A better implementation for the Range/NdRange class, with examples

numba_dpex/core/kernel_interface/dispatcher.py

@@ -33,11 +33,9 @@
 )
 from numba_dpex.core.kernel_interface.arg_pack_unpacker import Packer
 from numba_dpex.core.kernel_interface.spirv_kernel import SpirvKernel
-from numba_dpex.core.kernel_interface.utils import Ranges
+from numba_dpex.core.kernel_interface.utils import NdRange
 from numba_dpex.core.types import USMNdArray
 
-simplefilter("always", DeprecationWarning)
-
 
 def get_ordered_arg_access_types(pyfunc, access_types):
     """Deprecated and to be removed in next release."""
@@ -445,56 +443,6 @@ def _determine_kernel_launch_queue(self, args, argtypes):
             else:
                 raise ExecutionQueueInferenceError(self.kernel_name)
 
-    def _raise_invalid_kernel_enqueue_args(self):
-        error_message = (
-            "Incorrect number of arguments for enqueuing numba_dpex.kernel. "
-            "Usage: device_env, global size, local size. "
-            "The local size argument is optional."
-        )
-        raise InvalidKernelLaunchArgsError(error_message)
-
-    def _ensure_valid_work_item_grid(self, val):
-        if not isinstance(val, (tuple, list, int)):
-            error_message = (
-                "Cannot create work item dimension from provided argument"
-            )
-            raise ValueError(error_message)
-
-        if isinstance(val, int):
-            val = [val]
-
-        # TODO: we need some way to check the max dimensions
-        """
-        if len(val) > device_env.get_max_work_item_dims():
-            error_message = ("Unsupported number of work item dimensions ")
-            raise ValueError(error_message)
-        """
-
-        return list(
-            val[::-1]
-        )  # reversing due to sycl and opencl interop kernel range mismatch semantic
-
-    def _ensure_valid_work_group_size(self, val, work_item_grid):
-        if not isinstance(val, (tuple, list, int)):
-            error_message = (
-                "Cannot create work item dimension from provided argument"
-            )
-            raise ValueError(error_message)
-
-        if isinstance(val, int):
-            val = [val]
-
-        if len(val) != len(work_item_grid):
-            error_message = (
-                "Unsupported number of work item dimensions, "
-                + "dimensions of global and local work items has to be the same "
-            )
-            raise IllegalRangeValueError(error_message)
-
-        return list(
-            val[::-1]
-        )  # reversing due to sycl and opencl interop kernel range mismatch semantic
-
     def __getitem__(self, args):
         """Mimic's ``numba.cuda`` square-bracket notation for configuring the
         global_range and local_range settings when launching a kernel on a
@@ -523,9 +471,11 @@ def __getitem__(self, args):
 
         """
 
-        if isinstance(args, Ranges):
+        if isinstance(args, NdRange):
             self._global_range = list(args.global_range)[::-1]
-            self._local_range = list(args.local_range)[::-1]
+            self._local_range = (
+                list(args.local_range)[::-1] if args.local_range else None
+            )
         else:
             if (
                 isinstance(args, tuple)
@@ -540,38 +490,44 @@ def __getitem__(self, args):
                     + "i.e. <function>[(M,N), numba_dpex.DEFAULT_LOCAL_RANGE](<params>), "
                     + "otherwise your code might produce erroneous results.",
                     DeprecationWarning,
+                    stacklevel=2,
                 )
                 self._global_range = [args[0]]
                 self._local_range = [args[1]]
                 return self
 
-            if not isinstance(args, Iterable):
-                args = [args]
-
-            ls = None
+            args = [args] if not isinstance(args, Iterable) else args
             nargs = len(args)
+
             # Check if the kernel enquing arguments are sane
             if nargs < 1 or nargs > 2:
-                self._raise_invalid_kernel_enqueue_args()
+                raise InvalidKernelLaunchArgsError(kernel_name=self.kernel_name)
 
-            gs = self._ensure_valid_work_item_grid(args[0])
+            g_range = (
+                [args[0]] if not isinstance(args[0], Iterable) else args[0]
+            )
             # If the optional local size argument is provided
+            l_range = None
             if nargs == 2:
                 if args[1] != []:
-                    ls = self._ensure_valid_work_group_size(args[1], gs)
+                    l_range = (
+                        [args[1]]
+                        if not isinstance(args[1], Iterable)
+                        else args[1]
+                    )
                 else:
                     warn(
                         "Empty local_range calls will be deprecated in the future.",
                         DeprecationWarning,
+                        stacklevel=2,
                     )
 
-            self._global_range = list(gs)[::-1]
-            self._local_range = list(ls)[::-1] if ls else None
+            if len(g_range) < 1:
+                raise IllegalRangeValueError(kernel_name=self.kernel_name)
+
+            self._global_range = list(g_range)[::-1]
+            self._local_range = list(l_range)[::-1] if l_range else None
 
-            if self._global_range == [] and self._local_range is None:
-                raise IllegalRangeValueError(
-                    "Illegal range values for kernel launch parameters."
-                )
         return self
 
     def _check_ranges(self, device):