diff --git a/spgemm/mkl_shmem/mklspgemm.c b/spgemm/mkl_shmem/mklspgemm.c
index 44f32c059772c2764d75370d448b9542a6299e17..d3d006be40ff553b7d48de7369a9c96fb8205fba 100644
--- a/spgemm/mkl_shmem/mklspgemm.c
+++ b/spgemm/mkl_shmem/mklspgemm.c
@@ -359,7 +359,7 @@ for(i=0;i<=Bm;i++)BI[i]--;    for(i=0;i<Bnnz;i++)BJ[i]--;
 printmm_one(Am, Cval, CJ, CI);
 
 /** In order to write the output C matrix to file, uncomment the following line */
-//printfilemm_one(strpathC, Am, Bn, Cval, CJ, CI);
+printfilemm_one(strpathC, Am, Bn, Cval, CJ, CI);
 
 /** run my SpGEMM routine in order to find number of multiply-and-add operations */
 long nummult = 0;
diff --git a/spgemm/mkl_xphi/CMakeLists.txt b/spgemm/mkl_xphi/CMakeLists.txt
index ff0936ace9bf35d13014fe51b9b23d1338274b72..4c73cd3c19897aa26b4b08668be2d84b4200bd1f 100644
--- a/spgemm/mkl_xphi/CMakeLists.txt
+++ b/spgemm/mkl_xphi/CMakeLists.txt
@@ -13,10 +13,11 @@ set(CMAKE_VERBOSE_MAKEFILE ON)
 
 #find_package(MKL)
 
-SET(ICC_MKL_LINK_FLAGS    " -mkl -offload-option,mic,ld,\"-rpath /opt/intel/mic/myo/lib/ -rpath /opt/intel/mic/coi/device-linux-release/lib/\"")
+#SET(ICC_MKL_LINK_FLAGS    " -mkl -offload-option,mic,ld,\"-rpath /opt/intel/mic/myo/lib/ -rpath /opt/intel/mic/coi/device-linux-release/lib/\"")
+SET(ICC_MKL_LINK_FLAGS    " -mkl ")
 #SET( CMAKE_EXE_LINKER_FLAGS  "${ICC_MKL_LINK_FLAGS}" )
-
 set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} ${ICC_MKL_LINK_FLAGS}")
+#set(CMAKE_CC_COMPILER_FLAGS " -g3 ")
 
 include(${CMAKE_CURRENT_SOURCE_DIR}/../../../../cmake/common.cmake)
 
diff --git a/spgemm/mkl_xphi/mklspgemm.c b/spgemm/mkl_xphi/mklspgemm.c
index f047405743ad9cf1e792d54fa9f1a1ca477826e5..2b4873249c623bd0524010e338afff07a8c53c42 100644
--- a/spgemm/mkl_xphi/mklspgemm.c
+++ b/spgemm/mkl_xphi/mklspgemm.c
@@ -190,31 +190,6 @@ csr *csr_multiply(csi Am, csi An, csi Anzmax, const csi* Ap, const csi* Aj, cons
 }/*}}}*/
 
 /** Multiply two sparse matrices which are stored in CSR format. MKL is used */
-void mkl_cpu_spgemm(MKL_INT Am, MKL_INT An, MKL_INT Annz, double* Aval, MKL_INT* AJ, MKL_INT* AI, MKL_INT Bn, MKL_INT Bnnz, double* Bval, MKL_INT* BJ, MKL_INT* BI, double** pCval, MKL_INT** pCJ, MKL_INT** pCI, double* time) {  /*{{{*/
-
-MKL_INT* CJ = NULL;
-double* Cval = NULL;
-MKL_INT sort = 3;    // sort everything
-MKL_INT* CI = (MKL_INT*)mkl_malloc( (Am+2) * sizeof( MKL_INT ), 64 );
-MKL_INT nnzmax = 0;    // nnzmax is zero in case of symbolic&numeric usage of mkl_?csrmultcsr
-MKL_INT ierr;
-MKL_INT request = 1;    // symbolic multiplication
-mkl_dcsrmultcsr(&transa, &request, &sort, &Am, &An, &Bn, Aval, AJ, AI, Bval, BJ, BI, Cval, CJ, CI, &nnzmax, &ierr);
-
-request = 2;        //numeric multiplication
-int Cnnz = CI[Am]-1;
-int Cval_size = Cnnz + 1;
-CJ = (MKL_INT*)mkl_malloc( Cval_size    *    sizeof( MKL_INT ), 64 );
-Cval = (double*)mkl_malloc( Cval_size    *    sizeof( double ), 64 );
-double time_st = dsecnd();
-int i;
-for(i = 0; i < nrepeat; i++) {
-    mkl_dcsrmultcsr(&transa, &request, &sort, &Am, &An, &Bn, Aval, AJ, AI, Bval, BJ, BI, Cval, CJ, CI, &nnzmax, &ierr);
-}
-double time_end = dsecnd();
-*time = (time_end - time_st)/nrepeat;
-*pCval = Cval; *pCJ = CJ; *pCI = CI;
-} /*}}}*/ 
 void mkl_mic_spgemm(MKL_INT Am, MKL_INT An, MKL_INT Annz, double* Aval, MKL_INT* AJ, MKL_INT* AI, MKL_INT Bn, MKL_INT Bnnz, double* Bval, MKL_INT* BJ, MKL_INT* BI, double** pCval, MKL_INT** pCJ, MKL_INT** pCI, double* time) {/*{{{*/
 
 	int i;
@@ -225,7 +200,7 @@ void mkl_mic_spgemm(MKL_INT Am, MKL_INT An, MKL_INT Annz, double* Aval, MKL_INT*
 		} else {
 			nthreads = mkl_get_max_threads();
 		}
-	}  // printf("MIC started \n"); fflush(stdout);
+	} 
 	#pragma offload target(mic:micdev) \
 		in(transa) \
 		in(Am) \
@@ -238,13 +213,6 @@ void mkl_mic_spgemm(MKL_INT Am, MKL_INT An, MKL_INT Annz, double* Aval, MKL_INT*
 		in(BI:length(An+1)	free_if(0)) \
 		in(BJ:length(Bnnz)	free_if(0)) 
 	{}
-	
-
-	//void mkl_dcsrmultcsr (char *transa, MKL_INT *job, MKL_INT *sort, MKL_INT *m, MKL_INT *n, MKL_INT *k, double *a, MKL_INT *ja, MKL_INT *ia, double *b, MKL_INT *jb, MKL_INT *ib, double *c, MKL_INT *jc, MKL_INT *ic, MKL_INT *nnzmax, MKL_INT *ierr);
-
-
-//	double s_initial = dsecnd();
-//	double s_elapsed = dsecnd() - s_initial; // seconds
 
 	MKL_INT	Cnnz_host = -1;
 	MKL_INT* CI = NULL;
@@ -288,10 +256,10 @@ void mkl_mic_spgemm(MKL_INT Am, MKL_INT An, MKL_INT Annz, double* Aval, MKL_INT*
 		CJ		= (MKL_INT*)mkl_malloc( Cval_size	*	sizeof( MKL_INT ), 64 );
 		Cval		= (double*)mkl_malloc( Cval_size	*	sizeof( double ), 64 );
 		mkl_dcsrmultcsr(&transa, &request, &sort, &Am, &An, &An, Aval, AJ, AI, Bval, BJ, BI, Cval, CJ, CI, &nnzmax, &ierr);
-		printmm_one(Am, Cval, CJ, CI);//printf("Cnnz=%d\n", Cnnz);
+		printmm_one(Am, Cval, CJ, CI);
 		sort = 7;		// do not sort anything
 		request = 2;		// numeric multiplication
-	}//printf("Cnnz_mic: %d\n", Cnnz_host);exit(-1);
+	}
 	double time_mic_numeric_mm = 0.0;
 	#pragma offload target(mic:micdev) nocopy(request) nocopy(sort) nocopy(i) nocopy(ierr) nocopy(nnzmax)  nocopy(transa) nocopy(Am) nocopy(An) nocopy(Bn) \
 		out(time_mic_numeric_mm) \
@@ -305,13 +273,6 @@ void mkl_mic_spgemm(MKL_INT Am, MKL_INT An, MKL_INT Annz, double* Aval, MKL_INT*
 		nocopy(CJ:	alloc_if(0)	free_if(0)) \
 		nocopy(Cval:	alloc_if(0)	free_if(0))
 	{
-		//sort = 7
-		//request = 1; 2'ye gore sure iki katina cikiyor
-		//request = 0; 2'ye gore sure uc katina cikiyor
-
-		//request = 2
-		//sort = 3; 7'ye gore %30 daha yavas
-		//printf("sort:%d request:%d\n", sort, request); // prints sort:7 request:2
 		for(i = 0; i < 10; i++) {
 			mkl_dcsrmultcsr(&transa, &request, &sort, &Am, &An, &An, Aval, AJ, AI, Bval, BJ, BI, Cval, CJ, CI, &nnzmax, &ierr);
 		}
@@ -321,7 +282,8 @@ void mkl_mic_spgemm(MKL_INT Am, MKL_INT An, MKL_INT Annz, double* Aval, MKL_INT*
 		}
 		time_mic_numeric_mm = (dsecnd() - s_initial) / nrepeat; // seconds
 	}
-	if (0) {
+	*time = time_mic_numeric_mm;
+	if (1) { // Copy result matrix from MIC to Host
 	int nelm_CI = Am + 2; 
 	int nelm_CJ = Cnnz_host + 1; 
 	int nelm_Cval = Cnnz_host + 1; 
@@ -340,7 +302,7 @@ void mkl_mic_spgemm(MKL_INT Am, MKL_INT An, MKL_INT Annz, double* Aval, MKL_INT*
 		for(i = 0; i < nelm_CI; i++) CI_host[i] = CI[i];
 		for(i = 0; i < nelm_CJ; i++) {CJ_host[i] = CJ[i]; Cval_host[i] = Cval[i];}
 	}
-	*pCval = Cval_host; *pCJ = CJ_host; *pCI = CI_host; //printf("Cnnz_host:%d\n", Cnnz_host);
+	*pCval = Cval_host; *pCJ = CJ_host; *pCI = CI_host; 
 	}
 } /* ENDOF mkl_mic_spmm }}}*/
 
@@ -491,7 +453,7 @@ for(i=0;i<=Bm;i++)BI[i]--;    for(i=0;i<Bnnz;i++)BJ[i]--;
 printmm_one(Am, Cval, CJ, CI);
 
 /** In order to write the output C matrix to file, uncomment the following line */
-//printfilemm_one(strpathC, Am, Bn, Cval, CJ, CI);
+printfilemm_one(strpathC, Am, Bn, Cval, CJ, CI);
 
 /** run my SpGEMM routine in order to find number of multiply-and-add operations */
 long nummult = 0;