diff --git a/.travis.yml b/.travis.yml
index 122fc091e..203af5518 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -36,10 +36,12 @@ matrix:
       env: PRK_TARGET=allupc
            UPC_IMPL=bupc
            GASNET_CONDUIT=udp
+           PRK_FLAGS="-Wc,-O3"
     - os: osx
       env: PRK_TARGET=allupc
            UPC_IMPL=bupc
            GASNET_CONDUIT=smp
+           PRK_FLAGS="-Wc,-O3"
     # Mac issue with thread_t (see https://github.com/humairakamal/fgmpi/pull/1)
     - os: osx
       env: PRK_TARGET=allfgmpi
@@ -57,11 +59,13 @@ matrix:
     - env: PRK_TARGET=allupc
            UPC_IMPL=bupc
            GASNET_CONDUIT=ofi
+           PRK_FLAGS="-Wc,-O3"
     # UPC over MPICH on Mac hangs - may be async progress issue
     - os: osx
       env: PRK_TARGET=allupc
            UPC_IMPL=bupc
            GASNET_CONDUIT=mpi
+           PRK_FLAGS="-Wc,-O3"
     # probably because of old GCC (4.6)
     - os: linux
       compiler: gcc
@@ -126,15 +130,19 @@ env:
   - PRK_TARGET=allupc
     UPC_IMPL=bupc
     GASNET_CONDUIT=mpi
+    PRK_FLAGS="-Wc,-O3"
   - PRK_TARGET=allupc
     UPC_IMPL=bupc
     GASNET_CONDUIT=ofi
+    PRK_FLAGS="-Wc,-O3"
   - PRK_TARGET=allupc
     UPC_IMPL=bupc
     GASNET_CONDUIT=smp
+    PRK_FLAGS="-Wc,-O3"
   - PRK_TARGET=allupc
     UPC_IMPL=bupc
     GASNET_CONDUIT=udp
+    PRK_FLAGS="-Wc,-O3"
   - PRK_TARGET=allupc
     UPC_IMPL=gupc
   - PRK_TARGET=allampi
diff --git a/Makefile b/Makefile
index fdc781f22..261071d22 100644
--- a/Makefile
+++ b/Makefile
@@ -38,11 +38,13 @@ ifndef matrix_rank
   matrix_rank=5
 endif
 
-ifndef default_opt_flags
-  default_opt_flags=-O3
+ifndef PRK_FLAGS
+  PRK_FLAGS=-O3
 endif
 
-default:
+default: allserial allopenmp allmpi
+
+help:
 	@echo "Usage: \"make all\"          (re-)builds all targets"
 	@echo "       \"make allserial\"    (re-)builds all serial targets"
 	@echo "       \"make allopenmp\"    (re-)builds all OpenMP targets"
@@ -72,117 +74,117 @@ alldarwin: allserial allopenmp allmpi1 allfgmpi allmpiopenmp allmpirma allshmem
 allfreaks: allcharm++ allampi allgrappa 
 
 allmpi1:
-	cd MPI1/Synch_global;        $(MAKE) global    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/Synch_p2p;           $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/Sparse;              $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/Transpose;           $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/Stencil;             $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/DGEMM;               $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/Nstream;             $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/Reduce;              $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/Random;              $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPI1/Branch;              $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"  \
+	cd MPI1/Synch_global;        $(MAKE) global    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/Synch_p2p;           $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/Sparse;              $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/Transpose;           $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/Stencil;             $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/DGEMM;               $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/Nstream;             $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/Reduce;              $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/Random;              $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPI1/Branch;              $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"  \
                                                        "MATRIX_RANK         = $(matrix_rank)"        \
                                                        "NUMBER_OF_FUNCTIONS = $(number_of_functions)"
 
 allampi:
-	cd AMPI/Synch_global;        $(MAKE) global    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/Synch_p2p;           $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/Sparse;              $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/Transpose;           $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/Stencil;             $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/DGEMM;               $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/Nstream;             $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/Reduce;              $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/Random;              $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd AMPI/Branch;              $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"  \
+	cd AMPI/Synch_global;        $(MAKE) global    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/Synch_p2p;           $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/Sparse;              $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/Transpose;           $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/Stencil;             $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/DGEMM;               $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/Nstream;             $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/Reduce;              $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/Random;              $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd AMPI/Branch;              $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"  \
                                                        "MATRIX_RANK         = $(matrix_rank)"        \
                                                        "NUMBER_OF_FUNCTIONS = $(number_of_functions)"
 
 allfgmpi:
 	cd scripts/small;              $(MAKE) -f  Makefile_FG_MPI runfgmpi
 	cd scripts/wide;               $(MAKE) -f  Makefile_FG_MPI runfgmpi
-	cd FG_MPI/Synch_global;        $(MAKE) global    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/Synch_p2p;           $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/Sparse;              $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/Transpose;           $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/Stencil;             $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/DGEMM;               $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/Nstream;             $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/Reduce;              $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/Random;              $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd FG_MPI/Branch;              $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"  \
+	cd FG_MPI/Synch_global;        $(MAKE) global    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/Synch_p2p;           $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/Sparse;              $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/Transpose;           $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/Stencil;             $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/DGEMM;               $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/Nstream;             $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/Reduce;              $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/Random;              $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd FG_MPI/Branch;              $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"  \
                                                          "MATRIX_RANK         = $(matrix_rank)"        \
                                                          "NUMBER_OF_FUNCTIONS = $(number_of_functions)"
 
 allmpiopenmp:
-	cd MPIOPENMP/Nstream;       $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPIOPENMP/Synch_p2p;     $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPIOPENMP/Stencil;       $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPIOPENMP/Transpose;     $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
+	cd MPIOPENMP/Nstream;       $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPIOPENMP/Synch_p2p;     $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPIOPENMP/Stencil;       $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPIOPENMP/Transpose;     $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
 
 allmpiomp: allmpiopenmp
 
 allmpirma:
-	cd MPIRMA/Synch_p2p;        $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPIRMA/Stencil;          $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPIRMA/Transpose;        $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
+	cd MPIRMA/Synch_p2p;        $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPIRMA/Stencil;          $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPIRMA/Transpose;        $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
 
 allshmem:
-	cd SHMEM/Synch_p2p;         $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SHMEM/Stencil;           $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SHMEM/Transpose;         $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
+	cd SHMEM/Synch_p2p;         $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SHMEM/Stencil;           $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SHMEM/Transpose;         $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
 
 allmpishm:
-	cd MPISHM/Synch_p2p;        $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPISHM/Stencil;          $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd MPISHM/Transpose;        $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
+	cd MPISHM/Synch_p2p;        $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPISHM/Stencil;          $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd MPISHM/Transpose;        $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
 
 allmpi: allmpi1 allmpiomp allmpirma allmpishm
 
 allupc:
-	cd UPC/Synch_p2p;           $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd UPC/Stencil;             $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd UPC/Transpose;           $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
+	cd UPC/Synch_p2p;           $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd UPC/Stencil;             $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd UPC/Transpose;           $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
 
 allpgas: allshmem allupc allmpirma
 
 allopenmp:
-	cd OPENMP/DGEMM;            $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Nstream;          $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Reduce;           $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/RefCount_shared;  $(MAKE) shared    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/RefCount_private; $(MAKE) private   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Stencil;          $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Transpose;        $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Random;           $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Sparse;           $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Synch_global;     $(MAKE) global    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Synch_p2p;        $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd OPENMP/Branch;           $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"  \
+	cd OPENMP/DGEMM;            $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Nstream;          $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Reduce;           $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/RefCount_shared;  $(MAKE) shared    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/RefCount_private; $(MAKE) private   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Stencil;          $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Transpose;        $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Random;           $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Sparse;           $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Synch_global;     $(MAKE) global    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Synch_p2p;        $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd OPENMP/Branch;           $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"  \
                                                       "MATRIX_RANK         = $(matrix_rank)"        \
                                                       "NUMBER_OF_FUNCTIONS = $(number_of_functions)"
 
 allcharm++:
-	cd CHARM++/Synch_p2p;       $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd CHARM++/Stencil;         $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd CHARM++/Transpose;       $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
+	cd CHARM++/Synch_p2p;       $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd CHARM++/Stencil;         $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd CHARM++/Transpose;       $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
 
 allgrappa:
-	cd GRAPPA/Synch_p2p;       $(MAKE) p2p        "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd GRAPPA/Stencil;         $(MAKE) stencil    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd GRAPPA/Transpose;       $(MAKE) transpose  "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
+	cd GRAPPA/Synch_p2p;       $(MAKE) p2p        "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd GRAPPA/Stencil;         $(MAKE) stencil    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd GRAPPA/Transpose;       $(MAKE) transpose  "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
 
 allserial:
-	cd SERIAL/DGEMM;            $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SERIAL/Nstream;          $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SERIAL/Reduce;           $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SERIAL/Stencil;          $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SERIAL/Transpose;        $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SERIAL/Random;           $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SERIAL/Sparse;           $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SERIAL/Synch_p2p;        $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
-	cd SERIAL/Branch;           $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"  \
+	cd SERIAL/DGEMM;            $(MAKE) dgemm     "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SERIAL/Nstream;          $(MAKE) nstream   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SERIAL/Reduce;           $(MAKE) reduce    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SERIAL/Stencil;          $(MAKE) stencil   "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SERIAL/Transpose;        $(MAKE) transpose "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SERIAL/Random;           $(MAKE) random    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SERIAL/Sparse;           $(MAKE) sparse    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SERIAL/Synch_p2p;        $(MAKE) p2p       "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"
+	cd SERIAL/Branch;           $(MAKE) branch    "DEFAULT_OPT_FLAGS   = $(PRK_FLAGS)"  \
                                                       "MATRIX_RANK         = $(matrix_rank)"        \
                                                       "NUMBER_OF_FUNCTIONS = $(number_of_functions)"
 	cd SERIAL/PIC;              $(MAKE) pic       "DEFAULT_OPT_FLAGS   = $(default_opt_flags)"
diff --git a/OPENMP/DGEMM/dgemm.c b/OPENMP/DGEMM/dgemm.c
index 7f0805ed9..cd7fe666f 100755
--- a/OPENMP/DGEMM/dgemm.c
+++ b/OPENMP/DGEMM/dgemm.c
@@ -87,7 +87,7 @@ HISTORY: Written by Rob Van der Wijngaart, September 2006.
 
 #define forder (1.0*order)
 
-main(int argc, char **argv){
+int main(int argc, char **argv){
 
   int     iter, i,ii,j,jj,k,kk,ig,jg,kg; /* dummies                               */
   int     iterations;           /* number of times the multiplication is done     */
@@ -167,7 +167,7 @@ main(int argc, char **argv){
 
   #pragma omp parallel private (i,j,k,ii,jj,kk,ig,jg,kg,iter)
   {
-  double *AA, *BB, *CC;
+  double *AA=NULL, *BB=NULL, *CC=NULL;
 
   if (block > 0) {
     /* matrix blocks for local temporary copies                                     */
diff --git a/OPENMP/Nstream/nstream.c b/OPENMP/Nstream/nstream.c
index ab3142119..1299294be 100755
--- a/OPENMP/Nstream/nstream.c
+++ b/OPENMP/Nstream/nstream.c
@@ -230,8 +230,11 @@ int main(int argc, char **argv)
   }
   bail_out(num_error); 
 
+  /* FIXME Use OpenMP 4 via _Pragma */
   #pragma omp for
+#ifdef __INTEL_COMPILER
   #pragma vector always
+#endif
   for (j=0; j<length; j++) {
     a[j] = 0.0;
     b[j] = 2.0;
@@ -252,8 +255,11 @@ int main(int argc, char **argv)
       }
     }
  
+  /* FIXME Use OpenMP 4 via _Pragma */
     #pragma omp for
+#ifdef __INTEL_COMPILER
     #pragma vector always
+#endif
     for (j=0; j<length; j++) a[j] += b[j]+scalar*c[j];
  
   } /* end of iterations                                              */
diff --git a/OPENMP/Reduce/reduce.c b/OPENMP/Reduce/reduce.c
index 126e84907..2c20b4e8a 100755
--- a/OPENMP/Reduce/reduce.c
+++ b/OPENMP/Reduce/reduce.c
@@ -150,7 +150,7 @@ int main(int argc, char ** argv)
 
   vector_length  = atol(*++argv);
   if (vector_length < 1){
-    printf("ERROR: vector length must be >= 1 : %d \n",vector_length);
+    printf("ERROR: vector length must be >= 1 : %ld \n",vector_length);
     exit(EXIT_FAILURE);
   }
 
@@ -195,7 +195,7 @@ int main(int argc, char ** argv)
   } 
   else {
     printf("Number of threads              = %d\n",nthread_input);
-    printf("Vector length                  = %d\n", vector_length);
+    printf("Vector length                  = %ld\n", vector_length);
     printf("Reduction algorithm            = %s\n", algorithm);
     printf("Number of iterations           = %d\n", iterations);
   }
diff --git a/SERIAL/Branch/branch.c b/SERIAL/Branch/branch.c
index 4a56d8fed..bcb5c5c0c 100755
--- a/SERIAL/Branch/branch.c
+++ b/SERIAL/Branch/branch.c
@@ -145,7 +145,7 @@ int main(int argc, char ** argv)
            total_ref;       /* computed and stored verification values           */
   int * RESTRICT vector; 
   int * RESTRICT index;
-  int factor = -1;
+  /* UNUSED int factor = -1; */
 
 /**********************************************************************************
 ** process and test input parameters    
@@ -169,7 +169,7 @@ int main(int argc, char ** argv)
 
   vector_length  = atol(*++argv);
   if (vector_length < 1){
-     printf("ERROR: loop length must be >= 1 : %d \n",vector_length);
+     printf("ERROR: loop length must be >= 1 : %ld \n",vector_length);
      exit(EXIT_FAILURE);
   }
 
@@ -214,13 +214,17 @@ int main(int argc, char ** argv)
     case VECTOR_STOP:
       /* condition vector[index[i]]>0 inhibits vectorization                     */
       for (iter=0; iter<iterations; iter+=2) {
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) { 
           aux = -(3 - (i&7));
           if (vector[index[i]]>0) vector[i] -= 2*vector[i];
           else                    vector[i] -= 2*aux;
         }
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) { 
           aux = (3 - (i&7));
           if (vector[index[i]]>0) vector[i] -= 2*vector[i];
@@ -232,13 +236,17 @@ int main(int argc, char ** argv)
     case VECTOR_GO:
       /* condition aux>0 allows vectorization                                    */
       for (iter=0; iter<iterations; iter+=2) {
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) {
           aux = -(3 - (i&7));
           if (aux>0) vector[i] -= 2*vector[i];
           else       vector[i] -= 2*aux;
         }
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) {
           aux = (3 - (i&7));
           if (aux>0) vector[i] -= 2*vector[i];
@@ -250,13 +258,17 @@ int main(int argc, char ** argv)
     case NO_VECTOR:
       /* condition aux>0 allows vectorization, but indirect indexing inbibits it */
       for (iter=0; iter<iterations; iter+=2) {
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) {
           aux = -(3 - (i&7));
           if (aux>0) vector[i] -= 2*vector[index[i]];
           else       vector[i] -= 2*aux;
         }
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) {
           aux = (3 - (i&7));
           if (aux>0) vector[i] -= 2*vector[index[i]];
@@ -286,7 +298,9 @@ int main(int argc, char ** argv)
     case VECTOR_STOP:
     case VECTOR_GO:
       for (iter=0; iter<iterations; iter+=2) {
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) { 
           aux = -(3-(i&7)); 
           vector[i] -= (vector[i] + aux);
@@ -300,12 +314,16 @@ int main(int argc, char ** argv)
 
     case NO_VECTOR:
       for (iter=0; iter<iterations; iter+=2) {
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) {
           aux = -(3-(i&7));
           vector[i] -= (vector[index[i]]+aux); 
         }
+#ifdef __INTEL_COMPILER
         #pragma vector always
+#endif
         for (i=0; i<vector_length; i++) {
           aux = (3-(i&7));
           vector[i] -= (vector[index[i]]+aux); 
diff --git a/SERIAL/DGEMM/dgemm.c b/SERIAL/DGEMM/dgemm.c
index 1418316e7..ee670995a 100755
--- a/SERIAL/DGEMM/dgemm.c
+++ b/SERIAL/DGEMM/dgemm.c
@@ -80,8 +80,8 @@ HISTORY: Written by Rob Van der Wijngaart, February 2009.
 
 #define forder (1.0*order)
 
-main(int argc, char **argv){
-
+int main(int argc, char **argv)
+{
   int     iter, i,ii,j,jj,k,kk,ig,jg,kg; /* dummies                               */
   int     iterations;           /* number of times the multiplication is done     */
   double  dgemm_time,           /* timing parameters                              */
@@ -120,7 +120,7 @@ main(int argc, char **argv){
     order    = -order;
   }
   if (order < 1) {
-    printf("ERROR: Matrix order must be positive: %d\n", order);
+    printf("ERROR: Matrix order must be positive: %ld\n", order);
     exit(EXIT_FAILURE);
   }
 
@@ -130,11 +130,11 @@ main(int argc, char **argv){
   } else block = DEFAULTBLOCK;
 #endif
 
-  printf("Matrix order          = %d\n", order);
+  printf("Matrix order          = %ld\n", order);
   if (shortcut) 
     printf("Only doing initialization\n"); 
   if (block>0)
-    printf("Blocking factor       = %d\n", block);
+    printf("Blocking factor       = %ld\n", block);
   else
     printf("No blocking\n");
   printf("Number of iterations  = %d\n", iterations);
diff --git a/SERIAL/Nstream/nstream.c b/SERIAL/Nstream/nstream.c
index 831d83cb8..14c5160d3 100755
--- a/SERIAL/Nstream/nstream.c
+++ b/SERIAL/Nstream/nstream.c
@@ -206,7 +206,9 @@ int main(int argc, char **argv)
   printf("Offset               = %ld\n", offset);
   printf("Number of iterations = %d\n", iterations);
 
+#ifdef __INTEL_COMPILER
   #pragma vector always
+#endif
   for (j=0; j<length; j++) {
     a[j] = 0.0;
     b[j] = 2.0;
@@ -217,12 +219,16 @@ int main(int argc, char **argv)
  
   scalar = SCALAR;
  
+  nstream_time = 0.0; /* silence compiler warning */
+
   for (iter=0; iter<=iterations; iter++) {
  
     /* start timer after a warmup iteration */
     if (iter == 1) nstream_time = wtime();
  
+#ifdef __INTEL_COMPILER
     #pragma vector always
+#endif
     for (j=0; j<length; j++) a[j] += b[j]+scalar*c[j];
  
   }
diff --git a/SERIAL/Random/random.c b/SERIAL/Random/random.c
index e2f8f3e77..72015b8d6 100644
--- a/SERIAL/Random/random.c
+++ b/SERIAL/Random/random.c
@@ -250,7 +250,7 @@ int main(int argc, char **argv) {
   /* even though the table size can be represented, computing the space 
      required for the table may lead to overflow                            */
   tablespace = (size_t) tablesize*sizeof(u64Int);
-  if ((tablespace/sizeof(u64Int)) != tablesize || tablespace <=0) {
+  if ((signed)(tablespace/sizeof(u64Int)) != tablesize || tablespace <=0) {
     printf("Cannot represent space for table on this system; ");
     printf("reduce log2 tablesize\n");
     exit(EXIT_FAILURE);
diff --git a/SERIAL/Reduce/reduce.c b/SERIAL/Reduce/reduce.c
index 014fb0b6b..4b790cf3f 100755
--- a/SERIAL/Reduce/reduce.c
+++ b/SERIAL/Reduce/reduce.c
@@ -112,6 +112,8 @@ int main(int argc, char ** argv)
     ones[i]    = (double)1;
   }
 
+  reduce_time = 0.0; /* silence compiler warning */
+
   for (iter=0; iter<=iterations; iter++) {
 
     /* start timer after a warmup iteration */
diff --git a/SERIAL/Sparse/sparse.c b/SERIAL/Sparse/sparse.c
index cfc828241..4b794a5c1 100755
--- a/SERIAL/Sparse/sparse.c
+++ b/SERIAL/Sparse/sparse.c
@@ -146,7 +146,7 @@ int main(int argc, char **argv){
 
   /* emit error if (periodic) stencil overlaps with itself                        */
   if (size <2*radius+1) {
-    printf("ERROR: Grid extent %d smaller than stencil diameter 2*%d+1= %d\n",
+    printf("ERROR: Grid extent %lld smaller than stencil diameter 2*%lld+1= %lld\n",
            size, radius, radius*2+1);
     exit(EXIT_FAILURE);
   }
@@ -162,13 +162,13 @@ int main(int argc, char **argv){
   matrix_space = nent*sizeof(double);
   matrix = (double *) malloc(matrix_space);
   if (!matrix) {
-    printf("ERROR: Could not allocate space for sparse matrix: %ld\n", nent);
+    printf("ERROR: Could not allocate space for sparse matrix: %lld\n", nent);
     exit(EXIT_FAILURE);
   } 
 
   vector_space = 2*size2*sizeof(double);
-  if (vector_space/sizeof(double) != 2*size2) {
-    printf("ERROR: Cannot represent space for vectors: %ul\n", vector_space);
+  if (vector_space/sizeof(double) != (size_t)2*size2) {
+    printf("ERROR: Cannot represent space for vectors: %lu\n", vector_space);
     exit(EXIT_FAILURE);
   } 
 
@@ -180,8 +180,8 @@ int main(int argc, char **argv){
   result = vector + size2;
 
   index_space = nent*sizeof(s64Int);
-  if (index_space/sizeof(s64Int) != nent) {
-    printf("ERROR: Cannot represent space for column indices: %ul\n", index_space);
+  if (index_space/sizeof(s64Int) != (size_t)nent) {
+    printf("ERROR: Cannot represent space for column indices: %lu\n", index_space);
     exit(EXIT_FAILURE);
   } 
   colIndex = (s64Int *) malloc(index_space);
@@ -192,7 +192,7 @@ int main(int argc, char **argv){
   } 
 
   printf("Matrix order          = "FSTR64U"\n", size2);
-  printf("Stencil diameter      = %16d\n", 2*radius+1);
+  printf("Stencil diameter      = %16lld\n", 2*radius+1);
   printf("Sparsity              = %16.10lf\n", sparsity);
   printf("Number of iterations  = %16d\n", iterations);
 #ifdef SCRAMBLE
@@ -224,6 +224,8 @@ int main(int argc, char **argv){
       matrix[elm] = 1.0/(double)(colIndex[elm]+1);
   }
 
+  sparse_time = 0.0; /* silence compiler warning */
+
   for (iter=0; iter<=iterations; iter++) {
 
     if (iter==1) sparse_time = wtime();
@@ -234,7 +236,9 @@ int main(int argc, char **argv){
     /* do the actual matrix-vector multiplication                                 */
     for (row=0; row<size2; row++) {
       first = stencil_size*row; last = first+stencil_size-1;
+#ifdef __INTEL_COMPILER
       #pragma simd reduction(+:temp)
+#endif
       for (temp=0.0,col=first; col<=last; col++) {
         temp += matrix[col]*vector[colIndex[col]];
       }
diff --git a/SERIAL/Stencil/stencil.c b/SERIAL/Stencil/stencil.c
index 79a610365..d7af383f3 100755
--- a/SERIAL/Stencil/stencil.c
+++ b/SERIAL/Stencil/stencil.c
@@ -137,14 +137,14 @@ int main(int argc, char ** argv) {
   }
 
   if (2*RADIUS +1 > n) {
-    printf("ERROR: Stencil radius %d exceeds grid size %d\n", RADIUS, n);
+    printf("ERROR: Stencil radius %d exceeds grid size %ld\n", RADIUS, n);
     exit(EXIT_FAILURE);
   }
 
   /*  make sure the vector space can be represented                             */
   total_length = n*n*sizeof(DTYPE);
-  if (total_length/n != n*sizeof(DTYPE)) {
-    printf("ERROR: Space for %d x %d grid cannot be represented; ", n, n);
+  if (total_length/n != n*(signed)sizeof(DTYPE)) {
+    printf("ERROR: Space for %ld x %ld grid cannot be represented; ", n, n);
     exit(EXIT_FAILURE);
   }
 
@@ -181,7 +181,7 @@ int main(int argc, char ** argv) {
   norm = (DTYPE) 0.0;
   f_active_points = (DTYPE) (n-2*RADIUS)*(DTYPE) (n-2*RADIUS);
 
-  printf("Grid size            = %d\n", n);
+  printf("Grid size            = %ld\n", n);
   printf("Radius of stencil    = %d\n", RADIUS);
 #ifdef STAR
   printf("Type of stencil      = star\n");
@@ -208,6 +208,8 @@ int main(int argc, char ** argv) {
   for (j=RADIUS; j<n-RADIUS; j++) for (i=RADIUS; i<n-RADIUS; i++) 
     OUT(i,j) = (DTYPE)0.0;
 
+  stencil_time = 0.0; /* silence compiler warning */
+
   for (iter = 0; iter<=iterations; iter++){
 
     /* start timer after a warmup iteration */
diff --git a/SERIAL/Synch_p2p/p2p.c b/SERIAL/Synch_p2p/p2p.c
index 3979f5690..05ec170e9 100755
--- a/SERIAL/Synch_p2p/p2p.c
+++ b/SERIAL/Synch_p2p/p2p.c
@@ -117,6 +117,8 @@ int main(int argc, char ** argv) {
   for (j=0; j<n; j++) ARRAY(0,j) = (double) j;
   for (i=0; i<m; i++) ARRAY(i,0) = (double) i;
 
+  pipeline_time = 0.0; /* silence compiler warning */
+
   for (iter = 0; iter<=iterations; iter++){
 
     /* start timer after a warmup iteration */
diff --git a/SERIAL/Transpose/transpose.c b/SERIAL/Transpose/transpose.c
index f5ead244d..3aab90b03 100755
--- a/SERIAL/Transpose/transpose.c
+++ b/SERIAL/Transpose/transpose.c
@@ -64,7 +64,8 @@ HISTORY: Written by  Rob Van der Wijngaart, February 2009.
 #define A(i,j)        A_p[(i)+order*(j)]
 #define B(i,j)        B_p[(i)+order*(j)]
 
-static double test_results (int , double*);
+/* Never used...
+ * static double test_results (int , double*); */
 
 int main(int argc, char ** argv) {
 
@@ -101,7 +102,7 @@ int main(int argc, char ** argv) {
 
   order = atol(*++argv); 
   if (order < 0){
-    printf("ERROR: Matrix Order must be greater than 0 : %d \n", order);
+    printf("ERROR: Matrix Order must be greater than 0 : %ld \n", order);
     exit(EXIT_FAILURE);
   }
 
@@ -126,8 +127,8 @@ int main(int argc, char ** argv) {
 
   bytes = 2.0 * sizeof(double) * order * order;
 
-  printf("Matrix order          = %d\n", order);
-  if (tile_size < order) printf("Tile size             = %d\n", tile_size);
+  printf("Matrix order          = %ld\n", order);
+  if (tile_size < order) printf("Tile size             = %ld\n", tile_size);
   else                   printf("Untiled\n");
   printf("Number of iterations  = %d\n", iterations);
 
@@ -143,6 +144,8 @@ int main(int argc, char ** argv) {
     B(i,j) = 0.0;
   }
 
+  trans_time = 0.0; /* silence compiler warning */
+
   for (iter = 0; iter<=iterations; iter++){
 
     /* start timer after a warmup iteration                                        */