warp-lang 1.8.1__py3-none-win_amd64.whl → 1.9.1__py3-none-win_amd64.whl

warp/__init__.pyi

@@ -20,6 +20,7 @@ from typing import Tuple
 from typing import Callable
 from typing import TypeVar
 from typing import Generic
+from typing import Sequence
 from typing import overload as over
 
 Length = TypeVar("Length", bound=int)
@@ -36,122 +37,1718 @@ FabricArray = Generic[DType]
 IndexedFabricArray = Generic[DType]
 Tile = Generic[DType, Shape]
 
-from warp.types import array, array1d, array2d, array3d, array4d, constant, from_ptr
-from warp.types import indexedarray, indexedarray1d, indexedarray2d, indexedarray3d, indexedarray4d
-from warp.fabric import fabricarray, fabricarrayarray, indexedfabricarray, indexedfabricarrayarray
-from warp.types import tile
+from warp.types import array as array
+from warp.types import array1d as array1d
+from warp.types import array2d as array2d
+from warp.types import array3d as array3d
+from warp.types import array4d as array4d
+from warp.types import constant as constant
+from warp.types import from_ptr as from_ptr
+from warp.types import fixedarray as fixedarray
+from warp.types import indexedarray as indexedarray
+from warp.types import indexedarray1d as indexedarray1d
+from warp.types import indexedarray2d as indexedarray2d
+from warp.types import indexedarray3d as indexedarray3d
+from warp.types import indexedarray4d as indexedarray4d
+from warp.fabric import fabricarray as fabricarray
+from warp.fabric import fabricarrayarray as fabricarrayarray
+from warp.fabric import indexedfabricarray as indexedfabricarray
+from warp.fabric import indexedfabricarrayarray as indexedfabricarrayarray
+from warp.types import tile as tile
+
+from warp.types import bool as bool
+from warp.types import int8 as int8
+from warp.types import uint8 as uint8
+from warp.types import int16 as int16
+from warp.types import uint16 as uint16
+from warp.types import int32 as int32
+from warp.types import uint32 as uint32
+from warp.types import int64 as int64
+from warp.types import uint64 as uint64
+from warp.types import float16 as float16
+from warp.types import float32 as float32
+from warp.types import float64 as float64
+
+from warp.types import vec2 as vec2
+from warp.types import vec2b as vec2b
+from warp.types import vec2ub as vec2ub
+from warp.types import vec2s as vec2s
+from warp.types import vec2us as vec2us
+from warp.types import vec2i as vec2i
+from warp.types import vec2ui as vec2ui
+from warp.types import vec2l as vec2l
+from warp.types import vec2ul as vec2ul
+from warp.types import vec2h as vec2h
+from warp.types import vec2f as vec2f
+from warp.types import vec2d as vec2d
+
+from warp.types import vec3 as vec3
+from warp.types import vec3b as vec3b
+from warp.types import vec3ub as vec3ub
+from warp.types import vec3s as vec3s
+from warp.types import vec3us as vec3us
+from warp.types import vec3i as vec3i
+from warp.types import vec3ui as vec3ui
+from warp.types import vec3l as vec3l
+from warp.types import vec3ul as vec3ul
+from warp.types import vec3h as vec3h
+from warp.types import vec3f as vec3f
+from warp.types import vec3d as vec3d
+
+from warp.types import vec4 as vec4
+from warp.types import vec4b as vec4b
+from warp.types import vec4ub as vec4ub
+from warp.types import vec4s as vec4s
+from warp.types import vec4us as vec4us
+from warp.types import vec4i as vec4i
+from warp.types import vec4ui as vec4ui
+from warp.types import vec4l as vec4l
+from warp.types import vec4ul as vec4ul
+from warp.types import vec4h as vec4h
+from warp.types import vec4f as vec4f
+from warp.types import vec4d as vec4d
+
+from warp.types import mat22 as mat22
+from warp.types import mat22h as mat22h
+from warp.types import mat22f as mat22f
+from warp.types import mat22d as mat22d
+
+from warp.types import mat33 as mat33
+from warp.types import mat33h as mat33h
+from warp.types import mat33f as mat33f
+from warp.types import mat33d as mat33d
+
+from warp.types import mat44 as mat44
+from warp.types import mat44h as mat44h
+from warp.types import mat44f as mat44f
+from warp.types import mat44d as mat44d
+
+from warp.types import quat as quat
+from warp.types import quath as quath
+from warp.types import quatf as quatf
+from warp.types import quatd as quatd
+
+from warp.types import transform as transform
+from warp.types import transformh as transformh
+from warp.types import transformf as transformf
+from warp.types import transformd as transformd
+
+from warp.types import spatial_vector as spatial_vector
+from warp.types import spatial_vectorh as spatial_vectorh
+from warp.types import spatial_vectorf as spatial_vectorf
+from warp.types import spatial_vectord as spatial_vectord
+
+from warp.types import spatial_matrix as spatial_matrix
+from warp.types import spatial_matrixh as spatial_matrixh
+from warp.types import spatial_matrixf as spatial_matrixf
+from warp.types import spatial_matrixd as spatial_matrixd
+
+from warp.types import Int as Int
+from warp.types import Float as Float
+from warp.types import Scalar as Scalar
+
+from warp.types import Bvh as Bvh
+from warp.types import Mesh as Mesh
+from warp.types import HashGrid as HashGrid
+from warp.types import Volume as Volume
+from warp.types import BvhQuery as BvhQuery
+from warp.types import HashGridQuery as HashGridQuery
+from warp.types import MeshQueryAABB as MeshQueryAABB
+from warp.types import MeshQueryPoint as MeshQueryPoint
+from warp.types import MeshQueryRay as MeshQueryRay
+
+from warp.types import matmul as matmul
+from warp.types import adj_matmul as adj_matmul
+from warp.types import batched_matmul as batched_matmul
+from warp.types import adj_batched_matmul as adj_batched_matmul
 
-from warp.types import bool, int8, uint8, int16, uint16, int32, uint32, int64, uint64, float16, float32, float64
-from warp.types import vec2, vec2b, vec2ub, vec2s, vec2us, vec2i, vec2ui, vec2l, vec2ul, vec2h, vec2f, vec2d
-from warp.types import vec3, vec3b, vec3ub, vec3s, vec3us, vec3i, vec3ui, vec3l, vec3ul, vec3h, vec3f, vec3d
-from warp.types import vec4, vec4b, vec4ub, vec4s, vec4us, vec4i, vec4ui, vec4l, vec4ul, vec4h, vec4f, vec4d
-from warp.types import mat22, mat22h, mat22f, mat22d
-from warp.types import mat33, mat33h, mat33f, mat33d
-from warp.types import mat44, mat44h, mat44f, mat44d
-from warp.types import quat, quath, quatf, quatd
-from warp.types import transform, transformh, transformf, transformd
-from warp.types import spatial_vector, spatial_vectorh, spatial_vectorf, spatial_vectord
-from warp.types import spatial_matrix, spatial_matrixh, spatial_matrixf, spatial_matrixd
+from warp.types import vector as vec
+from warp.types import matrix as mat
+
+from warp.types import matrix_from_cols as matrix_from_cols
+from warp.types import matrix_from_rows as matrix_from_rows
+
+from warp.types import dtype_from_numpy as dtype_from_numpy
+from warp.types import dtype_to_numpy as dtype_to_numpy
+
+from warp.types import from_ipc_handle as from_ipc_handle
+
+from warp.context import init as init
+from warp.context import func as func
+from warp.context import func_grad as func_grad
+from warp.context import func_replay as func_replay
+from warp.context import func_native as func_native
+from warp.context import kernel as kernel
+from warp.context import struct as struct
+from warp.context import overload as overload
+
+from warp.context import is_cpu_available as is_cpu_available
+from warp.context import is_cuda_available as is_cuda_available
+from warp.context import is_device_available as is_device_available
+from warp.context import get_devices as get_devices
+from warp.context import get_preferred_device as get_preferred_device
+from warp.context import get_cuda_devices as get_cuda_devices
+from warp.context import get_cuda_device_count as get_cuda_device_count
+from warp.context import get_cuda_device as get_cuda_device
+from warp.context import map_cuda_device as map_cuda_device
+from warp.context import unmap_cuda_device as unmap_cuda_device
+from warp.context import get_device as get_device
+from warp.context import set_device as set_device
+from warp.context import synchronize_device as synchronize_device
+
+from warp.context import zeros as zeros
+from warp.context import zeros_like as zeros_like
+from warp.context import ones as ones
+from warp.context import ones_like as ones_like
+from warp.context import full as full
+from warp.context import full_like as full_like
+from warp.context import clone as clone
+from warp.context import empty as empty
+from warp.context import empty_like as empty_like
+from warp.context import copy as copy
+from warp.context import from_numpy as from_numpy
+
+from warp.context import launch as launch
+from warp.context import launch_tiled as launch_tiled
+from warp.context import synchronize as synchronize
+from warp.context import compile_aot_module as compile_aot_module
+from warp.context import force_load as force_load
+from warp.context import load_module as load_module
+from warp.context import load_aot_module as load_aot_module
+from warp.context import event_from_ipc_handle as event_from_ipc_handle
+
+from warp.context import set_module_options as set_module_options
+from warp.context import get_module_options as get_module_options
+from warp.context import get_module as get_module
+
+from warp.context import capture_begin as capture_begin
+from warp.context import capture_end as capture_end
+from warp.context import capture_launch as capture_launch
+from warp.context import capture_if as capture_if
+from warp.context import capture_while as capture_while
+from warp.context import capture_debug_dot_print as capture_debug_dot_print
+
+from warp.context import Kernel as Kernel
+from warp.context import Function as Function
+from warp.context import Launch as Launch
+
+from warp.context import Stream as Stream
+from warp.context import get_stream as get_stream
+from warp.context import set_stream as set_stream
+from warp.context import wait_stream as wait_stream
+from warp.context import synchronize_stream as synchronize_stream
+
+from warp.context import Event as Event
+from warp.context import record_event as record_event
+from warp.context import wait_event as wait_event
+from warp.context import synchronize_event as synchronize_event
+from warp.context import get_event_elapsed_time as get_event_elapsed_time
+
+from warp.context import RegisteredGLBuffer as RegisteredGLBuffer
+
+from warp.context import is_mempool_supported as is_mempool_supported
+from warp.context import is_mempool_enabled as is_mempool_enabled
+from warp.context import set_mempool_enabled as set_mempool_enabled
+
+from warp.context import set_mempool_release_threshold as set_mempool_release_threshold
+from warp.context import get_mempool_release_threshold as get_mempool_release_threshold
+from warp.context import get_mempool_used_mem_current as get_mempool_used_mem_current
+from warp.context import get_mempool_used_mem_high as get_mempool_used_mem_high
+
+from warp.context import is_mempool_access_supported as is_mempool_access_supported
+from warp.context import is_mempool_access_enabled as is_mempool_access_enabled
+from warp.context import set_mempool_access_enabled as set_mempool_access_enabled
+
+from warp.context import is_peer_access_supported as is_peer_access_supported
+from warp.context import is_peer_access_enabled as is_peer_access_enabled
+from warp.context import set_peer_access_enabled as set_peer_access_enabled
+
+from warp.tape import Tape as Tape
+
+from warp.utils import ScopedTimer as ScopedTimer
+from warp.utils import ScopedDevice as ScopedDevice
+from warp.utils import ScopedStream as ScopedStream
+from warp.utils import ScopedMempool as ScopedMempool
+from warp.utils import ScopedMempoolAccess as ScopedMempoolAccess
+from warp.utils import ScopedPeerAccess as ScopedPeerAccess
+from warp.utils import ScopedCapture as ScopedCapture
+
+from warp.utils import transform_expand as transform_expand
+from warp.utils import quat_between_vectors as quat_between_vectors
+
+from warp.utils import TimingResult as TimingResult
+from warp.utils import timing_begin as timing_begin
+from warp.utils import timing_end as timing_end
+from warp.utils import timing_print as timing_print
+
+from warp.utils import TIMING_KERNEL as TIMING_KERNEL
+from warp.utils import TIMING_KERNEL_BUILTIN as TIMING_KERNEL_BUILTIN
+from warp.utils import TIMING_MEMCPY as TIMING_MEMCPY
+from warp.utils import TIMING_MEMSET as TIMING_MEMSET
+from warp.utils import TIMING_GRAPH as TIMING_GRAPH
+from warp.utils import TIMING_ALL as TIMING_ALL
+
+from warp.utils import map as map
+
+from warp.marching_cubes import MarchingCubes as MarchingCubes
+
+from warp.torch import from_torch as from_torch
+from warp.torch import to_torch as to_torch
+from warp.torch import dtype_from_torch as dtype_from_torch
+from warp.torch import dtype_to_torch as dtype_to_torch
+from warp.torch import device_from_torch as device_from_torch
+from warp.torch import device_to_torch as device_to_torch
+from warp.torch import stream_from_torch as stream_from_torch
+from warp.torch import stream_to_torch as stream_to_torch
+
+from warp.jax import from_jax as from_jax
+from warp.jax import to_jax as to_jax
+from warp.jax import dtype_from_jax as dtype_from_jax
+from warp.jax import dtype_to_jax as dtype_to_jax
+from warp.jax import device_from_jax as device_from_jax
+from warp.jax import device_to_jax as device_to_jax
+
+from warp.dlpack import from_dlpack as from_dlpack
+from warp.dlpack import to_dlpack as to_dlpack
+
+from warp.paddle import from_paddle as from_paddle
+from warp.paddle import to_paddle as to_paddle
+from warp.paddle import dtype_from_paddle as dtype_from_paddle
+from warp.paddle import dtype_to_paddle as dtype_to_paddle
+from warp.paddle import device_from_paddle as device_from_paddle
+from warp.paddle import device_to_paddle as device_to_paddle
+from warp.paddle import stream_from_paddle as stream_from_paddle
+
+from warp.build import clear_kernel_cache as clear_kernel_cache
+from warp.build import clear_lto_cache as clear_lto_cache
+
+from warp.constants import *
+
+from . import builtins
+from warp.builtins import static as static
+
+from warp.math import *
+
+from . import config as config
+
+__version__ = config.version
+
+class vec2h:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2h) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float16, y: float16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2f:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2f) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float32, y: float32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2d:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2d) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float64, y: float64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2b:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2b) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int8, y: int8) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int8]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int8) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2ub:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2ub) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint8, y: uint8) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint8]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint8) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2s:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2s) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int16, y: int16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2us:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2us) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint16, y: uint16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2i:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2i) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int32, y: int32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2ui:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2ui) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint32, y: uint32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2l:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2l) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int64, y: int64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec2ul:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec2ul) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint64, y: uint64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+vec2 = vec2f
+
+class vec3h:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3h) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float16, y: float16, z: float16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3f:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3f) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float32, y: float32, z: float32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3d:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3d) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float64, y: float64, z: float64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3b:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3b) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int8, y: int8, z: int8) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int8]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int8) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3ub:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3ub) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint8, y: uint8, z: uint8) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint8]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint8) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3s:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3s) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int16, y: int16, z: int16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3us:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3us) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint16, y: uint16, z: uint16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3i:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3i) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int32, y: int32, z: int32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3ui:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3ui) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint32, y: uint32, z: uint32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3l:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3l) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int64, y: int64, z: int64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec3ul:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec3ul) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint64, y: uint64, z: uint64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+vec3 = vec3f
+
+class vec4h:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4h) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float16, y: float16, z: float16, w: float16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4f:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4f) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float32, y: float32, z: float32, w: float32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4d:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4d) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float64, y: float64, z: float64, w: float64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4b:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4b) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int8, y: int8, z: int8, w: int8) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int8]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int8) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4ub:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4ub) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint8, y: uint8, z: uint8, w: uint8) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint8]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint8) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4s:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4s) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int16, y: int16, z: int16, w: int16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4us:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4us) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint16, y: uint16, z: uint16, w: uint16) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint16]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint16) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4i:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4i) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int32, y: int32, z: int32, w: int32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4ui:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4ui) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint32, y: uint32, z: uint32, w: uint32) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint32]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint32) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4l:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4l) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: int64, y: int64, z: int64, w: int64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[int64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: int64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+class vec4ul:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized vector."""
+        ...
+
+    @over
+    def __init__(self, other: vec4ul) -> None:
+        """Construct a vector by copy."""
+        ...
+
+    @over
+    def __init__(self, x: uint64, y: uint64, z: uint64, w: uint64) -> None:
+        """Construct a vector from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[uint64]) -> None:
+        """Construct a vector from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: uint64) -> None:
+        """Construct a vector filled with a value."""
+        ...
+
+vec4 = vec4f
+
+class mat22h:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
+
+    @over
+    def __init__(self, other: mat22h) -> None:
+        """Construct a matrix by copy."""
+        ...
+
+    @over
+    def __init__(self, m00: float16, m01: float16, m10: float16, m11: float16) -> None:
+        """Construct a matrix from its component values."""
+        ...
+
+    @over
+    def __init__(self, v0: vec2h, v1: vec2h) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
 
-from warp.types import Int, Float, Scalar
+    @over
+    def __init__(self, args: Sequence[float16]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
 
-from warp.types import Bvh, Mesh, HashGrid, Volume, MarchingCubes
-from warp.types import BvhQuery, HashGridQuery, MeshQueryAABB, MeshQueryPoint, MeshQueryRay
+    @over
+    def __init__(self, value: float16) -> None:
+        """Construct a matrix filled with a value."""
+        ...
 
-from warp.types import matmul, adj_matmul, batched_matmul, adj_batched_matmul
+class mat22f:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
 
-from warp.types import vector as vec
-from warp.types import matrix as mat
+    @over
+    def __init__(self, other: mat22f) -> None:
+        """Construct a matrix by copy."""
+        ...
 
-from warp.types import dtype_from_numpy, dtype_to_numpy
-
-from warp.types import from_ipc_handle
-
-from warp.context import init, func, func_grad, func_replay, func_native, kernel, struct, overload
-from warp.context import is_cpu_available, is_cuda_available, is_device_available
-from warp.context import get_devices, get_preferred_device
-from warp.context import get_cuda_devices, get_cuda_device_count, get_cuda_device, map_cuda_device, unmap_cuda_device
-from warp.context import get_device, set_device, synchronize_device
-from warp.context import (
-    zeros,
-    zeros_like,
-    ones,
-    ones_like,
-    full,
-    full_like,
-    clone,
-    empty,
-    empty_like,
-    copy,
-    from_numpy,
-    launch,
-    launch_tiled,
-    synchronize,
-    force_load,
-    load_module,
-    event_from_ipc_handle,
-)
-from warp.context import set_module_options, get_module_options, get_module
-from warp.context import capture_begin, capture_end, capture_launch, capture_if, capture_while, capture_debug_dot_print
-from warp.context import Kernel, Function, Launch
-from warp.context import Stream, get_stream, set_stream, wait_stream, synchronize_stream
-from warp.context import Event, record_event, wait_event, synchronize_event, get_event_elapsed_time
-from warp.context import RegisteredGLBuffer
-from warp.context import is_mempool_supported, is_mempool_enabled, set_mempool_enabled
-from warp.context import (
-    set_mempool_release_threshold,
-    get_mempool_release_threshold,
-    get_mempool_used_mem_current,
-    get_mempool_used_mem_high,
-)
-from warp.context import is_mempool_access_supported, is_mempool_access_enabled, set_mempool_access_enabled
-from warp.context import is_peer_access_supported, is_peer_access_enabled, set_peer_access_enabled
-
-from warp.tape import Tape
-from warp.utils import ScopedTimer, ScopedDevice, ScopedStream
-from warp.utils import ScopedMempool, ScopedMempoolAccess, ScopedPeerAccess
-from warp.utils import ScopedCapture
-from warp.utils import transform_expand, quat_between_vectors
-from warp.utils import TimingResult, timing_begin, timing_end, timing_print
-from warp.utils import (
-    TIMING_KERNEL,
-    TIMING_KERNEL_BUILTIN,
-    TIMING_MEMCPY,
-    TIMING_MEMSET,
-    TIMING_GRAPH,
-    TIMING_ALL,
-)
-from warp.utils import map
-
-from warp.torch import from_torch, to_torch
-from warp.torch import dtype_from_torch, dtype_to_torch
-from warp.torch import device_from_torch, device_to_torch
-from warp.torch import stream_from_torch, stream_to_torch
-
-from warp.jax import from_jax, to_jax
-from warp.jax import dtype_from_jax, dtype_to_jax
-from warp.jax import device_from_jax, device_to_jax
-
-from warp.dlpack import from_dlpack, to_dlpack
-
-from warp.paddle import from_paddle, to_paddle
-from warp.paddle import dtype_from_paddle, dtype_to_paddle
-from warp.paddle import device_from_paddle, device_to_paddle
-from warp.paddle import stream_from_paddle
-
-from warp.build import clear_kernel_cache
-from warp.build import clear_lto_cache
+    @over
+    def __init__(self, m00: float32, m01: float32, m10: float32, m11: float32) -> None:
+        """Construct a matrix from its component values."""
+        ...
 
-from warp.constants import *
+    @over
+    def __init__(self, v0: vec2f, v1: vec2f) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
 
-from . import builtins
-from warp.builtins import static
+    @over
+    def __init__(self, args: Sequence[float32]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
 
-from warp.math import *
+    @over
+    def __init__(self, value: float32) -> None:
+        """Construct a matrix filled with a value."""
+        ...
 
-import warp.config as config
+class mat22d:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
 
-__version__ = config.version
+    @over
+    def __init__(self, other: mat22d) -> None:
+        """Construct a matrix by copy."""
+        ...
+
+    @over
+    def __init__(self, m00: float64, m01: float64, m10: float64, m11: float64) -> None:
+        """Construct a matrix from its component values."""
+        ...
+
+    @over
+    def __init__(self, v0: vec2d, v1: vec2d) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float64]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float64) -> None:
+        """Construct a matrix filled with a value."""
+        ...
+
+mat22 = mat22f
+
+class mat33h:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
+
+    @over
+    def __init__(self, other: mat33h) -> None:
+        """Construct a matrix by copy."""
+        ...
+
+    @over
+    def __init__(
+        self,
+        m00: float16,
+        m01: float16,
+        m02: float16,
+        m10: float16,
+        m11: float16,
+        m12: float16,
+        m20: float16,
+        m21: float16,
+        m22: float16,
+    ) -> None:
+        """Construct a matrix from its component values."""
+        ...
+
+    @over
+    def __init__(self, v0: vec3h, v1: vec3h, v2: vec3h) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float16]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float16) -> None:
+        """Construct a matrix filled with a value."""
+        ...
+
+class mat33f:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
+
+    @over
+    def __init__(self, other: mat33f) -> None:
+        """Construct a matrix by copy."""
+        ...
+
+    @over
+    def __init__(
+        self,
+        m00: float32,
+        m01: float32,
+        m02: float32,
+        m10: float32,
+        m11: float32,
+        m12: float32,
+        m20: float32,
+        m21: float32,
+        m22: float32,
+    ) -> None:
+        """Construct a matrix from its component values."""
+        ...
+
+    @over
+    def __init__(self, v0: vec3f, v1: vec3f, v2: vec3f) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float32]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float32) -> None:
+        """Construct a matrix filled with a value."""
+        ...
+
+class mat33d:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
+
+    @over
+    def __init__(self, other: mat33d) -> None:
+        """Construct a matrix by copy."""
+        ...
+
+    @over
+    def __init__(
+        self,
+        m00: float64,
+        m01: float64,
+        m02: float64,
+        m10: float64,
+        m11: float64,
+        m12: float64,
+        m20: float64,
+        m21: float64,
+        m22: float64,
+    ) -> None:
+        """Construct a matrix from its component values."""
+        ...
+
+    @over
+    def __init__(self, v0: vec3d, v1: vec3d, v2: vec3d) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float64]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float64) -> None:
+        """Construct a matrix filled with a value."""
+        ...
+
+mat33 = mat33f
+
+class mat44h:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
+
+    @over
+    def __init__(self, other: mat44h) -> None:
+        """Construct a matrix by copy."""
+        ...
+
+    @over
+    def __init__(
+        self,
+        m00: float16,
+        m01: float16,
+        m02: float16,
+        m03: float16,
+        m10: float16,
+        m11: float16,
+        m12: float16,
+        m13: float16,
+        m20: float16,
+        m21: float16,
+        m22: float16,
+        m23: float16,
+        m30: float16,
+        m31: float16,
+        m32: float16,
+        m33: float16,
+    ) -> None:
+        """Construct a matrix from its component values."""
+        ...
+
+    @over
+    def __init__(self, v0: vec4h, v1: vec4h, v2: vec4h, v3: vec4h) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float16]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float16) -> None:
+        """Construct a matrix filled with a value."""
+        ...
+
+class mat44f:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
+
+    @over
+    def __init__(self, other: mat44f) -> None:
+        """Construct a matrix by copy."""
+        ...
+
+    @over
+    def __init__(
+        self,
+        m00: float32,
+        m01: float32,
+        m02: float32,
+        m03: float32,
+        m10: float32,
+        m11: float32,
+        m12: float32,
+        m13: float32,
+        m20: float32,
+        m21: float32,
+        m22: float32,
+        m23: float32,
+        m30: float32,
+        m31: float32,
+        m32: float32,
+        m33: float32,
+    ) -> None:
+        """Construct a matrix from its component values."""
+        ...
+
+    @over
+    def __init__(self, v0: vec4f, v1: vec4f, v2: vec4f, v3: vec4f) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float32]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float32) -> None:
+        """Construct a matrix filled with a value."""
+        ...
+
+class mat44d:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized matrix."""
+        ...
+
+    @over
+    def __init__(self, other: mat44d) -> None:
+        """Construct a matrix by copy."""
+        ...
+
+    @over
+    def __init__(
+        self,
+        m00: float64,
+        m01: float64,
+        m02: float64,
+        m03: float64,
+        m10: float64,
+        m11: float64,
+        m12: float64,
+        m13: float64,
+        m20: float64,
+        m21: float64,
+        m22: float64,
+        m23: float64,
+        m30: float64,
+        m31: float64,
+        m32: float64,
+        m33: float64,
+    ) -> None:
+        """Construct a matrix from its component values."""
+        ...
+
+    @over
+    def __init__(self, v0: vec4d, v1: vec4d, v2: vec4d, v3: vec4d) -> None:
+        """Construct a matrix from its row vectors."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float64]) -> None:
+        """Construct a matrix from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float64) -> None:
+        """Construct a matrix filled with a value."""
+        ...
+
+mat44 = mat44f
+
+class quath:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized quaternion."""
+        ...
+
+    @over
+    def __init__(self, other: quath) -> None:
+        """Construct a quaternion by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float16, y: float16, z: float16, w: float16) -> None:
+        """Construct a quaternion from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float16]) -> None:
+        """Construct a quaternion from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float16) -> None:
+        """Construct a quaternion filled with a value."""
+        ...
+
+class quatf:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized quaternion."""
+        ...
+
+    @over
+    def __init__(self, other: quatf) -> None:
+        """Construct a quaternion by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float32, y: float32, z: float32, w: float32) -> None:
+        """Construct a quaternion from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float32]) -> None:
+        """Construct a quaternion from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float32) -> None:
+        """Construct a quaternion filled with a value."""
+        ...
+
+class quatd:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized quaternion."""
+        ...
+
+    @over
+    def __init__(self, other: quatd) -> None:
+        """Construct a quaternion by copy."""
+        ...
+
+    @over
+    def __init__(self, x: float64, y: float64, z: float64, w: float64) -> None:
+        """Construct a quaternion from its component values."""
+        ...
+
+    @over
+    def __init__(self, args: Sequence[float64]) -> None:
+        """Construct a quaternion from a sequence of values."""
+        ...
+
+    @over
+    def __init__(self, value: float64) -> None:
+        """Construct a quaternion filled with a value."""
+        ...
+
+quat = quatf
+
+class transformh:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized transformation."""
+        ...
+
+    @over
+    def __init__(self, other: transformh) -> None:
+        """Construct a transformation by copy."""
+        ...
+
+    @over
+    def __init__(self, p: vec3h, q: quath) -> None:
+        """Construct a transformation from its p and q components."""
+        ...
+
+    @over
+    def __init__(
+        self, px: float16, py: float16, pz: float16, qx: float16, qy: float16, qz: float16, qw: float16
+    ) -> None:
+        """Construct a transformation from its component values."""
+        ...
+
+    @over
+    def __init__(self, p: Sequence[float16], q: Sequence[float16]) -> None:
+        """Construct a transformation from two sequences of values."""
+        ...
+
+    @over
+    def __init__(self, value: float16) -> None:
+        """Construct a transformation filled with a value."""
+        ...
+
+class transformf:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized transformation."""
+        ...
+
+    @over
+    def __init__(self, other: transformf) -> None:
+        """Construct a transformation by copy."""
+        ...
+
+    @over
+    def __init__(self, p: vec3f, q: quatf) -> None:
+        """Construct a transformation from its p and q components."""
+        ...
+
+    @over
+    def __init__(
+        self, px: float32, py: float32, pz: float32, qx: float32, qy: float32, qz: float32, qw: float32
+    ) -> None:
+        """Construct a transformation from its component values."""
+        ...
+
+    @over
+    def __init__(self, p: Sequence[float32], q: Sequence[float32]) -> None:
+        """Construct a transformation from two sequences of values."""
+        ...
+
+    @over
+    def __init__(self, value: float32) -> None:
+        """Construct a transformation filled with a value."""
+        ...
+
+class transformd:
+    @over
+    def __init__(self) -> None:
+        """Construct a zero-initialized transformation."""
+        ...
+
+    @over
+    def __init__(self, other: transformd) -> None:
+        """Construct a transformation by copy."""
+        ...
+
+    @over
+    def __init__(self, p: vec3d, q: quatd) -> None:
+        """Construct a transformation from its p and q components."""
+        ...
+
+    @over
+    def __init__(
+        self, px: float64, py: float64, pz: float64, qx: float64, qy: float64, qz: float64, qw: float64
+    ) -> None:
+        """Construct a transformation from its component values."""
+        ...
+
+    @over
+    def __init__(self, p: Sequence[float64], q: Sequence[float64]) -> None:
+        """Construct a transformation from two sequences of values."""
+        ...
+
+    @over
+    def __init__(self, value: float64) -> None:
+        """Construct a transformation filled with a value."""
+        ...
+
+transform = transformf
 
 @over
 def min(a: Scalar, b: Scalar) -> Scalar:
@@ -924,7 +2521,7 @@ def tile_arange(*args: Scalar, dtype: Scalar, storage: str) -> Tile[Scalar, Tupl
 
 @over
 def tile_load(
-    a: Array[Any], shape: Tuple[int, ...], offset: Tuple[int, ...], storage: str
+    a: Array[Any], shape: Tuple[int, ...], offset: Tuple[int, ...], storage: str, bounds_check: bool
 ) -> Tile[Any, Tuple[int, ...]]:
     """Loads a tile from a global memory array.
 
@@ -935,12 +2532,80 @@ def tile_load(
     :param offset: Offset in the source array to begin reading from (optional)
     :param storage: The storage location for the tile: ``"register"`` for registers
       (default) or ``"shared"`` for shared memory.
+    :param bounds_check: Needed for unaligned tiles, but can disable for memory-aligned tiles for faster load times
+    :returns: A tile with shape as specified and data type the same as the source array
+    """
+    ...
+
+@over
+def tile_load_indexed(
+    a: Array[Any],
+    indices: Tile[int32, Tuple[int]],
+    shape: Tuple[int, ...],
+    offset: Tuple[int, ...],
+    axis: int32,
+    storage: str,
+) -> Tile[Any, Tuple[int, ...]]:
+    """Loads a tile from a global memory array, with loads along a specified axis mapped according to a 1D tile of indices.
+
+    :param a: The source array in global memory
+    :param indices: A 1D tile of integer indices mapping to elements in ``a``.
+    :param shape: Shape of the tile to load, must have the same number of dimensions as ``a``, and along ``axis``, it must have the same number of elements as the ``indices`` tile.
+    :param offset: Offset in the source array to begin reading from (optional)
+    :param axis: Axis of ``a`` that indices refer to
+    :param storage: The storage location for the tile: ``"register"`` for registers (default) or ``"shared"`` for shared memory.
     :returns: A tile with shape as specified and data type the same as the source array
+
+    This example shows how to select and store the even indexed rows from a 2D array:
+
+    .. code-block:: python
+
+        TILE_M = wp.constant(2)
+        TILE_N = wp.constant(2)
+        HALF_M = wp.constant(TILE_M // 2)
+        HALF_N = wp.constant(TILE_N // 2)
+
+        @wp.kernel
+        def compute(x: wp.array2d(dtype=float), y: wp.array2d(dtype=float)):
+            i, j = wp.tid()
+
+            evens = wp.tile_arange(HALF_M, dtype=int, storage="shared") * 2
+
+            t0 = wp.tile_load_indexed(
+                x, indices=evens, shape=(HALF_M, TILE_N), offset=(i * TILE_M, j * TILE_N), axis=0, storage="register"
+            )
+            wp.tile_store(y, t0, offset=(i * HALF_M, j * TILE_N))
+
+        M = TILE_M * 2
+        N = TILE_N * 2
+
+        arr = np.arange(M * N).reshape(M, N)
+
+        x = wp.array(arr, dtype=float)
+        y = wp.zeros((M // 2, N), dtype=float)
+
+        wp.launch_tiled(compute, dim=[2, 2], inputs=[x], outputs=[y], block_dim=32, device=device)
+
+        print(x.numpy())
+        print(y.numpy())
+
+    Prints:
+
+    .. code-block:: text
+
+        [[ 0.  1.  2.  3.]
+         [ 4.  5.  6.  7.]
+         [ 8.  9. 10. 11.]
+         [12. 13. 14. 15.]]
+
+        [[ 0.  1.  2.  3.]
+         [ 8.  9. 10. 11.]]
+
     """
     ...
 
 @over
-def tile_store(a: Array[Any], t: Tile[Any, Tuple[int, ...]], offset: Tuple[int, ...]):
+def tile_store(a: Array[Any], t: Tile[Any, Tuple[int, ...]], offset: Tuple[int, ...], bounds_check: bool):
     """Store a tile to a global memory array.
 
     This method will cooperatively store a tile to global memory using all threads in the block.
@@ -948,22 +2613,147 @@ def tile_store(a: Array[Any], t: Tile[Any, Tuple[int, ...]], offset: Tuple[int,
     :param a: The destination array in global memory
     :param t: The source tile to store data from, must have the same data type and number of dimensions as the destination array
     :param offset: Offset in the destination array (optional)
+    :param bounds_check: Needed for unaligned tiles, but can disable for memory-aligned tiles for faster write times
+
+    """
+    ...
+
+@over
+def tile_store_indexed(
+    a: Array[Any], indices: Tile[int32, Tuple[int]], t: Tile[Any, Tuple[int, ...]], offset: Tuple[int, ...], axis: int32
+):
+    """Store a tile to a global memory array, with storage along a specified axis mapped according to a 1D tile of indices.
+
+    :param a: The destination array in global memory
+    :param indices: A 1D tile of integer indices mapping to elements in ``a``.
+    :param t: The source tile to store data from, must have the same data type and number of dimensions as the destination array, and along ``axis``, it must have the same number of elements as the ``indices`` tile.
+    :param offset: Offset in the destination array (optional)
+    :param axis: Axis of ``a`` that indices refer to
+
+    This example shows how to map tile rows to the even rows of a 2D array:
+
+    .. code-block:: python
+
+        TILE_M = wp.constant(2)
+        TILE_N = wp.constant(2)
+        TWO_M = wp.constant(TILE_M * 2)
+        TWO_N = wp.constant(TILE_N * 2)
+
+        @wp.kernel
+        def compute(x: wp.array2d(dtype=float), y: wp.array2d(dtype=float)):
+            i, j = wp.tid()
+
+            t = wp.tile_load(x, shape=(TILE_M, TILE_N), offset=(i * TILE_M, j * TILE_N), storage="register")
+
+            evens_M = wp.tile_arange(TILE_M, dtype=int, storage="shared") * 2
+
+            wp.tile_store_indexed(y, indices=evens_M, t=t, offset=(i * TWO_M, j * TILE_N), axis=0)
+
+        M = TILE_M * 2
+        N = TILE_N * 2
+
+        arr = np.arange(M * N, dtype=float).reshape(M, N)
+
+        x = wp.array(arr, dtype=float, requires_grad=True, device=device)
+        y = wp.zeros((M * 2, N), dtype=float, requires_grad=True, device=device)
+
+        wp.launch_tiled(compute, dim=[2, 2], inputs=[x], outputs=[y], block_dim=32, device=device)
+
+        print(x.numpy())
+        print(y.numpy())
+
+    Prints:
+
+    .. code-block:: text
+
+        [[ 0.  1.  2.  3.]
+         [ 4.  5.  6.  7.]
+         [ 8.  9. 10. 11.]
+         [12. 13. 14. 15.]]
+
+        [[ 0.  1.  2.  3.]
+         [ 0.  0.  0.  0.]
+         [ 4.  5.  6.  7.]
+         [ 0.  0.  0.  0.]
+         [ 8.  9. 10. 11.]
+         [ 0.  0.  0.  0.]
+         [12. 13. 14. 15.]
+         [ 0.  0.  0.  0.]]
+
     """
     ...
 
 @over
 def tile_atomic_add(
-    a: Array[Any], t: Tile[Any, Tuple[int, ...]], offset: Tuple[int, ...]
+    a: Array[Any], t: Tile[Any, Tuple[int, ...]], offset: Tuple[int, ...], bounds_check: bool
 ) -> Tile[Any, Tuple[int, ...]]:
     """Atomically add a tile onto the array `a`, each element will be updated atomically.
 
     :param a: Array in global memory, should have the same ``dtype`` as the input tile
     :param t: Source tile to add to the destination array
     :param offset: Offset in the destination array (optional)
+    :param bounds_check: Needed for unaligned tiles, but can disable for memory-aligned tiles for faster write times
     :returns: A tile with the same dimensions and data type as the source tile, holding the original value of the destination elements
     """
     ...
 
+@over
+def tile_atomic_add_indexed(
+    a: Array[Any], indices: Tile[int32, Tuple[int]], t: Tile[Any, Tuple[int, ...]], offset: Tuple[int, ...], axis: int32
+) -> Tile[Any, Tuple[int, ...]]:
+    """Atomically add a tile to a global memory array, with storage along a specified axis mapped according to a 1D tile of indices.
+
+    :param a: The destination array in global memory
+    :param indices: A 1D tile of integer indices mapping to elements in ``a``.
+    :param t: The source tile to extract data from, must have the same data type and number of dimensions as the destination array, and along ``axis``, it must have the same number of elements as the ``indices`` tile.
+    :param offset: Offset in the destination array (optional)
+    :param axis: Axis of ``a`` that indices refer to
+
+    This example shows how to compute a blocked, row-wise reduction:
+
+    .. code-block:: python
+
+        TILE_M = wp.constant(2)
+        TILE_N = wp.constant(2)
+
+        @wp.kernel
+        def tile_atomic_add_indexed(x: wp.array2d(dtype=float), y: wp.array2d(dtype=float)):
+            i, j = wp.tid()
+
+            t = wp.tile_load(x, shape=(TILE_M, TILE_N), offset=(i * TILE_M, j * TILE_N), storage="register")
+
+            zeros = wp.tile_zeros(TILE_M, dtype=int, storage="shared")
+
+            wp.tile_atomic_add_indexed(y, indices=zeros, t=t, offset=(i, j * TILE_N), axis=0)
+
+        M = TILE_M * 2
+        N = TILE_N * 2
+
+        arr = np.arange(M * N, dtype=float).reshape(M, N)
+
+        x = wp.array(arr, dtype=float, requires_grad=True, device=device)
+        y = wp.zeros((2, N), dtype=float, requires_grad=True, device=device)
+
+        wp.launch_tiled(tile_atomic_add_indexed, dim=[2, 2], inputs=[x], outputs=[y], block_dim=32, device=device)
+
+        print(x.numpy())
+        print(y.numpy())
+
+    Prints:
+
+    .. code-block:: text
+
+        [[ 0.  1.  2.  3.]
+         [ 4.  5.  6.  7.]
+         [ 8.  9. 10. 11.]
+         [12. 13. 14. 15.]]
+
+        [[ 4.  6.  8. 10.]
+         [20. 22. 24. 26.]]
+
+    """
+    ...
+
 @over
 def tile_view(
     t: Tile[Any, Tuple[int, ...]], offset: Tuple[int, ...], shape: Tuple[int, ...]
@@ -1370,7 +3160,7 @@ def tile_map(op: Callable, a: Tile[Scalar, Tuple[int, ...]]) -> Tile[Scalar, Tup
 
     :param op: A callable function that accepts one argument and returns one argument, may be a user function or builtin
     :param a: The input tile, the operator (or one of its overloads) must be able to accept the tile's data type
-    :returns: A tile with the same dimensions and data type as the input tile.
+    :returns: A tile with the same dimensions as the input tile. Its datatype is specified by the return type of op
 
     Example:
 
@@ -1401,12 +3191,12 @@ def tile_map(
     """Apply a binary function onto the tile.
 
     This function cooperatively applies a binary function to each element of the tiles using all threads in the block.
-    Both input tiles must have the same dimensions and datatype.
+    Both input tiles must have the same dimensions, and if using a builtin op, the same datatypes.
 
     :param op: A callable function that accepts two arguments and returns one argument, all of the same type, may be a user function or builtin
     :param a: The first input tile, the operator (or one of its overloads) must be able to accept the tile's dtype
     :param b: The second input tile, the operator (or one of its overloads) must be able to accept the tile's dtype
-    :returns: A tile with the same dimensions and datatype as the input tiles.
+    :returns: A tile with the same dimensions as the input tiles. Its datatype is specified by the return type of op
 
     Example:
 
@@ -1551,7 +3341,7 @@ def mesh_query_ray(id: uint64, start: vec3f, dir: vec3f, max_t: float32) -> Mesh
 def mesh_query_aabb(id: uint64, low: vec3f, high: vec3f) -> MeshQueryAABB:
     """Construct an axis-aligned bounding box query against a :class:`Mesh`.
 
-    This query can be used to iterate over all triangles inside a volume.
+    This query can be used to iterate over all bounding boxes of the triangles inside a volume.
 
     :param id: The mesh identifier
     :param low: The lower bound of the bounding box in mesh space
@@ -1561,7 +3351,7 @@ def mesh_query_aabb(id: uint64, low: vec3f, high: vec3f) -> MeshQueryAABB:
 
 @over
 def mesh_query_aabb_next(query: MeshQueryAABB, index: int32) -> bool:
-    """Move to the next triangle overlapping the query bounding box.
+    """Move to the next triangle whose bounding box overlaps the query bounding box.
 
     The index of the current face is stored in ``index``, returns ``False`` if there are no more overlapping triangles.
     """
@@ -2971,7 +4761,7 @@ def mod(a: Scalar, b: Scalar) -> Scalar:
     ...
 
 @over
-def mod(a: Vector[Any, Scalar], b: Vector[Any, Scalar]) -> Scalar:
+def mod(a: Vector[Any, Scalar], b: Vector[Any, Scalar]) -> Vector[Any, Scalar]:
     """Modulo operation using truncated division."""
     ...