diff --git a/demos/ARM7-LPC214x-GCC/main.c b/demos/ARM7-LPC214x-GCC/main.c
index 10cae19da..d83a386d5 100644
--- a/demos/ARM7-LPC214x-GCC/main.c
+++ b/demos/ARM7-LPC214x-GCC/main.c
@@ -131,8 +131,8 @@ int main(int argc, char **argv) {
    * are not started in order to make accurate benchmarks.
    */
   if ((IO0PIN & 0x00018000) == 0x00018000) {
-    chThdCreate(NORMALPRIO, 0, waThread1, sizeof(waThread1), Thread1, NULL);
-    chThdCreate(NORMALPRIO, 0, waThread2, sizeof(waThread2), Thread2, NULL);
+    chThdCreateFast(NORMALPRIO, waThread1, sizeof(waThread1), Thread1);
+    chThdCreateFast(NORMALPRIO, waThread2, sizeof(waThread2), Thread2);
   }
 
   /*
diff --git a/readme.txt b/readme.txt
index 51f0be398..5fbee71f4 100644
--- a/readme.txt
+++ b/readme.txt
@@ -75,10 +75,14 @@ Win32-MinGW            - ChibiOS/RT simulator and demo into a WIN32 process,
 *****************************************************************************
 
 *** 0.6.7 ***
+- NEW: New chThdCreateFast() API, it is a simplified form of chThdCreate()
+       that allows even faster threads creation. The new API does not support
+       the "mode" and "arg" parameters (still available in the old API).
 - OPT: Removed an unrequired initialization from the chThdCreate().
 - OPT: Improvements to the test framework, now a virtual timer is used instead
   of software loops into the bechmarks in order to have more stable results.
 - Added the C++ wrapper entries to the documentation.
+- Fixed the documentation entry for the chThdCreate() API.
 
 *** 0.6.6 ***
 - NEW: Improved test suite, now the suite is divided in modules and the code
diff --git a/src/chinit.c b/src/chinit.c
index ec7626e6a..5b2c0cd24 100644
--- a/src/chinit.c
+++ b/src/chinit.c
@@ -55,7 +55,8 @@ void chSysInit(void) {
    * serve interrupts in its context while keeping the lowest energy saving
    * mode compatible with the system status.
    */
-  chThdCreate(IDLEPRIO, 0, waIdleThread, sizeof(waIdleThread), (tfunc_t)_IdleThread, NULL);
+  chThdCreateFast(IDLEPRIO, waIdleThread,
+                  sizeof(waIdleThread), (tfunc_t)_IdleThread);
 }
 
 /**
diff --git a/src/chthreads.c b/src/chthreads.c
index 33c75dd45..d12bced10 100644
--- a/src/chthreads.c
+++ b/src/chthreads.c
@@ -62,9 +62,8 @@ static void memfill(uint8_t *p, uint32_t n, uint8_t v) {
 /**
  * Creates a new thread.
  * @param prio the priority level for the new thread. Usually the threads are
- *             created with priority \p NORMALPRIO (128), priorities
- *             can range from \p LOWPRIO (1) to \p HIGHPRIO
- *             (255).
+ *             created with priority \p NORMALPRIO, priorities
+ *             can range from \p LOWPRIO to \p HIGHPRIO.
  * @param mode the creation option flags for the thread. The following options
  *             can be OR'ed in this parameter:<br>
  *             <ul>
@@ -106,9 +105,7 @@ Thread *chThdCreate(tprio_t prio, tmode_t mode, void *workspace,
   else {
 #endif
     chSysLock();
-
     chSchWakeupS(tp, RDY_OK);
-
     chSysUnlock();
 #ifdef CH_USE_RESUME
   }
@@ -116,6 +113,38 @@ Thread *chThdCreate(tprio_t prio, tmode_t mode, void *workspace,
   return tp;
 }
 
+/**
+ * Creates a new thread.
+ * @param prio the priority level for the new thread. Usually the threads are
+ *             created with priority \p NORMALPRIO, priorities
+ *             can range from \p LOWPRIO to \p HIGHPRIO.
+ * @param workspace pointer to a working area dedicated to the thread stack
+ * @param wsize size of the working area.
+ * @param pf the thread function. Returning from this function automatically
+ *           terminates the thread.
+ * @return the pointer to the \p Thread structure allocated for the
+ *         thread into the working space area.
+ * @note A thread can terminate by calling \p chThdExit() or by simply
+ *       returning from its main function.
+ */
+Thread *chThdCreateFast(tprio_t prio, void *workspace,
+                        size_t wsize, tfunc_t pf) {
+  Thread *tp = workspace;
+
+  chDbgAssert((wsize >= UserStackSize(0)) && (prio <= HIGHPRIO) &&
+              (workspace != NULL) && (pf != NULL),
+              "chthreads.c, chThdCreateFast()");
+#ifdef CH_USE_DEBUG
+  memfill(workspace, wsize, MEM_FILL_PATTERN);
+#endif
+  init_thread(prio, 0, tp);
+  SETUP_CONTEXT(workspace, wsize, pf, NULL);
+  chSysLock();
+  chSchWakeupS(tp, RDY_OK);
+  chSysUnlock();
+  return tp;
+}
+
 /**
  * Changes the thread priority, reschedules if necessary.
  * @param newprio the new priority of the invoking thread
diff --git a/src/include/threads.h b/src/include/threads.h
index 297c56a68..5bfd51423 100644
--- a/src/include/threads.h
+++ b/src/include/threads.h
@@ -170,6 +170,8 @@ extern "C" {
 #endif
   Thread *chThdCreate(tprio_t prio, tmode_t mode, void *workspace,
                       size_t wsize, tfunc_t pf, void *arg);
+  Thread *chThdCreateFast(tprio_t prio, void *workspace,
+                          size_t wsize, tfunc_t pf);
   void chThdSetPriority(tprio_t newprio);
   void chThdExit(msg_t msg);
 #ifdef CH_USE_RESUME
diff --git a/test/testbmk.c b/test/testbmk.c
index 8631138d1..a75ccddc2 100644
--- a/test/testbmk.c
+++ b/test/testbmk.c
@@ -61,7 +61,7 @@ static void bmk1_teardown(void) {
 static void bmk1_execute(void) {
   uint32_t n;
 
-  threads[0] = chThdCreate(chThdGetPriority()-1, 0, wa[0], STKSIZE, thread1, 0);
+  threads[0] = chThdCreateFast(chThdGetPriority()-1, wa[0], STKSIZE, thread1);
   n = msg_loop_test(threads[0]);
   chThdTerminate(threads[0]);
   test_wait_threads();
@@ -93,7 +93,7 @@ static void bmk2_teardown(void) {
 static void bmk2_execute(void) {
   uint32_t n;
 
-  threads[0] = chThdCreate(chThdGetPriority()+1, 0, wa[0], STKSIZE, thread1, 0);
+  threads[0] = chThdCreateFast(chThdGetPriority()+1, wa[0], STKSIZE, thread1);
   n = msg_loop_test(threads[0]);
   chThdTerminate(threads[0]);
   test_wait_threads();
@@ -130,11 +130,11 @@ static void bmk3_teardown(void) {
 static void bmk3_execute(void) {
   uint32_t n;
 
-  threads[0] = chThdCreate(chThdGetPriority()+1, 0, wa[0], STKSIZE, thread1, 0);
-  threads[1] = chThdCreate(chThdGetPriority()-2, 0, wa[1], STKSIZE, thread2, 0);
-  threads[2] = chThdCreate(chThdGetPriority()-3, 0, wa[2], STKSIZE, thread2, 0);
-  threads[3] = chThdCreate(chThdGetPriority()-4, 0, wa[3], STKSIZE, thread2, 0);
-  threads[4] = chThdCreate(chThdGetPriority()-5, 0, wa[4], STKSIZE, thread2, 0);
+  threads[0] = chThdCreateFast(chThdGetPriority()+1, wa[0], STKSIZE, thread1);
+  threads[1] = chThdCreateFast(chThdGetPriority()-2, wa[1], STKSIZE, thread2);
+  threads[2] = chThdCreateFast(chThdGetPriority()-3, wa[2], STKSIZE, thread2);
+  threads[3] = chThdCreateFast(chThdGetPriority()-4, wa[3], STKSIZE, thread2);
+  threads[4] = chThdCreateFast(chThdGetPriority()-5, wa[4], STKSIZE, thread2);
   n = msg_loop_test(threads[0]);
   chThdTerminate(threads[0]);
   test_wait_threads();
@@ -171,7 +171,7 @@ static void bmk4_execute(void) {
   test_wait_tick();
   test_start_timer(1000);
   do {
-    chThdWait(chThdCreate(prio, 0, wap, STKSIZE, thread2, NULL));
+    chThdWait(chThdCreateFast(prio, wap, STKSIZE, thread2));
     n++;
 #if defined(WIN32)
     ChkIntSources();