Numpy actx: cache execuctor

Author: inducer

arraycontext/context.py

@@ -339,7 +339,7 @@ def to_numpy(self,
 
     @abstractmethod
     def call_loopy(self,
-                   program: "loopy.TranslationUnit",
+                   t_unit: "loopy.TranslationUnit",
                    **kwargs: Any) -> Dict[str, Array]:
         """Execute the :mod:`loopy` program *program* on the arguments
         *kwargs*.

arraycontext/impl/numpy/__init__.py

@@ -1,3 +1,6 @@
+from __future__ import annotations
+
+
 """
 .. currentmodule:: arraycontext
 
@@ -30,6 +33,7 @@
 THE SOFTWARE.
 """
 
+from collections.abc import Mapping
 from typing import Any, Dict
 
 import numpy as np
@@ -39,6 +43,7 @@
 
 from arraycontext.container.traversal import rec_map_array_container, with_array_context
 from arraycontext.context import (
+    Array,
     ArrayContext,
     ArrayOrContainerOrScalar,
     ArrayOrContainerOrScalarT,
@@ -62,10 +67,12 @@ class NumpyArrayContext(ArrayContext):
 
     .. automethod:: __init__
     """
+
+    _loopy_transform_cache: dict[lp.TranslationUnit, lp.ExecutorBase]
+
     def __init__(self) -> None:
         super().__init__()
-        self._loopy_transform_cache: \
-                Dict[lp.TranslationUnit, lp.TranslationUnit] = {}
+        self._loopy_transform_cache = {}
 
     array_types = (NumpyNonObjectArray,)
 
@@ -88,17 +95,18 @@ def to_numpy(self,
                  ) -> NumpyOrContainerOrScalar:
         return array
 
-    def call_loopy(self, t_unit, **kwargs):
+    def call_loopy(
+                self,
+                t_unit: lp.TranslationUnit, **kwargs: Any
+            ) -> dict[str, Array]:
         t_unit = t_unit.copy(target=lp.ExecutableCTarget())
         try:
-            t_unit = self._loopy_transform_cache[t_unit]
+            executor = self._loopy_transform_cache[t_unit]
         except KeyError:
-            orig_t_unit = t_unit
-            t_unit = self.transform_loopy_program(t_unit)
-            self._loopy_transform_cache[orig_t_unit] = t_unit
-            del orig_t_unit
+            executor = self.transform_loopy_program(t_unit).executor()
+            self._loopy_transform_cache[t_unit] = executor
 
-        _, result = t_unit(**kwargs)
+        _, result = executor(**kwargs)
 
         return result