Cache numba stuff

.github/workflows/test.yml

@@ -193,7 +193,7 @@ jobs:
           else
             micromamba install --yes -q "python~=${PYTHON_VERSION}" mkl "numpy${NUMPY_VERSION}" scipy pip mkl-service graphviz cython pytest coverage pytest-cov pytest-benchmark pytest-mock;
           fi
-          if [[ $INSTALL_NUMBA == "1" ]]; then micromamba install --yes -q -c conda-forge "python~=${PYTHON_VERSION}" "numba>=0.57"; fi
+          micromamba install --yes -q -c conda-forge "python~=${PYTHON_VERSION}" "numba>=0.57"; fi
           if [[ $INSTALL_JAX == "1" ]]; then micromamba install --yes -q -c conda-forge "python~=${PYTHON_VERSION}" jax jaxlib numpyro && pip install tensorflow-probability; fi
           if [[ $INSTALL_TORCH == "1" ]]; then micromamba install --yes -q -c conda-forge "python~=${PYTHON_VERSION}" pytorch pytorch-cuda=12.1 "mkl<=2024.0" -c pytorch -c nvidia; fi
           pip install pytest-sphinx

pytensor/compile/mode.py

@@ -63,9 +63,8 @@ def register_linker(name, linker):
 # If a string is passed as the optimizer argument in the constructor
 # for Mode, it will be used as the key to retrieve the real optimizer
 # in this dictionary
-exclude = []
-if not config.cxx:
-    exclude = ["cxx_only"]
+
+exclude = ["cxx_only", "BlasOpt"]
 OPT_NONE = RewriteDatabaseQuery(include=[], exclude=exclude)
 # Even if multiple merge optimizer call will be there, this shouldn't
 # impact performance.
@@ -346,6 +345,11 @@ def __setstate__(self, state):
             optimizer = predefined_optimizers[optimizer]
         if isinstance(optimizer, RewriteDatabaseQuery):
             self.provided_optimizer = optimizer
+
+        # Force numba-required rewrites if using NumbaLinker
+        if isinstance(linker, NumbaLinker):
+            optimizer = optimizer.including("numba")
+
         self._optimizer = optimizer
         self.call_time = 0
         self.fn_time = 0
@@ -443,16 +447,20 @@ def clone(self, link_kwargs=None, optimizer="", **kwargs):
 # string as the key
 # Use VM_linker to allow lazy evaluation by default.
 FAST_COMPILE = Mode(
-    VMLinker(use_cloop=False, c_thunks=False),
-    RewriteDatabaseQuery(include=["fast_compile", "py_only"]),
+    NumbaLinker(),
+    # TODO: Fast_compile should just use python code, CHANGE ME!
+    RewriteDatabaseQuery(
+        include=["fast_compile", "numba"],
+        exclude=["cxx_only", "BlasOpt", "local_careduce_fusion"],
+    ),
+)
+FAST_RUN = Mode(
+    NumbaLinker(),
+    RewriteDatabaseQuery(
+        include=["fast_run", "numba"],
+        exclude=["cxx_only", "BlasOpt", "local_careduce_fusion"],
+    ),
 )
-if config.cxx:
-    FAST_RUN = Mode("cvm", "fast_run")
-else:
-    FAST_RUN = Mode(
-        "vm",
-        RewriteDatabaseQuery(include=["fast_run", "py_only"]),
-    )
 
 NUMBA = Mode(
     NumbaLinker(),
@@ -565,6 +573,7 @@ def register_mode(name, mode):
     Add a `Mode` which can be referred to by `name` in `function`.
 
     """
+    # TODO: Remove me
     if name in predefined_modes:
         raise ValueError(f"Mode name already taken: {name}")
     predefined_modes[name] = mode

pytensor/configdefaults.py

@@ -370,11 +370,21 @@ def add_compile_configvars():
 
     if rc == 0 and config.cxx != "":
         # Keep the default linker the same as the one for the mode FAST_RUN
-        linker_options = ["c|py", "py", "c", "c|py_nogc", "vm", "vm_nogc", "cvm_nogc"]
+        linker_options = [
+            "cvm",
+            "c|py",
+            "py",
+            "c",
+            "c|py_nogc",
+            "vm",
+            "vm_nogc",
+            "cvm_nogc",
+            "jax",
+        ]
     else:
         # g++ is not present or the user disabled it,
         # linker should default to python only.
-        linker_options = ["py", "vm_nogc"]
+        linker_options = ["py", "vm", "vm_nogc", "jax"]
         if type(config).cxx.is_default:
             # If the user provided an empty value for cxx, do not warn.
             _logger.warning(
@@ -388,7 +398,7 @@ def add_compile_configvars():
         "linker",
         "Default linker used if the pytensor flags mode is Mode",
         # Not mutable because the default mode is cached after the first use.
-        EnumStr("cvm", linker_options, mutable=False),
+        EnumStr("numba", linker_options, mutable=False),
         in_c_key=False,
     )
 

pytensor/link/jax/linker.py

@@ -87,3 +87,6 @@ def create_thunk_inputs(self, storage_map):
             thunk_inputs.append(sinput)
 
         return thunk_inputs
+
+    def __repr__(self):
+        return "JAXLinker()"

pytensor/link/numba/dispatch/basic.py

@@ -40,6 +40,7 @@
 from pytensor.tensor.slinalg import Solve
 from pytensor.tensor.type import TensorType
 from pytensor.tensor.type_other import MakeSlice, NoneConst
+from pytensor.typed_list import TypedListType
 
 
 def global_numba_func(func):
@@ -135,6 +136,8 @@ def get_numba_type(
             return CSCMatrixType(numba_dtype)
 
         raise NotImplementedError()
+    elif isinstance(pytensor_type, TypedListType):
+        return numba.types.List(get_numba_type(pytensor_type.ttype))
     else:
         raise NotImplementedError(f"Numba type not implemented for {pytensor_type}")
 
@@ -481,11 +484,11 @@ def numba_funcify_SpecifyShape(op, node, **kwargs):
     shape_input_names = ["shape_" + str(i) for i in range(len(shape_inputs))]
 
     func_conditions = [
-        f"assert x.shape[{i}] == {shape_input_names}"
-        for i, (shape_input, shape_input_names) in enumerate(
+        f"assert x.shape[{i}] == {eval_dim_name}, f'SpecifyShape: dim {{{i}}} of input has shape {{x.shape[{i}]}}, expected {{{eval_dim_name}.item()}}.'"
+        for i, (node_dim_input, eval_dim_name) in enumerate(
             zip(shape_inputs, shape_input_names, strict=True)
         )
-        if shape_input is not NoneConst
+        if node_dim_input is not NoneConst
     ]
 
     func = dedent(

pytensor/link/numba/dispatch/elemwise.py

@@ -264,6 +264,12 @@ def axis_apply_fn(x):
 
 @numba_funcify.register(Elemwise)
 def numba_funcify_Elemwise(op, node, **kwargs):
+    # op = getattr(np, str(op.scalar_op).lower())
+    # @numba_njit
+    # def elemwise_is_numpy(x):
+    #     return op(x)
+    # return elemwise_is_numpy
+
     scalar_inputs = [get_scalar_type(dtype=input.dtype)() for input in node.inputs]
     scalar_node = op.scalar_op.make_node(*scalar_inputs)
 
@@ -276,7 +282,7 @@ def numba_funcify_Elemwise(op, node, **kwargs):
 
     nin = len(node.inputs)
     nout = len(node.outputs)
-    core_op_fn = store_core_outputs(scalar_op_fn, nin=nin, nout=nout)
+    core_op_fn = store_core_outputs(scalar_op_fn, op.scalar_op, nin=nin, nout=nout)
 
     input_bc_patterns = tuple(inp.type.broadcastable for inp in node.inputs)
     output_bc_patterns = tuple(out.type.broadcastable for out in node.outputs)

pytensor/link/numba/dispatch/scalar.py

@@ -23,6 +23,7 @@
     Composite,
     Identity,
     Mul,
+    Pow,
     Reciprocal,
     ScalarOp,
     Second,
@@ -154,6 +155,21 @@ def numba_funcify_Switch(op, node, **kwargs):
     return numba_basic.global_numba_func(switch)
 
 
+@numba_funcify.register(Pow)
+def numba_funcify_Pow(op, node, **kwargs):
+    pow_dtype = node.inputs[1].type.dtype
+
+    def pow(x, y):
+        return x**y
+
+    # Work-around https://github.com/numba/numba/issues/9554
+    # fast-math casuse kernel crash
+    patch_kwargs = {}
+    if pow_dtype.startswith("int"):
+        patch_kwargs["fastmath"] = False
+    return numba_basic.numba_njit(**patch_kwargs)(pow)
+
+
 def binary_to_nary_func(inputs: list[Variable], binary_op_name: str, binary_op: str):
     """Create a Numba-compatible N-ary function from a binary function."""
     unique_names = unique_name_generator(["binary_op_name"], suffix_sep="_")

pytensor/link/numba/dispatch/tensor_basic.py

@@ -68,7 +68,7 @@ def numba_funcify_Alloc(op, node, **kwargs):
     shape_var_item_names = [f"{name}_item" for name in shape_var_names]
     shapes_to_items_src = indent(
         "\n".join(
-            f"{item_name} = to_scalar({shape_name})"
+            f"{item_name} = {shape_name}.item()"
             for item_name, shape_name in zip(
                 shape_var_item_names, shape_var_names, strict=True
             )
@@ -86,7 +86,7 @@ def numba_funcify_Alloc(op, node, **kwargs):
 
     alloc_def_src = f"""
 def alloc(val, {", ".join(shape_var_names)}):
-    val_np = np.asarray(val)
+    val_np = val
 {shapes_to_items_src}
     scalar_shape = {create_tuple_string(shape_var_item_names)}
 {check_runtime_broadcast_src}

pytensor/link/numba/dispatch/vectorize_codegen.py

@@ -3,6 +3,7 @@
 import base64
 import pickle
 from collections.abc import Callable, Sequence
+from hashlib import sha256
 from textwrap import indent
 from typing import Any, cast
 
@@ -15,15 +16,19 @@
 from numba.core.types.misc import NoneType
 from numba.np import arrayobj
 
+from pytensor.graph.op import HasInnerGraph
 from pytensor.link.numba.dispatch import basic as numba_basic
-from pytensor.link.utils import compile_function_src
+from pytensor.link.numba.super_utils import compile_function_src2
+from pytensor.scalar import ScalarOp
 
 
 def encode_literals(literals: Sequence) -> str:
     return base64.encodebytes(pickle.dumps(literals)).decode()
 
 
-def store_core_outputs(core_op_fn: Callable, nin: int, nout: int) -> Callable:
+def store_core_outputs(
+    core_op_fn: Callable, core_op: ScalarOp, nin: int, nout: int
+) -> Callable:
     """Create a Numba function that wraps a core function and stores its vectorized outputs.
 
     @njit
@@ -52,9 +57,14 @@ def store_core_outputs({inp_signature}, {out_signature}):
 {indent(store_outputs, " " * 4)}
 """
     global_env = {"core_op_fn": core_op_fn}
-    func = compile_function_src(
-        func_src, "store_core_outputs", {**globals(), **global_env}
-    )
+    # func = compile_function_src(
+    #     func_src, "store_core_outputs", {**globals(), **global_env},
+    # )
+    if isinstance(core_op, HasInnerGraph):
+        key = sha256(core_op.c_code_template.encode()).hexdigest()
+    else:
+        key = str(core_op)
+    func = compile_function_src2(key, func_src, "store_core_outputs", global_env)
     return cast(Callable, numba_basic.numba_njit(func))
 
 

pytensor/link/numba/linker.py

@@ -35,3 +35,6 @@ def create_thunk_inputs(self, storage_map):
             thunk_inputs.append(sinput)
 
         return thunk_inputs
+
+    def __repr__(self):
+        return "NumbaLinker()"