nj, int nk, int nl, int nm,
   for (i = 0; i < _PB_NI; i++){
     for (j = 0; j < _PB_NJ; j++)
     	E[i][j] = SCALAR_VAL(0.0);
-    for (j = 0; j < _PB_NL; j++)
-	    G[i][j] = SCALAR_VAL(0.0);
     for (k = 0; k < _PB_NK; ++k)
       for (j = 0; j < _PB_NJ; j++)
         E[i][j] += A[i][k] * B[k][j];
+    for (j = 0; j < _PB_NL; j++)
+	    G[i][j] = SCALAR_VAL(0.0);
     for (k = 0; k < _PB_NJ; ++k)
       for (j = 0; j < _PB_NL; j++)
         G[i][j] += E[i][k] * F[k][j];