Introduce dpctl.SyclQueue.submit_async

The SyclQueue.submit has become synchronosing, although
it still returns a SyclEvent (with exectuion_status always complete)

Author: oleksandr-pavlyk

dpctl/_sycl_queue.pxd

@@ -75,6 +75,14 @@ cdef public api class SyclQueue (_SyclQueue) [
         object args,
         list dEvents
     )
+    cpdef SyclEvent submit_async(
+        self,
+        SyclKernel kernel,
+        list args,
+        list gS,
+        list lS=*,
+        list dEvents=*
+    )
     cpdef SyclEvent submit(
         self,
         SyclKernel kernel,

dpctl/_sycl_queue.pyx

@@ -776,7 +776,7 @@ cdef class SyclQueue(_SyclQueue):
         return SyclEvent._create(htERef)
 
 
-    cpdef SyclEvent submit(
+    cpdef SyclEvent submit_async(
         self,
         SyclKernel kernel,
         list args,
@@ -908,6 +908,18 @@ cdef class SyclQueue(_SyclQueue):
 
         return SyclEvent._create(Eref)
 
+    cpdef SyclEvent submit(
+        self,
+        SyclKernel kernel,
+        list args,
+        list gS,
+        list lS=None,
+        list dEvents=None
+    ):
+        cdef SyclEvent e = self.submit_async(kernel, args, gS, lS, dEvents)
+        e.wait()
+        return e
+
     cpdef void wait(self):
         with nogil: DPCTLQueue_Wait(self._queue_ref)
 

dpctl/tests/test_sycl_kernel_submit.py

@@ -114,7 +114,7 @@ def test_create_program_from_source(ctype_str, dtype, ctypes_ctor):
         )
 
 
-def test_async_submit():
+def test_submit_async():
     try:
         q = dpctl.SyclQueue("opencl")
     except dpctl.SyclQueueCreationError:
@@ -182,7 +182,7 @@ def test_async_submit():
 
     async_detected = False
     for attempt in range(5):
-        e1 = q.submit(
+        e1 = q.submit_async(
             kern1Kernel,
             [
                 first_row,
@@ -192,7 +192,7 @@ def test_async_submit():
                 n,
             ],
         )
-        e2 = q.submit(
+        e2 = q.submit_async(
             kern2Kernel,
             [
                 second_row,
@@ -202,7 +202,7 @@ def test_async_submit():
                 n,
             ],
         )
-        e3 = q.submit(
+        e3 = q.submit_async(
             kern3Kernel,
             [third_row, first_row, second_row],
             [