diff -Nur MFCG.orig/Makefile MFCG/Makefile
--- MFCG.orig/Makefile	2025-07-31 14:05:35.400907372 +0200
+++ MFCG/Makefile	2025-07-31 13:54:51.241425520 +0200
@@ -11,12 +11,12 @@
 	FLAGS_S = -std=c++0x -Wall -Winline -Wshadow -W -O3 -qopenmp-stubs -xHOST -pg
 endif
 
-INCLUDES =-Iinclude
-LDFLAGS =
-LIBS =
+INCLUDES =-Iinclude $(LIKWID_INC)
+LDFLAGS = $(LIKWID_LIB)
+LIBS = -llikwid
 
 #EXTRA_FLAGS :=
-CXXFLAGS = $(FLAGS) $(EXTRA_FLAGS)
+CXXFLAGS = -DLIKWID_PERFMON $(FLAGS) $(EXTRA_FLAGS)
 CXXFLAGS_S = $(FLAGS_S) $(EXTRA_FLAGS)
 
 all: perf perf-serial
diff -Nur MFCG.orig/src/Grid.cpp MFCG/src/Grid.cpp
--- MFCG.orig/src/Grid.cpp	2025-07-31 14:05:35.124111300 +0200
+++ MFCG/src/Grid.cpp	2025-07-31 12:26:38.527982296 +0200
@@ -3,6 +3,7 @@
 #include <omp.h>
 
 
+#include <likwid-marker.h>
 // ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
 // ////////////////////////////////////////////////////   Class Member Functions      /////////////////////////////////////////////////////////////
 // ////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
@@ -257,7 +258,11 @@
 
     int shift = halo?0:HALO;
 
-#pragma omp parallel for
+
+#pragma omp parallel
+{
+    LIKWID_MARKER_START("axpby");
+#pragma omp for
     for(int yIndex=shift; yIndex<res->numGrids_y(true)-shift; ++yIndex)
     {
 #pragma omp simd 
@@ -266,6 +271,8 @@
             (*res)(yIndex,xIndex) = (a*(*x)(yIndex,xIndex)) + (b*(*y)(yIndex,xIndex));
         }
     }
+    LIKWID_MARKER_STOP("axpby");
+}
 
     STOP_TIMER(AXPBY);
 }
@@ -306,7 +313,10 @@
     int shift = halo?0:HALO;
 
     double l2_sq = 0;
-#pragma omp parallel for reduction(+:l2_sq)
+#pragma omp parallel
+{
+    //LIKWID_MARKER_START("dotProduct");
+#pragma omp for reduction(+:l2_sq)
     for(int yIndex=shift; yIndex<x->numGrids_y(true)-shift; ++yIndex)
     {
 #pragma omp simd reduction(+:l2_sq)
@@ -315,7 +325,8 @@
             l2_sq += (*x)(yIndex,xIndex)*(*y)(yIndex,xIndex);
         }
     }
-
+    //LIKWID_MARKER_STOP("dotProduct");
+}
     STOP_TIMER(DOT_PRODUCT);
     return l2_sq;
 }
diff -Nur MFCG.orig/src/PDE.cpp MFCG/src/PDE.cpp
--- MFCG.orig/src/PDE.cpp	2025-07-31 14:05:35.304498410 +0200
+++ MFCG/src/PDE.cpp	2025-07-31 12:23:53.964004166 +0200
@@ -2,6 +2,7 @@
 #include <math.h>
 #include <iostream>
 #include <omp.h>
+#include <likwid.h>
 
 //default boundary function as in ex01
 double defaultBoundary(int i, int j, double h_x, double h_y)
@@ -97,7 +98,7 @@
 void PDE::applyStencil(Grid* res, Grid* u)
 {
     START_TIMER(APPLY_STENCIL);
-
+    
 #if DEBUG
     assert((res->numGrids_y(true)==grids_y) && (res->numGrids_x(true)==grids_x));
     assert((u->numGrids_y(true)==grids_y) && (u->numGrids_x(true)==grids_x));
@@ -109,7 +110,9 @@
     const double w_y = 1.0/(h_y*h_y);
     const double w_c = 2.0*w_x + 2.0*w_y;
 
-#pragma omp parallel for
+#pragma omp parallel
+{
+#pragma omp for
     for ( int j=1; j<ySize-1; ++j)
     {
 #pragma omp simd
@@ -118,7 +121,7 @@
             (*res)(j,i) = w_c*(*u)(j,i) - w_y*((*u)(j+1,i) + (*u)(j-1,i)) - w_x*((*u)(j,i+1) + (*u)(j,i-1));
         }
     }
-
+}
     STOP_TIMER(APPLY_STENCIL);
 }
 
@@ -163,7 +166,7 @@
     
     
 #pragma omp parallel
-  {
+{
     int nthreads = omp_get_num_threads();
     int tid = omp_get_thread_num();
     int istart = (xSize-2)/nthreads*tid+1;
@@ -171,6 +174,7 @@
     if(tid == nthreads-1)
       iend = xSize-2;
     //forward substitution begin
+    LIKWID_MARKER_START("precon-forward");
     for ( int j=1; j<ySize-1+nthreads-1; ++j)
     {
       int jj = j-tid;
@@ -182,8 +186,10 @@
       }
 #pragma omp barrier
     }
+    LIKWID_MARKER_STOP("precon-forward");
     //forward substitution end
     //backward substitution begin
+    LIKWID_MARKER_START("precon-backward");
     for ( int j=ySize-2+nthreads-1; j>0; --j)
     {
       int jj = j-tid;
@@ -195,6 +201,7 @@
       }
 #pragma omp barrier
     }
+    LIKWID_MARKER_STOP("precon-backward");
     //backward substitution end
   }
 
diff -Nur MFCG.orig/src/perf.cpp MFCG/src/perf.cpp
--- MFCG.orig/src/perf.cpp	2025-07-31 14:05:35.348015393 +0200
+++ MFCG/src/perf.cpp	2025-07-31 12:36:51.642220802 +0200
@@ -3,7 +3,7 @@
 #include "Solver.h"
 #include "test_macros.h"
 #include <omp.h>
-
+#include <likwid-marker.h>   
 double uSineFunc(int i, int j, double h_x, double h_y)
 {
     return sin(M_PI*i*h_x)*sin(M_PI*j*h_y);
@@ -47,9 +47,14 @@
         fprintf(stderr, "Dimensions for x and y must be greater than zero\n");
         return 0;
     }
-
+    LIKWID_MARKER_INIT;
+    LIKWID_MARKER_REGISTER("PCG");
 #pragma omp parallel
     {
+        LIKWID_MARKER_REGISTER("axpby");
+        
+        LIKWID_MARKER_REGISTER("precon-forward");
+        LIKWID_MARKER_REGISTER("precon-backward");
 #pragma omp single
         {
             numThreads = omp_get_num_threads();
@@ -92,4 +97,6 @@
     TESTS_END;
 
     PRINT_TIME_SUMMARY;
+    
+LIKWID_MARKER_CLOSE;
 }
diff -Nur MFCG.orig/src/Solver.cpp MFCG/src/Solver.cpp
--- MFCG.orig/src/Solver.cpp	2025-07-31 14:05:35.092939190 +0200
+++ MFCG/src/Solver.cpp	2025-07-31 12:22:19.267223886 +0200
@@ -1,5 +1,6 @@
 #include "Solver.h"
 #include "Grid.h"
+#include <likwid-marker.h>
 
 SolverClass::SolverClass(PDE *pde_, Grid *x_, Grid *b_):pde(pde_),x(x_),b(b_)
 {
@@ -27,6 +28,7 @@
 
     // PCG loop begin
     START_TIMER(PCG);
+    LIKWID_MARKER_START("PCG");
     while( (iter<niter) && (res_norm_sq>tol*tol) )
     {
         pde->applyStencil(v,p);
@@ -48,6 +50,7 @@
 #endif
         ++iter;
     }
+    LIKWID_MARKER_STOP("PCG");
     STOP_TIMER(PCG);
     // PCG loop end