diff --git a/Makefile.rule b/Makefile.rule
index 17815096e..255d1da46 100644
--- a/Makefile.rule
+++ b/Makefile.rule
@@ -58,6 +58,12 @@ VERSION = 0.3.7.dev
 # For force setting for multi  threaded, specify USE_THREAD = 1
 # USE_THREAD = 0
 
+# If you want to build a single-threaded OpenBLAS, but expect to call this
+# from several concurrent threads in some other program, comment this in for
+# thread safety. (This is done automatically for USE_THREAD=1 , and should not
+# be necessary when USE_OPENMP=1)
+# USE_LOCKING = 1
+
 # If you're going to use this library with OpenMP, please comment it in.
 # This flag is always set for POWER8. Don't set USE_OPENMP = 0 if you're targeting POWER8.
 # USE_OPENMP = 1
diff --git a/Makefile.system b/Makefile.system
index a95d6190f..f574edf88 100644
--- a/Makefile.system
+++ b/Makefile.system
@@ -237,6 +237,10 @@ SMP = 1
 endif
 endif
 
+ifeq ($(SMP), 1)
+USE_LOCKING = 
+endif
+
 ifndef NEED_PIC
 NEED_PIC = 1
 endif
@@ -388,6 +392,12 @@ ifneq ($(MAX_STACK_ALLOC), 0)
 CCOMMON_OPT	+= -DMAX_STACK_ALLOC=$(MAX_STACK_ALLOC)
 endif
 
+ifdef USE_LOCKING
+ifneq ($(USE_LOCKING), 0)
+CCOMMON_OPT	+= -DUSE_LOCKING
+endif
+endif
+
 #
 #  Architecture dependent settings
 #
diff --git a/cmake/system.cmake b/cmake/system.cmake
index d0f560872..adedd32cc 100644
--- a/cmake/system.cmake
+++ b/cmake/system.cmake
@@ -136,6 +136,10 @@ endif ()
 
 if (USE_THREAD)
   message(STATUS "Multi-threading enabled with ${NUM_THREADS} threads.")
+else()
+  if (${USE_LOCKING})
+    set(CCOMMON_OPT "${CCOMMON_OPT} -DUSE_LOCKING")
+  endif ()
 endif ()
 
 include("${PROJECT_SOURCE_DIR}/cmake/prebuild.cmake")
diff --git a/common.h b/common.h
index 0ac74bb20..a9fe8d911 100644
--- a/common.h
+++ b/common.h
@@ -131,7 +131,7 @@ extern "C" {
 #include <time.h>
 #include <unistd.h>
 #include <math.h>
-#ifdef SMP
+#if defined(SMP) || defined(USE_LOCKING)
 #include <pthread.h>
 #endif
 #endif
@@ -200,7 +200,7 @@ extern "C" {
 #error "You can't specify both LOCK operation!"
 #endif
 
-#ifdef SMP
+#if defined(SMP) || defined(USE_LOCKING)
 #define USE_PTHREAD_LOCK
 #undef	USE_PTHREAD_SPINLOCK
 #endif
diff --git a/driver/others/memory.c b/driver/others/memory.c
index 02352b3ae..adb1ec86c 100644
--- a/driver/others/memory.c
+++ b/driver/others/memory.c
@@ -2062,13 +2062,13 @@ static void *alloc_mmap(void *address){
   }
 
   if (map_address != (void *)-1) {
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
     LOCK_COMMAND(&alloc_lock);
 #endif    
     release_info[release_pos].address = map_address;
     release_info[release_pos].func    = alloc_mmap_free;
     release_pos ++;
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
     UNLOCK_COMMAND(&alloc_lock);
 #endif    
   }
@@ -2214,13 +2214,13 @@ static void *alloc_mmap(void *address){
 #endif
 
   if (map_address != (void *)-1) {
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
     LOCK_COMMAND(&alloc_lock);
 #endif
     release_info[release_pos].address = map_address;
     release_info[release_pos].func    = alloc_mmap_free;
     release_pos ++;
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
     UNLOCK_COMMAND(&alloc_lock);
 #endif
   }
@@ -2701,7 +2701,7 @@ void *blas_memory_alloc(int procpos){
 
   position = 0;
 
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
   LOCK_COMMAND(&alloc_lock);
 #endif
   do {
@@ -2718,7 +2718,7 @@ void *blas_memory_alloc(int procpos){
     position ++;
 
   } while (position < NUM_BUFFERS);
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
   UNLOCK_COMMAND(&alloc_lock);	
 #endif
   goto error;
@@ -2730,7 +2730,7 @@ void *blas_memory_alloc(int procpos){
 #endif
 
   memory[position].used = 1;
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
   UNLOCK_COMMAND(&alloc_lock);
 #else
   blas_unlock(&memory[position].lock);	
@@ -2779,11 +2779,11 @@ void *blas_memory_alloc(int procpos){
 
     } while ((BLASLONG)map_address == -1);
 
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
     LOCK_COMMAND(&alloc_lock);
 #endif    
     memory[position].addr = map_address;
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
     UNLOCK_COMMAND(&alloc_lock);
 #endif
 
@@ -2839,7 +2839,7 @@ void blas_memory_free(void *free_area){
 #endif
 
   position = 0;
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
   LOCK_COMMAND(&alloc_lock);
 #endif
   while ((position < NUM_BUFFERS) && (memory[position].addr != free_area))
@@ -2855,7 +2855,7 @@ void blas_memory_free(void *free_area){
   WMB;
 
   memory[position].used = 0;
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
   UNLOCK_COMMAND(&alloc_lock);
 #endif
 
@@ -2872,7 +2872,7 @@ void blas_memory_free(void *free_area){
   for (position = 0; position < NUM_BUFFERS; position++)
     printf("%4ld  %p : %d\n", position, memory[position].addr, memory[position].used);
 #endif
-#if defined(SMP) && !defined(USE_OPENMP)
+#if (defined(SMP) || defined(USE_LOCKING)) && !defined(USE_OPENMP)
   UNLOCK_COMMAND(&alloc_lock);
 #endif
   return;
@@ -2924,7 +2924,7 @@ void blas_shutdown(void){
 
 #if defined(OS_LINUX) && !defined(NO_WARMUP)
 
-#ifdef SMP
+#if defined(SMP) || defined(USE_LOCKING)
 #if   defined(USE_PTHREAD_LOCK)
 static pthread_mutex_t    init_lock = PTHREAD_MUTEX_INITIALIZER;
 #elif defined(USE_PTHREAD_SPINLOCK)
@@ -2949,7 +2949,7 @@ static void _touch_memory(blas_arg_t *arg, BLASLONG *range_m, BLASLONG *range_n,
     if (hot_alloc != 2) {
 #endif
 
-#ifdef SMP
+#if defined(SMP) || defined(USE_LOCKING)
   LOCK_COMMAND(&init_lock);
 #endif
 
@@ -2959,7 +2959,7 @@ static void _touch_memory(blas_arg_t *arg, BLASLONG *range_m, BLASLONG *range_n,
     size    -= PAGESIZE;
   }
 
-#ifdef SMP
+#if defined(SMP) || defined(USE_LOCKING)
   UNLOCK_COMMAND(&init_lock);
 #endif