warp-lang 1.0.2__py3-none-win_amd64.whl → 1.2.0__py3-none-win_amd64.whl

warp/jax_experimental.py

@@ -1,339 +1,341 @@
-# Copyright (c) 2024 NVIDIA CORPORATION.  All rights reserved.
-# NVIDIA CORPORATION and its licensors retain all intellectual property
-# and proprietary rights in and to this software, related documentation
-# and any modifications thereto.  Any use, reproduction, disclosure or
-# distribution of this software and related documentation without an express
-# license agreement from NVIDIA CORPORATION is strictly prohibited.
-
-import ctypes
-import warp as wp
-from warp.types import array_t, launch_bounds_t, strides_from_shape
-from warp.context import type_str
-import jax
-import jax.numpy as jp
-
-_jax_warp_p = None
-
-# Holder for the custom callback to keep it alive.
-_cc_callback = None
-_registered_kernels = [None]
-_registered_kernel_to_id = {}
-
-
-def jax_kernel(wp_kernel):
-    """Create a Jax primitive from a Warp kernel.
-
-    NOTE: This is an experimental feature under development.
-
-    Current limitations:
-    - All kernel arguments must be arrays.
-    - Kernel launch dimensions are inferred from the shape of the first argument.
-    - Input arguments are followed by output arguments in the Warp kernel definition.
-    - There must be at least one input argument and at least one output argument.
-    - Output shapes must match the launch dimensions (i.e., output shapes must match the shape of the first argument).
-    - All arrays must be contiguous.
-    - Only the CUDA backend is supported.
-    """
-
-    if _jax_warp_p == None:
-        # Create and register the primitive
-        _create_jax_warp_primitive()
-    if not wp_kernel in _registered_kernel_to_id:
-        id = len(_registered_kernels)
-        _registered_kernels.append(wp_kernel)
-        _registered_kernel_to_id[wp_kernel] = id
-    else:
-        id = _registered_kernel_to_id[wp_kernel]
-
-    def bind(*args):
-        return _jax_warp_p.bind(*args, kernel=id)
-
-    return bind
-
-
-def _warp_custom_callback(stream, buffers, opaque, opaque_len):
-    # The descriptor is the form
-    # <kernel-id>|<launch-dims>|<arg-dims-list>
-    # Example:  42|16,32|16,32;100;16,32
-    kernel_id_str, dim_str, args_str = opaque.decode().split("|")
-
-    # Get the kernel from the registry.
-    kernel_id = int(kernel_id_str)
-    kernel = _registered_kernels[kernel_id]
-
-    # Parse launch dimensions.
-    dims = [int(d) for d in dim_str.split(",")]
-    bounds = launch_bounds_t(dims)
-
-    # Parse arguments.
-    arg_strings = args_str.split(";")
-    num_args = len(arg_strings)
-    assert num_args == len(kernel.adj.args), "Incorrect number of arguments"
-
-    # First param is the launch bounds.
-    kernel_params = (ctypes.c_void_p * (1 + num_args))()
-    kernel_params[0] = ctypes.addressof(bounds)
-
-    # Parse array descriptors.
-    args = []
-    for i in range(num_args):
-        dtype = kernel.adj.args[i].type.dtype
-        shape = [int(d) for d in arg_strings[i].split(",")]
-        strides = strides_from_shape(shape, dtype)
-
-        arr = array_t(buffers[i], 0, len(shape), shape, strides)
-        args.append(arr)  # keep a reference
-        arg_ptr = ctypes.addressof(arr)
-
-        kernel_params[i + 1] = arg_ptr
-
-    # Get current device.
-    device = wp.device_from_jax(_get_jax_device())
-
-    # Get kernel hooks.
-    # Note: module was loaded during jit lowering.
-    hooks = kernel.module.get_kernel_hooks(kernel, device)
-    assert hooks.forward, "Failed to find kernel entry point"
-
-    # Launch the kernel.
-    wp.context.runtime.core.cuda_launch_kernel(
-        device.context, hooks.forward, bounds.size, 0, kernel_params, stream
-    )
-
-
-# TODO: is there a simpler way of getting the Jax "current" device?
-def _get_jax_device():
-    # check if jax.default_device() context manager is active
-    device = jax.config.jax_default_device
-    # if default device is not set, use first device
-    if device is None:
-        device = jax.devices()[0]
-    return device
-
-
-def _create_jax_warp_primitive():
-    from functools import reduce
-    import jax
-    from jax._src.interpreters import batching
-    from jax.interpreters import mlir
-    from jax.interpreters.mlir import ir
-    from jaxlib.hlo_helpers import custom_call
-
-    global _jax_warp_p
-    global _cc_callback
-
-    # Create and register the primitive.
-    # TODO add default implementation that calls the kernel via warp.
-    _jax_warp_p = jax.core.Primitive("jax_warp")
-    _jax_warp_p.multiple_results = True
-
-    # TODO Just launch the kernel directly, but make sure the argument
-    # shapes are massaged the same way as below so that vmap works.
-    def impl(*args):
-        raise Exception("Not implemented")
-
-    _jax_warp_p.def_impl(impl)
-
-    # Auto-batching. Make sure all the arguments are fully broadcasted
-    # so that Warp is not confused about dimensions.
-    def vectorized_multi_batcher(args, dims, **params):
-        # Figure out the number of outputs.
-        wp_kernel = _registered_kernels[params["kernel"]]
-        output_count = len(wp_kernel.adj.args) - len(args)
-        shape, dim = next((a.shape, d) for a, d in zip(args, dims) if d is not None)
-        size = shape[dim]
-        args = [batching.bdim_at_front(a, d, size) if len(a.shape) else a for a, d in zip(args, dims)]
-        # Create the batched primitive.
-        return _jax_warp_p.bind(*args, **params), [dims[0]] * output_count
-
-    batching.primitive_batchers[_jax_warp_p] = vectorized_multi_batcher
-
-    def get_vecmat_shape(warp_type):
-        if hasattr(warp_type.dtype, "_shape_"):
-            return warp_type.dtype._shape_
-        return []
-
-    def strip_vecmat_dimensions(warp_arg, actual_shape):
-        shape = get_vecmat_shape(warp_arg.type)
-        for i, s in enumerate(reversed(shape)):
-            item = actual_shape[-i - 1]
-            if s != item:
-                raise Exception(f"The vector/matrix shape for argument {warp_arg.label} does not match")
-        return actual_shape[: len(actual_shape) - len(shape)]
-
-    def collapse_into_leading_dimension(warp_arg, actual_shape):
-        if len(actual_shape) < warp_arg.type.ndim:
-            raise Exception(f"Argument {warp_arg.label} has too few non-matrix/vector dimensions")
-        index_rest = len(actual_shape) - warp_arg.type.ndim + 1
-        leading_size = reduce(lambda x, y: x * y, actual_shape[:index_rest])
-        return [leading_size] + actual_shape[index_rest:]
-
-    # Infer array dimensions from input type.
-    def infer_dimensions(warp_arg, actual_shape):
-        actual_shape = strip_vecmat_dimensions(warp_arg, actual_shape)
-        return collapse_into_leading_dimension(warp_arg, actual_shape)
-
-    def base_type_to_jax(warp_dtype):
-        if hasattr(warp_dtype, "_wp_scalar_type_"):
-            return wp.dtype_to_jax(warp_dtype._wp_scalar_type_)
-        return wp.dtype_to_jax(warp_dtype)
-
-    def base_type_to_jax_ir(warp_dtype):
-        warp_to_jax_dict = {
-            wp.float16: ir.F16Type.get(),
-            wp.float32: ir.F32Type.get(),
-            wp.float64: ir.F64Type.get(),
-            wp.int8: ir.IntegerType.get_signless(8),
-            wp.int16: ir.IntegerType.get_signless(16),
-            wp.int32: ir.IntegerType.get_signless(32),
-            wp.int64: ir.IntegerType.get_signless(64),
-            wp.uint8: ir.IntegerType.get_unsigned(8),
-            wp.uint16: ir.IntegerType.get_unsigned(16),
-            wp.uint32: ir.IntegerType.get_unsigned(32),
-            wp.uint64: ir.IntegerType.get_unsigned(64),
-        }
-        if hasattr(warp_dtype, "_wp_scalar_type_"):
-            warp_dtype = warp_dtype._wp_scalar_type_
-        jax_dtype = warp_to_jax_dict.get(warp_dtype)
-        if jax_dtype is None:
-            raise TypeError(f"Invalid or unsupported data type: {warp_dtype}")
-        return jax_dtype
-
-    def base_type_is_compatible(warp_type, jax_ir_type):
-        jax_ir_to_warp = {
-            "f16": wp.float16,
-            "f32": wp.float32,
-            "f64": wp.float64,
-            "i8": wp.int8,
-            "i16": wp.int16,
-            "i32": wp.int32,
-            "i64": wp.int64,
-            "ui8": wp.uint8,
-            "ui16": wp.uint16,
-            "ui32": wp.uint32,
-            "ui64": wp.uint64,
-        }
-        expected_warp_type = jax_ir_to_warp.get(str(jax_ir_type))
-        if expected_warp_type is not None:
-            if hasattr(warp_type, "_wp_scalar_type_"):
-                return warp_type._wp_scalar_type_ == expected_warp_type
-            else:
-                return warp_type == expected_warp_type
-        else:
-            raise TypeError(f"Invalid or unsupported data type: {jax_ir_type}")
-
-    # Abstract evaluation.
-    def jax_warp_abstract(*args, kernel=None):
-        wp_kernel = _registered_kernels[kernel]
-        # All the extra arguments to the warp kernel are outputs.
-        warp_outputs = [o.type for o in wp_kernel.adj.args[len(args) :]]
-        # TODO. Let's just use the first input dimension to infer the output's dimensions.
-        dims = strip_vecmat_dimensions(wp_kernel.adj.args[0], list(args[0].shape))
-        jax_outputs = []
-        for o in warp_outputs:
-            shape = list(dims) + list(get_vecmat_shape(o))
-            dtype = base_type_to_jax(o.dtype)
-            jax_outputs.append(jax.core.ShapedArray(shape, dtype))
-        return jax_outputs
-
-    _jax_warp_p.def_abstract_eval(jax_warp_abstract)
-
-    # Lowering to MLIR.
-
-    # Create python-land custom call target.
-    CCALLFUNC = ctypes.CFUNCTYPE(
-        ctypes.c_voidp, ctypes.c_void_p, ctypes.POINTER(ctypes.c_void_p), ctypes.c_char_p, ctypes.c_size_t
-    )
-    _cc_callback = CCALLFUNC(_warp_custom_callback)
-    ccall_address = ctypes.cast(_cc_callback, ctypes.c_void_p)
-
-    # Put the custom call into a capsule, as required by XLA.
-    PyCapsule_Destructor = ctypes.CFUNCTYPE(None, ctypes.py_object)
-    PyCapsule_New = ctypes.pythonapi.PyCapsule_New
-    PyCapsule_New.restype = ctypes.py_object
-    PyCapsule_New.argtypes = (ctypes.c_void_p, ctypes.c_char_p, PyCapsule_Destructor)
-    capsule = PyCapsule_New(ccall_address.value, b"xla._CUSTOM_CALL_TARGET", PyCapsule_Destructor(0))
-
-    # Register the callback in XLA.
-    jax.lib.xla_client.register_custom_call_target("warp_call", capsule, platform="gpu")
-
-    def default_layout(shape):
-        return range(len(shape) - 1, -1, -1)
-
-    def warp_call_lowering(ctx, *args, kernel=None):
-        if not kernel:
-            raise Exception("Unknown kernel id " + str(kernel))
-        wp_kernel = _registered_kernels[kernel]
-
-        # TODO This may not be necessary, but it is perhaps better not to be
-        # mucking with kernel loading while already running the workload.
-        module = wp_kernel.module
-        device = wp.device_from_jax(_get_jax_device())
-        if not module.load(device):
-            raise Exception("Could not load kernel on device")
-
-        # Infer dimensions from the first input.
-        warp_arg0 = wp_kernel.adj.args[0]
-        actual_shape0 = ir.RankedTensorType(args[0].type).shape
-        dims = strip_vecmat_dimensions(warp_arg0, actual_shape0)
-        warp_dims = collapse_into_leading_dimension(warp_arg0, dims)
-
-        # Figure out the types and shapes of the input arrays.
-        arg_strings = []
-        operand_layouts = []
-        for actual, warg in zip(args, wp_kernel.adj.args):
-            wtype = warg.type
-            rtt = ir.RankedTensorType(actual.type)
-
-            if not isinstance(wtype, wp.array):
-                raise Exception("Only contiguous arrays are supported for Jax kernel arguments")
-
-            if not base_type_is_compatible(wtype.dtype, rtt.element_type):
-                raise TypeError(f"Incompatible data type for argument '{warg.label}', expected {type_str(wtype.dtype)}, got {rtt.element_type}")
-
-            # Infer array dimension (by removing the vector/matrix dimensions and
-            # collapsing the initial dimensions).
-            shape = infer_dimensions(warg, rtt.shape)
-
-            if len(shape) != wtype.ndim:
-                raise TypeError(f"Incompatible array dimensionality for argument '{warg.label}'")
-
-            arg_strings.append(",".join([str(d) for d in shape]))
-            operand_layouts.append(default_layout(rtt.shape))
-
-        # Figure out the types and shapes of the output arrays.
-        result_types = []
-        result_layouts = []
-        for warg in wp_kernel.adj.args[len(args) :]:
-            wtype = warg.type
-
-            if not isinstance(wtype, wp.array):
-                raise Exception("Only contiguous arrays are supported for Jax kernel arguments")
-
-            # Infer dimensions from the first input.
-            arg_strings.append(",".join([str(d) for d in warp_dims]))
-
-            result_shape = list(dims) + list(get_vecmat_shape(wtype))
-            result_types.append(ir.RankedTensorType.get(result_shape, base_type_to_jax_ir(wtype.dtype)))
-            result_layouts.append(default_layout(result_shape))
-
-        # Build opaque descriptor for callback.
-        shape_str = ",".join([str(d) for d in warp_dims])
-        args_str = ";".join(arg_strings)
-        descriptor = f"{kernel}|{shape_str}|{args_str}"
-
-        out = custom_call(
-            b"warp_call",
-            result_types=result_types,
-            operands=args,
-            backend_config=descriptor.encode("utf-8"),
-            operand_layouts=operand_layouts,
-            result_layouts=result_layouts,
-        ).results
-        return out
-
-    mlir.register_lowering(
-        _jax_warp_p,
-        warp_call_lowering,
-        platform="gpu",
-    )
+# Copyright (c) 2024 NVIDIA CORPORATION.  All rights reserved.
+# NVIDIA CORPORATION and its licensors retain all intellectual property
+# and proprietary rights in and to this software, related documentation
+# and any modifications thereto.  Any use, reproduction, disclosure or
+# distribution of this software and related documentation without an express
+# license agreement from NVIDIA CORPORATION is strictly prohibited.
+
+import ctypes
+
+import jax
+
+import warp as wp
+from warp.context import type_str
+from warp.types import array_t, launch_bounds_t, strides_from_shape
+
+_jax_warp_p = None
+
+# Holder for the custom callback to keep it alive.
+_cc_callback = None
+_registered_kernels = [None]
+_registered_kernel_to_id = {}
+
+
+def jax_kernel(wp_kernel):
+    """Create a Jax primitive from a Warp kernel.
+
+    NOTE: This is an experimental feature under development.
+
+    Current limitations:
+    - All kernel arguments must be arrays.
+    - Kernel launch dimensions are inferred from the shape of the first argument.
+    - Input arguments are followed by output arguments in the Warp kernel definition.
+    - There must be at least one input argument and at least one output argument.
+    - Output shapes must match the launch dimensions (i.e., output shapes must match the shape of the first argument).
+    - All arrays must be contiguous.
+    - Only the CUDA backend is supported.
+    """
+
+    if _jax_warp_p is None:
+        # Create and register the primitive
+        _create_jax_warp_primitive()
+    if wp_kernel not in _registered_kernel_to_id:
+        id = len(_registered_kernels)
+        _registered_kernels.append(wp_kernel)
+        _registered_kernel_to_id[wp_kernel] = id
+    else:
+        id = _registered_kernel_to_id[wp_kernel]
+
+    def bind(*args):
+        return _jax_warp_p.bind(*args, kernel=id)
+
+    return bind
+
+
+def _warp_custom_callback(stream, buffers, opaque, opaque_len):
+    # The descriptor is the form
+    # <kernel-id>|<launch-dims>|<arg-dims-list>
+    # Example:  42|16,32|16,32;100;16,32
+    kernel_id_str, dim_str, args_str = opaque.decode().split("|")
+
+    # Get the kernel from the registry.
+    kernel_id = int(kernel_id_str)
+    kernel = _registered_kernels[kernel_id]
+
+    # Parse launch dimensions.
+    dims = [int(d) for d in dim_str.split(",")]
+    bounds = launch_bounds_t(dims)
+
+    # Parse arguments.
+    arg_strings = args_str.split(";")
+    num_args = len(arg_strings)
+    assert num_args == len(kernel.adj.args), "Incorrect number of arguments"
+
+    # First param is the launch bounds.
+    kernel_params = (ctypes.c_void_p * (1 + num_args))()
+    kernel_params[0] = ctypes.addressof(bounds)
+
+    # Parse array descriptors.
+    args = []
+    for i in range(num_args):
+        dtype = kernel.adj.args[i].type.dtype
+        shape = [int(d) for d in arg_strings[i].split(",")]
+        strides = strides_from_shape(shape, dtype)
+
+        arr = array_t(buffers[i], 0, len(shape), shape, strides)
+        args.append(arr)  # keep a reference
+        arg_ptr = ctypes.addressof(arr)
+
+        kernel_params[i + 1] = arg_ptr
+
+    # Get current device.
+    device = wp.device_from_jax(_get_jax_device())
+
+    # Get kernel hooks.
+    # Note: module was loaded during jit lowering.
+    hooks = kernel.module.get_kernel_hooks(kernel, device)
+    assert hooks.forward, "Failed to find kernel entry point"
+
+    # Launch the kernel.
+    wp.context.runtime.core.cuda_launch_kernel(device.context, hooks.forward, bounds.size, 0, kernel_params, stream)
+
+
+# TODO: is there a simpler way of getting the Jax "current" device?
+def _get_jax_device():
+    # check if jax.default_device() context manager is active
+    device = jax.config.jax_default_device
+    # if default device is not set, use first device
+    if device is None:
+        device = jax.devices()[0]
+    return device
+
+
+def _create_jax_warp_primitive():
+    from functools import reduce
+
+    import jax
+    from jax._src.interpreters import batching
+    from jax.interpreters import mlir
+    from jax.interpreters.mlir import ir
+    from jaxlib.hlo_helpers import custom_call
+
+    global _jax_warp_p
+    global _cc_callback
+
+    # Create and register the primitive.
+    # TODO add default implementation that calls the kernel via warp.
+    _jax_warp_p = jax.core.Primitive("jax_warp")
+    _jax_warp_p.multiple_results = True
+
+    # TODO Just launch the kernel directly, but make sure the argument
+    # shapes are massaged the same way as below so that vmap works.
+    def impl(*args):
+        raise Exception("Not implemented")
+
+    _jax_warp_p.def_impl(impl)
+
+    # Auto-batching. Make sure all the arguments are fully broadcasted
+    # so that Warp is not confused about dimensions.
+    def vectorized_multi_batcher(args, dims, **params):
+        # Figure out the number of outputs.
+        wp_kernel = _registered_kernels[params["kernel"]]
+        output_count = len(wp_kernel.adj.args) - len(args)
+        shape, dim = next((a.shape, d) for a, d in zip(args, dims) if d is not None)
+        size = shape[dim]
+        args = [batching.bdim_at_front(a, d, size) if len(a.shape) else a for a, d in zip(args, dims)]
+        # Create the batched primitive.
+        return _jax_warp_p.bind(*args, **params), [dims[0]] * output_count
+
+    batching.primitive_batchers[_jax_warp_p] = vectorized_multi_batcher
+
+    def get_vecmat_shape(warp_type):
+        if hasattr(warp_type.dtype, "_shape_"):
+            return warp_type.dtype._shape_
+        return []
+
+    def strip_vecmat_dimensions(warp_arg, actual_shape):
+        shape = get_vecmat_shape(warp_arg.type)
+        for i, s in enumerate(reversed(shape)):
+            item = actual_shape[-i - 1]
+            if s != item:
+                raise Exception(f"The vector/matrix shape for argument {warp_arg.label} does not match")
+        return actual_shape[: len(actual_shape) - len(shape)]
+
+    def collapse_into_leading_dimension(warp_arg, actual_shape):
+        if len(actual_shape) < warp_arg.type.ndim:
+            raise Exception(f"Argument {warp_arg.label} has too few non-matrix/vector dimensions")
+        index_rest = len(actual_shape) - warp_arg.type.ndim + 1
+        leading_size = reduce(lambda x, y: x * y, actual_shape[:index_rest])
+        return [leading_size] + actual_shape[index_rest:]
+
+    # Infer array dimensions from input type.
+    def infer_dimensions(warp_arg, actual_shape):
+        actual_shape = strip_vecmat_dimensions(warp_arg, actual_shape)
+        return collapse_into_leading_dimension(warp_arg, actual_shape)
+
+    def base_type_to_jax(warp_dtype):
+        if hasattr(warp_dtype, "_wp_scalar_type_"):
+            return wp.dtype_to_jax(warp_dtype._wp_scalar_type_)
+        return wp.dtype_to_jax(warp_dtype)
+
+    def base_type_to_jax_ir(warp_dtype):
+        warp_to_jax_dict = {
+            wp.float16: ir.F16Type.get(),
+            wp.float32: ir.F32Type.get(),
+            wp.float64: ir.F64Type.get(),
+            wp.int8: ir.IntegerType.get_signless(8),
+            wp.int16: ir.IntegerType.get_signless(16),
+            wp.int32: ir.IntegerType.get_signless(32),
+            wp.int64: ir.IntegerType.get_signless(64),
+            wp.uint8: ir.IntegerType.get_unsigned(8),
+            wp.uint16: ir.IntegerType.get_unsigned(16),
+            wp.uint32: ir.IntegerType.get_unsigned(32),
+            wp.uint64: ir.IntegerType.get_unsigned(64),
+        }
+        if hasattr(warp_dtype, "_wp_scalar_type_"):
+            warp_dtype = warp_dtype._wp_scalar_type_
+        jax_dtype = warp_to_jax_dict.get(warp_dtype)
+        if jax_dtype is None:
+            raise TypeError(f"Invalid or unsupported data type: {warp_dtype}")
+        return jax_dtype
+
+    def base_type_is_compatible(warp_type, jax_ir_type):
+        jax_ir_to_warp = {
+            "f16": wp.float16,
+            "f32": wp.float32,
+            "f64": wp.float64,
+            "i8": wp.int8,
+            "i16": wp.int16,
+            "i32": wp.int32,
+            "i64": wp.int64,
+            "ui8": wp.uint8,
+            "ui16": wp.uint16,
+            "ui32": wp.uint32,
+            "ui64": wp.uint64,
+        }
+        expected_warp_type = jax_ir_to_warp.get(str(jax_ir_type))
+        if expected_warp_type is not None:
+            if hasattr(warp_type, "_wp_scalar_type_"):
+                return warp_type._wp_scalar_type_ == expected_warp_type
+            else:
+                return warp_type == expected_warp_type
+        else:
+            raise TypeError(f"Invalid or unsupported data type: {jax_ir_type}")
+
+    # Abstract evaluation.
+    def jax_warp_abstract(*args, kernel=None):
+        wp_kernel = _registered_kernels[kernel]
+        # All the extra arguments to the warp kernel are outputs.
+        warp_outputs = [o.type for o in wp_kernel.adj.args[len(args) :]]
+        # TODO. Let's just use the first input dimension to infer the output's dimensions.
+        dims = strip_vecmat_dimensions(wp_kernel.adj.args[0], list(args[0].shape))
+        jax_outputs = []
+        for o in warp_outputs:
+            shape = list(dims) + list(get_vecmat_shape(o))
+            dtype = base_type_to_jax(o.dtype)
+            jax_outputs.append(jax.core.ShapedArray(shape, dtype))
+        return jax_outputs
+
+    _jax_warp_p.def_abstract_eval(jax_warp_abstract)
+
+    # Lowering to MLIR.
+
+    # Create python-land custom call target.
+    CCALLFUNC = ctypes.CFUNCTYPE(
+        ctypes.c_voidp, ctypes.c_void_p, ctypes.POINTER(ctypes.c_void_p), ctypes.c_char_p, ctypes.c_size_t
+    )
+    _cc_callback = CCALLFUNC(_warp_custom_callback)
+    ccall_address = ctypes.cast(_cc_callback, ctypes.c_void_p)
+
+    # Put the custom call into a capsule, as required by XLA.
+    PyCapsule_Destructor = ctypes.CFUNCTYPE(None, ctypes.py_object)
+    PyCapsule_New = ctypes.pythonapi.PyCapsule_New
+    PyCapsule_New.restype = ctypes.py_object
+    PyCapsule_New.argtypes = (ctypes.c_void_p, ctypes.c_char_p, PyCapsule_Destructor)
+    capsule = PyCapsule_New(ccall_address.value, b"xla._CUSTOM_CALL_TARGET", PyCapsule_Destructor(0))
+
+    # Register the callback in XLA.
+    jax.lib.xla_client.register_custom_call_target("warp_call", capsule, platform="gpu")
+
+    def default_layout(shape):
+        return range(len(shape) - 1, -1, -1)
+
+    def warp_call_lowering(ctx, *args, kernel=None):
+        if not kernel:
+            raise Exception("Unknown kernel id " + str(kernel))
+        wp_kernel = _registered_kernels[kernel]
+
+        # TODO This may not be necessary, but it is perhaps better not to be
+        # mucking with kernel loading while already running the workload.
+        module = wp_kernel.module
+        device = wp.device_from_jax(_get_jax_device())
+        if not module.load(device):
+            raise Exception("Could not load kernel on device")
+
+        # Infer dimensions from the first input.
+        warp_arg0 = wp_kernel.adj.args[0]
+        actual_shape0 = ir.RankedTensorType(args[0].type).shape
+        dims = strip_vecmat_dimensions(warp_arg0, actual_shape0)
+        warp_dims = collapse_into_leading_dimension(warp_arg0, dims)
+
+        # Figure out the types and shapes of the input arrays.
+        arg_strings = []
+        operand_layouts = []
+        for actual, warg in zip(args, wp_kernel.adj.args):
+            wtype = warg.type
+            rtt = ir.RankedTensorType(actual.type)
+
+            if not isinstance(wtype, wp.array):
+                raise Exception("Only contiguous arrays are supported for Jax kernel arguments")
+
+            if not base_type_is_compatible(wtype.dtype, rtt.element_type):
+                raise TypeError(
+                    f"Incompatible data type for argument '{warg.label}', expected {type_str(wtype.dtype)}, got {rtt.element_type}"
+                )
+
+            # Infer array dimension (by removing the vector/matrix dimensions and
+            # collapsing the initial dimensions).
+            shape = infer_dimensions(warg, rtt.shape)
+
+            if len(shape) != wtype.ndim:
+                raise TypeError(f"Incompatible array dimensionality for argument '{warg.label}'")
+
+            arg_strings.append(",".join([str(d) for d in shape]))
+            operand_layouts.append(default_layout(rtt.shape))
+
+        # Figure out the types and shapes of the output arrays.
+        result_types = []
+        result_layouts = []
+        for warg in wp_kernel.adj.args[len(args) :]:
+            wtype = warg.type
+
+            if not isinstance(wtype, wp.array):
+                raise Exception("Only contiguous arrays are supported for Jax kernel arguments")
+
+            # Infer dimensions from the first input.
+            arg_strings.append(",".join([str(d) for d in warp_dims]))
+
+            result_shape = list(dims) + list(get_vecmat_shape(wtype))
+            result_types.append(ir.RankedTensorType.get(result_shape, base_type_to_jax_ir(wtype.dtype)))
+            result_layouts.append(default_layout(result_shape))
+
+        # Build opaque descriptor for callback.
+        shape_str = ",".join([str(d) for d in warp_dims])
+        args_str = ";".join(arg_strings)
+        descriptor = f"{kernel}|{shape_str}|{args_str}"
+
+        out = custom_call(
+            b"warp_call",
+            result_types=result_types,
+            operands=args,
+            backend_config=descriptor.encode("utf-8"),
+            operand_layouts=operand_layouts,
+            result_layouts=result_layouts,
+        ).results
+        return out
+
+    mlir.register_lowering(
+        _jax_warp_p,
+        warp_call_lowering,
+        platform="gpu",
+    )