diff --git a/Makefile b/Makefile
index bc0e037..2e05f9a 100644
--- a/Makefile
+++ b/Makefile
@@ -68,5 +68,5 @@ clean:
 
 distclean: clean
 	@echo "===>  DIST CLEAN"
-	@rm -f $(TARGET)
+	@rm -f $(TARGET)*
 	@rm -f tags
diff --git a/config.mk b/config.mk
index a67dd4d..04b48dd 100644
--- a/config.mk
+++ b/config.mk
@@ -1,9 +1,9 @@
 # Supported: GCC, CLANG, ICC
-TAG ?= GCC
+TAG ?= ICC
 # SP or DP
 DATA_TYPE ?= DP
 # AOS or SOA
-DATA_LAYOUT ?= SOA
+DATA_LAYOUT ?= AOS
 
 #Feature options
 OPTIONS +=  -DALIGNMENT=64
diff --git a/include_GCC.mk b/include_GCC.mk
index d29cca9..954ac5d 100644
--- a/include_GCC.mk
+++ b/include_GCC.mk
@@ -10,6 +10,6 @@ ANSI_CFLAGS += -Wextra
 CFLAGS   = -O3 -march=znver1  -ffast-math -funroll-loops # -fopenmp
 ASFLAGS  =  -masm=intel
 LFLAGS   =
-DEFINES  = -D_GNU_SOURCE
-INCLUDES =
-LIBS     = -lm
+DEFINES  = -D_GNU_SOURCE -DLIKWID_PERFMON
+INCLUDES = $(LIKWID_INC)
+LIBS     = -lm $(LIKWID_LIB) -llikwid
diff --git a/include_ICC.mk b/include_ICC.mk
index ddb61ac..37b017f 100644
--- a/include_ICC.mk
+++ b/include_ICC.mk
@@ -3,15 +3,15 @@ LINKER = $(CC)
 
 OPENMP  = #-qopenmp
 PROFILE  = #-profile-functions -g  -pg
-# OPTS     = -fast -xCORE-AVX512 -qopt-zmm-usage=high $(PROFILE)
+ OPTS     = -Ofast -xCORE-AVX512 -qopt-zmm-usage=high $(PROFILE)
 #OPTS     = -fast -xCORE-AVX2  $(PROFILE)
 #OPTS     = -fast -xAVX  $(PROFILE)
 #OPTS     = -fast -xSSE4.2 $(PROFILE)
 #OPTS     = -fast -no-vec $(PROFILE)
-OPTS     = -fast -xHost $(PROFILE)
+#OPTS     = -fast -xHost $(PROFILE)
 CFLAGS   = $(PROFILE) -restrict $(OPENMP) $(OPTS)
-ASFLAGS  = -masm=intel
+ASFLAGS  = #-masm=intel
 LFLAGS   = $(PROFILE) $(OPTS) $(OPENMP)
-DEFINES  = -D_GNU_SOURCE # -DALIGNMENT=64  -DLIKWID_PERFMON   -DPRECISION=1
+DEFINES  = -D_GNU_SOURCE #-DLIKWID_PERFMON
 INCLUDES = #$(LIKWID_INC)
 LIBS     = -lm #$(LIKWID_LIB) -llikwid
diff --git a/src/force.c b/src/force.c
index 7ce9d8a..24a433f 100644
--- a/src/force.c
+++ b/src/force.c
@@ -39,9 +39,8 @@ double computeForce(
     MD_FLOAT sigma6 = param->sigma6;
     MD_FLOAT epsilon = param->epsilon;
     MD_FLOAT* fx = atom->fx; MD_FLOAT* fy = atom->fy; MD_FLOAT* fz = atom->fz;
-    MD_FLOAT S, E;
+    double S, E;
 
-    S = getTimeStamp();
     for(int i = 0; i < Nlocal; i++) {
         fx[i] = 0.0;
         fy[i] = 0.0;
@@ -49,7 +48,7 @@ double computeForce(
     }
 
     if(profile) {
-        LIKWID_MARKER_START("force");
+    //    LIKWID_MARKER_START("force");
     }
 
 #pragma omp parallel for
@@ -64,6 +63,8 @@ double computeForce(
         MD_FLOAT fiy = 0;
         MD_FLOAT fiz = 0;
 
+//	printf("%d: %d\n", i, numneighs);
+
         for(int k = 0; k < numneighs; k++) {
             int j = neighs[k];
             MD_FLOAT delx = xtmp - atom_x(j);
@@ -87,9 +88,8 @@ double computeForce(
     }
 
     if(profile) {
-        LIKWID_MARKER_STOP("force");
+     //   LIKWID_MARKER_STOP("force");
     }
 
-    E = getTimeStamp();
-    return E-S;
+    return 0.0;
 }
diff --git a/src/main-stub.c b/src/main-stub.c
index f2bf8fb..6178584 100644
--- a/src/main-stub.c
+++ b/src/main-stub.c
@@ -158,11 +158,17 @@ int main(int argc, const char *argv[]) {
         }
     }
 
-    const double estim_volume = (double)(atom->Nlocal * 6 * sizeof(MD_FLOAT) + (atoms_per_unit_cell - 1 + 2) * sizeof(int)) / 1000.0;
+    const double estim_volume = (double)
+	    (atom->Nlocal * 6 * sizeof(MD_FLOAT) +
+	     atom->Nlocal * (atoms_per_unit_cell - 1 + 2) * sizeof(int)) / 1000.0;
     printf("System size (unit cells): %dx%dx%d\n", param.nx, param.ny, param.nz);
     printf("Atoms per unit cell: %d\n", atoms_per_unit_cell);
     printf("Total number of atoms: %d\n", atom->Nlocal);
-    printf("Estimated memory volume (kB): %.4f\n", estim_volume);
+    printf("Estimated total data volume (kB): %.4f\n", estim_volume );
+    printf("Estimated atom data volume (kB): %.4f\n",
+		    (double)(atom->Nlocal * 3 * sizeof(MD_FLOAT)  / 1000.0));
+    printf("Estimated neighborlist data volume (kB): %.4f\n",
+		    (double)(atom->Nlocal * (atoms_per_unit_cell - 1 + 2) * sizeof(int)) / 1000.0);
 
     DEBUG("Initializing neighbor lists...\n");
     initNeighbor(&neighbor, &param);
@@ -173,12 +179,18 @@ int main(int argc, const char *argv[]) {
     DEBUG("Computing forces...\n");
     computeForce(&param, atom, &neighbor, 0);
 
-    double T_accum = 0.0;
+    double S, E;
+    S = getTimeStamp();
+    LIKWID_MARKER_START("force");
     for(int i = 0; i < param.ntimes; i++) {
-        T_accum += computeForce(&param, atom, &neighbor, 1);
+        computeForce(&param, atom, &neighbor, 1);
     }
+    LIKWID_MARKER_STOP("force");
+    E = getTimeStamp();
+    double T_accum = E-S;
 
-    printf("Total time: %.4f, Time/force: %.4f\n", T_accum, T_accum / param.ntimes);
+    printf("Total time: %.4f, Mega atom updates/s: %.4f\n",
+		    T_accum, atom->Nlocal * param.ntimes/T_accum/1.E6);
     LIKWID_MARKER_CLOSE;
     return EXIT_SUCCESS;
 }