ai-content-maker/.venv/Lib/site-packages/torch/_inductor/lowering.py

6007 lines
190 KiB
Python

import functools
import itertools
import logging
import os
import warnings
from collections import defaultdict
from typing import Any, Callable, Dict, List, Optional, Set, Tuple, Union
import sympy
import torch
import torch.ao.quantization.fx._decomposed
import torch.fx
import torch.utils._pytree as pytree
from torch._higher_order_ops.triton_kernel_wrap import (
triton_kernel_wrapper_functional,
triton_kernel_wrapper_mutation,
)
from torch._prims_common import (
canonicalize_dim,
canonicalize_dims,
check,
dtype_to_type,
elementwise_dtypes,
ELEMENTWISE_TYPE_PROMOTION_KIND,
get_computation_dtype,
is_boolean_dtype,
is_float_dtype,
is_integer_dtype,
Number,
)
from torch.fx.experimental.sym_node import magic_methods, method_to_operator
from torch.utils._sympy.functions import CeilDiv, FloorDiv, ModularIndexing
from .._dynamo.utils import import_submodule
from . import config, inductor_prims, ir, test_operators # NOQA: F401
from .decomposition import decompositions, get_decompositions
from .ir import (
ExpandView,
IndexingConstant,
is_triton,
ops_wrapper,
PermuteView,
Pointwise,
Reduction,
SqueezeView,
TensorBox,
validate_ir,
View,
)
from .utils import (
ceildiv,
decode_device,
is_dynamic,
is_pointwise_use,
pad_listlike,
parallel_num_threads,
sympy_product,
)
from .virtualized import ops, V
log = logging.getLogger(__name__)
lowerings: Dict[torch._ops.OpOverload, Callable[..., Any]] = {}
layout_constraints: Dict[torch._ops.OpOverload, Callable[..., Any]] = {}
fallbacks: Set[torch._ops.OpOverload] = set()
aten = torch.ops.aten
tr_c10d = torch.ops.tr_c10d
prims = torch.ops.prims
needs_realized_inputs: Set[torch._ops.OpOverload] = set()
foreach_ops: Set[torch._ops.OpOverload] = set()
inplace_foreach_ops: Set[torch._ops.OpOverload] = set()
inplaceable_foreach_ops: Dict[torch._ops.OpOverload, torch._ops.OpOverload] = dict()
quantized_decomposed = torch.ops.quantized_decomposed
def assert_nyi(cond, msg):
if not cond:
raise NotImplementedError(f"inductor does not support {msg}")
def add_needs_realized_inputs(fn):
if isinstance(fn, (list, tuple, set)):
return [add_needs_realized_inputs(x) for x in fn]
needs_realized_inputs.add(fn)
if isinstance(fn, torch._ops.OpOverloadPacket):
for overload in fn.overloads():
needs_realized_inputs.add(getattr(fn, overload))
def add_layout_constraint(fn, constraint):
if isinstance(fn, torch._ops.OpOverloadPacket):
for overload in fn.overloads():
layout_constraints[getattr(fn, overload)] = constraint
else:
layout_constraints[fn] = constraint
add_needs_realized_inputs(
[
aten.as_strided,
aten.avg_pool2d,
aten.avg_pool2d_backward,
aten.bmm,
aten.convolution,
aten.convolution_backward,
aten.max_pool2d_with_indices,
aten.max_pool2d_with_indices_backward,
aten.mm,
aten.upsample_nearest2d,
aten._upsample_nearest_exact2d,
aten.upsample_bicubic2d,
aten._int_mm,
]
)
# TODO(jansel): ezyang says we won't need this in the future, try removing it
# based on https://github.com/pytorch/pytorch/blob/9e3eb329df8f701/c10/core/ScalarType.h#L28
DTYPE_ID_LOOKUP = {
0: torch.uint8,
1: torch.int8,
2: torch.int16,
3: torch.int32,
4: torch.int64,
5: torch.float16,
6: torch.float32,
7: torch.float64,
8: torch.complex32,
9: torch.complex64,
10: torch.complex32,
11: torch.bool,
15: torch.bfloat16,
# TODO(jansel): add quantized types?
# _(c10::qint8, QInt8) /* 12 */
# _(c10::quint8, QUInt8) /* 13 */
# _(c10::qint32, QInt32) /* 14 */
# _(c10::quint4x2, QUInt4x2) /* 16 */
# _(c10::quint2x4, QUInt2x4) /* 17 */
}
def decode_dtype(dtype: int):
if not isinstance(dtype, int):
return dtype
assert dtype in DTYPE_ID_LOOKUP, f"id {dtype} missing from DTYPE_ID_LOOKUP"
dtype = DTYPE_ID_LOOKUP[dtype]
return dtype
def is_integer_type(x):
if isinstance(x, TensorBox):
return is_integer_dtype(x.get_dtype()) or is_boolean_dtype(x.get_dtype())
elif isinstance(x, sympy.Expr):
return x.is_integer is True # type: ignore[attr-defined]
else:
return isinstance(x, int)
def is_boolean_type(x):
if isinstance(x, TensorBox):
return is_boolean_dtype(x.get_dtype())
else:
return isinstance(x, bool)
def get_promoted_dtype(*args, type_promotion_kind: ELEMENTWISE_TYPE_PROMOTION_KIND):
def construct_input(inp):
if isinstance(inp, (Number, sympy.Expr)):
return inp
else:
assert hasattr(inp, "get_dtype")
dim = len(inp.get_size())
# construct a tmp tensor to feed into torch.result_type
return torch.zeros([1] * dim, dtype=inp.get_dtype())
inps = [construct_input(arg) for arg in args]
_, dtype = elementwise_dtypes(*inps, type_promotion_kind=type_promotion_kind)
return dtype
def get_overloads(aten_fn):
if not isinstance(aten_fn, (list, tuple)):
aten_fn = [aten_fn]
else:
aten_fn = list(aten_fn)
for fn in list(aten_fn):
if isinstance(fn, torch._ops.OpOverloadPacket):
for overload in fn.overloads():
other_fn = getattr(fn, overload)
if other_fn not in lowerings:
aten_fn.append(other_fn)
return aten_fn
def transform_args(args, broadcast, type_promotion_kind, convert_input_to_bool):
indices = [i for i, x in enumerate(args) if isinstance(x, TensorBox)]
if (type_promotion_kind or convert_input_to_bool) and indices:
if convert_input_to_bool:
dtype = torch.bool
else:
# FIXME that's a crude approximation for promoting args
promoting_args = [
a
for a in args
if isinstance(a, (Number, sympy.Expr)) or hasattr(a, "dtype")
]
dtype = get_promoted_dtype(
*promoting_args, type_promotion_kind=type_promotion_kind
)
# sometimes args are an immutable list so we can't mutate them
def promote(arg):
if isinstance(arg, TensorBox):
return to_dtype(arg, dtype)
elif isinstance(arg, ir.Constant):
return ir.Constant(arg.value, dtype, args[indices[0]].get_device())
else:
return arg
args = [promote(a) for a in args]
if broadcast and indices:
for i, x in zip(indices, broadcast_tensors(*[args[i] for i in indices])):
args[i] = x
for i in range(len(args)):
if isinstance(args[i], ir.Constant):
args[i] = ExpandView.create(args[i], list(args[indices[0]].get_size()))
return args
def _register_foreach_lowering(aten_fn, decomp_fn):
"""
Add a foreach lowering to lowerings dict.
Arguments:
aten_fn: torch.ops.aten.* fn we are lowering
decomp_fn: alternate implementation on our IR
broadcast: True to apply broadcasting to tensor inputs
type_promotion_kind: kind of type promotion applied to tensor inputs, `None` means no type promotion
convert_input_to_bool: some logical ops require inputs are converted to bool
"""
@functools.wraps(decomp_fn)
def wrapped(*args, **kwargs):
assert len(args) <= 2
out = decomp_fn(*args, **kwargs)
validate_ir(out)
return out
aten_fns = get_overloads(aten_fn)
foreach_ops.update(aten_fns)
lowerings.update(dict.fromkeys(aten_fns, wrapped))
return wrapped
def _register_lowering(
aten_fn, decomp_fn, broadcast, type_promotion_kind, convert_input_to_bool
):
"""
Add a lowering to lowerings dict
Arguments:
aten_fn: torch.ops.aten.* fn we are lowering
decomp_fn: alternate implementation on our IR
broadcast: True to apply broadcasting to tensor inputs
type_promotion_kind: kind of type promotion applied to tensor inputs, `None` means no type promotion
convert_input_to_bool: some logical ops require inputs are converted to bool
"""
@functools.wraps(decomp_fn)
def wrapped(*args, **kwargs):
args: Union[List[Any], Tuple[Any, ...], Dict[Any, Any]] = list(args)
unpacked = False
# TODO maybe we need to use pytrees here
if len(args) == 1 and isinstance(args[0], (list, tuple)):
unpacked = True
args = args[0]
# explicitly assert for "out=" ops for better error messages
assert not any(
x == "out" for x in kwargs.keys()
), "out= ops aren't yet supported"
# kwargs tensors not supported yet unless it's a fallback op
assert not any(isinstance(x, TensorBox) for x in kwargs.values()) or all(
fn in fallbacks for fn in aten_fn
)
args = transform_args(
args, broadcast, type_promotion_kind, convert_input_to_bool
)
if unpacked:
args = [args]
out = decomp_fn(*args, **kwargs)
validate_ir(out)
return out
aten_fn = get_overloads(aten_fn)
lowerings.update(dict.fromkeys(aten_fn, wrapped))
return wrapped
def register_lowering(
aten_fn,
broadcast=False,
type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.DEFAULT,
convert_input_to_bool=False,
):
"""
Shim to support decorator syntax.
"""
return functools.partial(
_register_lowering,
aten_fn,
broadcast=broadcast,
type_promotion_kind=type_promotion_kind,
convert_input_to_bool=convert_input_to_bool,
)
def broadcast_symbolic_shapes(a, b):
"""
Broadcasting logic based on symbolic shapes.
We give the shapes 0 and 1 concrete values, while all other shapes
are symbolic sympy formulas.
"""
output = []
for x, y in itertools.zip_longest(
reversed(a), reversed(b), fillvalue=sympy.Integer(1)
):
if y == 1:
output.append(x)
elif x == 1:
output.append(y)
else:
V.graph.sizevars.guard_equals(x, y)
if len(sympy.expand(y).free_symbols) < len(sympy.expand(x).free_symbols):
output.append(y) # prefer shorter formula
else:
output.append(x)
return tuple(reversed(output))
def promote_constants(inputs, override_return_dtype=None, type_promotion_kind=None):
assert (
override_return_dtype is None or type_promotion_kind is None
), "only one of override_return_dtype or type_promotion_kind may be given"
if override_return_dtype is None and type_promotion_kind is None:
type_promotion_kind = ELEMENTWISE_TYPE_PROMOTION_KIND.DEFAULT
if not any(isinstance(x, (sympy.Expr, int, float)) for x in inputs):
return inputs
if all(isinstance(x, (int, float, sympy.Expr)) for x in inputs):
dtype = override_return_dtype or get_promoted_dtype(
*inputs, type_promotion_kind=type_promotion_kind
)
def const_func(x):
if isinstance(x, sympy.Expr):
return ir.IndexingConstant(x, dtype, decode_device(None))
else:
return ir.Constant(x, dtype, decode_device(None))
return [const_func(x) for x in inputs]
ex = next(x for x in inputs if isinstance(x, (TensorBox, ExpandView)))
out = []
for x in inputs:
if isinstance(x, (int, float)):
out.append(
ExpandView.create(
ir.Constant(x, ex.get_dtype(), ex.get_device()), list(ex.get_size())
)
)
elif isinstance(x, sympy.Expr):
out.append(
ExpandView.create(
IndexingConstant(x, ex.get_dtype(), ex.get_device()),
list(ex.get_size()),
)
)
else:
out.append(x)
return out
def make_pointwise(
fn,
override_return_dtype=None,
override_device=None,
override_fn_when_input_bool=None,
override_fn_when_cuda_float64=None,
allow_alpha=False,
triton_fallback=None,
):
def inner(*inputs: List[TensorBox], alpha=None):
if triton_fallback is not None and any(map(is_triton, inputs)):
assert not allow_alpha # not implemented
return triton_fallback(*inputs)
inputs = promote_constants(inputs, override_return_dtype)
if allow_alpha:
if alpha is not None and alpha != 1:
inputs = list(inputs)
inputs[-1] = mul(inputs[-1], alpha)
else:
assert alpha is None
loaders = [x.make_loader() for x in inputs]
ranges = inputs[0].get_size()
dtype = override_return_dtype or inputs[0].get_dtype()
is_cuda = decode_device(inputs[0].get_device()).type == "cuda"
for other in inputs[1:]:
assert isinstance(other, ir.BaseConstant) or len(ranges) == len(
other.get_size()
), f"ndim mismatch {fn} {ranges} {other.get_size()}"
def inner_fn(index):
assert len(index) == len(ranges), f"wrong ndim {index} {ranges}"
if dtype == torch.bool and override_fn_when_input_bool is not None:
return override_fn_when_input_bool(*[load(index) for load in loaders])
elif override_fn_when_cuda_float64 and is_cuda and dtype == torch.float64:
return override_fn_when_cuda_float64(*[load(index) for load in loaders])
else:
return fn(*[load(index) for load in loaders])
if not override_device:
device = None
for i in inputs:
if i.get_device().type == "cuda":
device = i.get_device()
break
if not device:
device = inputs[0].get_device()
device = override_device or device
return Pointwise.create(
device=device,
dtype=dtype,
inner_fn=inner_fn,
ranges=ranges,
)
return inner
def make_foreach_pointwise(pw_fn, allow_alpha=False):
def inner(*inputs: List[List[TensorBox]], alpha=1):
# group by device, whether any of the inputs are dynamic, and whether their types match
# (proxy for type promotion)
def group_args(arg_pairs):
out = defaultdict(list)
for i, args in enumerate(arg_pairs):
use_foreach = not is_dynamic(*args)
device = None
for t in args:
if isinstance(t, TensorBox):
device = t.data.get_device()
break
assert (
device is not None
), "foreach op should have at least one tensor arg"
out[(device, use_foreach)].append((i, args))
return out
realize_outputs = (
len(V.graph.current_node.users) == 0
or V.graph.current_node.target in inplace_foreach_ops
)
for node in V.graph.current_node.users:
for user in node.users:
if not (user.op == "call_function" and (user.target in foreach_ops)):
realize_outputs = True
a_list_input = None
for input in inputs:
if isinstance(input, (list, tuple)):
a_list_input = input
break
assert (
a_list_input is not None
), "at least one input must be a list to a foreach op"
# broadcast scalar inputs to match length of list inputs
broadcast_inputs = []
for input in inputs:
if not isinstance(input, (list, tuple)):
broadcast_inputs.append([input] * len(a_list_input))
else:
broadcast_inputs.append(input)
groups = group_args(zip(*broadcast_inputs))
outputs = [None] * len(a_list_input)
for (device, use_foreach), group in groups.items():
buffer_list = []
for (
output_ind,
args,
) in group:
if allow_alpha:
output = pw_fn(*args, alpha=alpha)
else:
output = pw_fn(*args)
outputs[output_ind] = output
if device.type == "cuda" and use_foreach and realize_outputs:
buffer_list.append(output.realize())
if buffer_list:
V.graph.register_list(buffer_list)
assert all(x is not None for x in outputs)
return outputs
return inner
def to_dtype(x: TensorBox, dtype: torch.dtype, copy=False):
src_dtype = x.get_dtype()
if src_dtype == dtype:
return clone(x) if copy else x
def _to_dtype(x):
return ops.to_dtype(x, dtype, src_dtype=src_dtype)
return make_pointwise(_to_dtype, override_return_dtype=dtype)(x)
@register_lowering(prims.convert_element_type, type_promotion_kind=None)
def _convert_element_type(x: TensorBox, dtype: torch.dtype):
if dtype.is_complex or x.get_dtype().is_complex:
if x.get_size():
# Decompose since aa aten fallback is more friendly for c++ codegen.
# This decompostion doesn't work for empty tensor, which needs more investigation.
dst = empty_like(x, dtype=dtype)
ir.InplaceCopyFallback.create(dst, x)
return dst
else:
return fallback_handler(
prims.convert_element_type.default, add_to_fallback_set=False
)(x, dtype)
return to_dtype(x, dtype, copy=True)
def to_dtype_bitcast(x: TensorBox, dtype: torch.dtype, *, copy=False):
x_dtype = x.get_dtype()
if x_dtype == dtype:
return clone(x) if copy else x
def _get_primitive_bitwidth(dtype):
if dtype.is_floating_point:
return torch.finfo(dtype).bits
else:
return torch.iinfo(dtype).bits
src_bits = _get_primitive_bitwidth(x_dtype)
dst_bits = _get_primitive_bitwidth(dtype)
if src_bits != dst_bits:
raise NotImplementedError(
f"bitcast {x_dtype} to different bitwidth type {dtype} is not supported yet."
)
def _to_dtype_bitcast(x):
# Because we may promote tensor type from float16 or bfloat16
# to float, we will need to pass the original src dtype (i.e. x_dtype),
# which is used for correctly constructing type conversion before bitcast,
# which requires the bitwidth of the input tensor type is the same as the
# target type.
return ops.to_dtype_bitcast(x, dtype, x_dtype)
return make_pointwise(_to_dtype_bitcast, override_return_dtype=dtype)(x)
@register_lowering(aten.view.dtype, type_promotion_kind=None)
def _view_dtype(x: TensorBox, dtype: torch.dtype):
if dtype.is_complex or x.get_dtype().is_complex:
return TensorBox.create(
ir.ComplexView.create(torch.ops.aten.view.dtype, x, dtype)
)
return to_dtype_bitcast(x, dtype, copy=True)
def to_device(x: TensorBox, device: torch.device, *, copy=False):
device = decode_device(device)
if x.get_device() == device:
return clone(x) if copy else x
return TensorBox.create(ir.DeviceCopy.create(x, device))
@register_lowering(prims.device_put, type_promotion_kind=None)
def _device_put(x: TensorBox, device: torch.device):
return to_device(x, device, copy=True)
def register_pointwise(
aten_fn,
name=None,
broadcast=True,
type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.DEFAULT,
convert_input_to_bool=False,
override_return_dtype=None,
override_fn_when_input_bool=None,
allow_alpha=False,
use_libdevice_for_f64=False,
triton_fallback=None,
):
"""A pointwise function that maps ops.{name} to inputs"""
name = name or aten_fn.__name__
fn = ops_wrapper(name)
if use_libdevice_for_f64:
fn_libdevice = ops_wrapper("libdevice_" + name)
if override_fn_when_input_bool is not None:
override_fn_when_input_bool = ops_wrapper(override_fn_when_input_bool)
fn = make_pointwise(
fn,
override_return_dtype=override_return_dtype,
override_fn_when_input_bool=override_fn_when_input_bool,
override_fn_when_cuda_float64=fn_libdevice if use_libdevice_for_f64 else None, # type: ignore[possibly-undefined]
allow_alpha=allow_alpha,
triton_fallback=triton_fallback,
)
fn = register_lowering(
aten_fn,
broadcast=broadcast,
type_promotion_kind=type_promotion_kind,
convert_input_to_bool=convert_input_to_bool,
)(fn)
if hasattr(prims, name):
register_lowering(
getattr(prims, name),
type_promotion_kind=None,
convert_input_to_bool=convert_input_to_bool,
)(fn)
return fn
def register_frexp():
"""A pointwise function that maps ops.frexp to inputs"""
name = "frexp"
frexp = ops_wrapper("frexp")
def frexp0(*args, **kwargs):
return frexp(*args, **kwargs)[0]
def frexp1(*args, **kwargs):
return frexp(*args, **kwargs)[1]
pw_fns = [
make_pointwise(frexp0),
make_pointwise(frexp1, override_return_dtype=torch.int32),
]
def fn(*args, **kwargs):
return pw_fns[0](*args, **kwargs), pw_fns[1](*args, **kwargs)
fn = register_lowering(
aten.frexp,
)(fn)
if hasattr(prims, name):
register_lowering(
getattr(prims, name),
type_promotion_kind=None,
)(fn)
return fn
register_frexp()
def register_foreach_pointwise(
aten_fn,
pointwise_lowering_fn,
allow_alpha=False,
):
fn = make_foreach_pointwise(pointwise_lowering_fn, allow_alpha=allow_alpha)
fn = _register_foreach_lowering(aten_fn, fn)
return fn
@register_lowering(aten.where, broadcast=False, type_promotion_kind=None)
def where(cond, a, b):
def fn(*args):
return ops.where(*args)
if isinstance(a, (float, int)):
a = constant_like(a)(b)
if isinstance(b, (float, int)):
b = constant_like(b)(a)
args = [cond, a, b]
dtype = get_promoted_dtype(
args[1], args[2], type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.DEFAULT
)
indices = [i for i, x in enumerate(args) if isinstance(x, TensorBox)]
for i, x in zip(indices, broadcast_tensors(*[args[i] for i in indices])):
args[i] = x
for i in range(len(args)):
if isinstance(args[i], ir.Constant):
args[i] = ExpandView.create(args[i], list(args[indices[0]].get_size()))
return make_pointwise(fn, override_return_dtype=dtype)(
args[0], to_dtype(args[1], dtype), to_dtype(args[2], dtype)
)
@register_lowering(aten.broadcast_tensors, broadcast=False, type_promotion_kind=None)
def broadcast_tensors(*inputs):
if len(inputs) == 1 and isinstance(inputs[0], (list, tuple)):
return broadcast_tensors(*inputs[0])
target: List[sympy.Expr] = functools.reduce(
broadcast_symbolic_shapes, [x.get_size() for x in inputs], []
)
outputs = []
for x in inputs:
sizes = x.get_size()
if len(sizes) != len(target) or any(
((a == 1 and b != 1) or (a != 1 and b == 1)) for a, b in zip(sizes, target)
):
x = expand(x, target)
outputs.append(x)
return outputs
@register_lowering([aten.alias, aten.detach, aten.detach_, aten.lift, prims.view_of])
def nop(x):
return x # AOT autograd handles this for us
if hasattr(aten, "lift_fresh"):
register_lowering(aten.lift_fresh)(nop)
@register_lowering(aten.squeeze, type_promotion_kind=None)
def squeeze(x, dim=None):
assert isinstance(x, TensorBox)
if dim is None:
return TensorBox(SqueezeView.create(x.data))
dim = canonicalize_dims(len(x.get_size()), dim)
dims = set((dim,) if not isinstance(dim, tuple) else dim)
new_shape = []
for d, s in enumerate(x.get_size()):
if not (d in dims and V.graph.sizevars.evaluate_expr(sympy.Eq(s, 1))):
new_shape.append(s)
# squeeze does nothing if the size isn't 1
return view(x, new_shape) if new_shape != x.get_size() else x
@register_lowering(aten.squeeze_copy, type_promotion_kind=None)
def squeeze_copy(x, dim=None):
return clone(squeeze(x, dim))
@register_lowering([aten.squeeze_])
def squeeze_(x, dim=None):
val = squeeze(x, dim)
assert isinstance(x, TensorBox)
assert isinstance(val, TensorBox)
x.data = val.data
return x
@register_lowering(aten.isinf)
def isinf(x):
if is_integer_type(x):
return full_like(x, False, dtype=torch.bool)
fn = ops_wrapper("isinf")
return make_pointwise(fn, override_return_dtype=torch.bool)(x)
@register_lowering(aten.isnan)
def isnan(x):
if is_integer_type(x):
return full_like(x, False, dtype=torch.bool)
fn = ops_wrapper("isnan")
return make_pointwise(fn, override_return_dtype=torch.bool)(x)
@register_lowering(aten.ceil)
def ceil(x):
if is_integer_type(x):
return clone(x)
fn = ops_wrapper("ceil")
return make_pointwise(fn)(x)
@register_lowering(aten.floor)
def floor(x):
if is_integer_type(x):
return clone(x)
fn = ops_wrapper("floor")
return make_pointwise(fn)(x)
@register_lowering(aten.round.default)
def round(x):
if is_integer_type(x):
return clone(x)
else:
fn = ops_wrapper("round")
return make_pointwise(fn)(x)
@register_lowering(aten.trunc)
def trunc(x):
if is_integer_type(x):
return clone(x)
fn = ops_wrapper("trunc")
return make_pointwise(fn)(x)
@register_lowering(aten.expand, type_promotion_kind=None)
def expand(x, sizes):
(x,) = promote_constants([x])
if isinstance(x, ir.BaseConstant):
return ExpandView.create(x, tuple(sizes))
assert isinstance(x, TensorBox)
assert isinstance(sizes, (list, tuple))
if tuple(x.get_size()) == tuple(sizes):
return x
if not any(V.graph.sizevars.shape_env.is_unbacked_symint(s) for s in x.get_size()):
x_size_product = V.graph.sizevars.size_hint(sympy_product(x.get_size()))
# TODO: It would be better to realize the input if any of its sizes
# are unbacked, because typically the size will be non-zero. However,
# this cannot be done directly as below as we'll choke on the size_hint
# here
if x_size_product > 0 and not any(
V.graph.sizevars.shape_env.is_unbacked_symint(s) for s in sizes
):
# maybe realize input before broadcasting it
x.mark_reuse(
V.graph.sizevars.size_hint(sympy_product(sizes)) // x_size_product
)
return TensorBox(ExpandView.create(x.data, tuple(sizes)))
@register_lowering(prims.broadcast_in_dim, type_promotion_kind=None)
def broadcast_in_dim(a, shape, broadcast_dimensions):
s = list(shape)
for broadcast_dimension in broadcast_dimensions:
s[broadcast_dimension] = -1
v = a
for idx, x in enumerate(s):
if x != -1:
v = unsqueeze(v, idx)
return expand(v, shape)
@register_lowering(aten.expand_as, type_promotion_kind=None)
def expand_as(x, y):
return expand(x, y.get_size())
@register_lowering(aten.repeat)
def repeat(x, repeats):
old_size = list(x.get_size())
if len(repeats) > len(old_size):
old_size = [sympy.Integer(1)] * (len(repeats) - len(old_size)) + old_size
x = view(x, list(old_size))
assert len(repeats) == len(x.get_size())
new_size = list(x.get_size())
zero_tensor = False
for i in range(len(repeats)):
if repeats[i] == 0:
zero_tensor = True
new_size[i] = new_size[i] * repeats[i]
if zero_tensor:
return empty(new_size, dtype=x.get_dtype(), device=x.get_device())
if all((a == 1 or b == 1) for a, b in zip(repeats, old_size)):
return expand(x, new_size)
x_loader: Callable[[Any], Any]
def inner_fn(index):
assert len(index) == len(repeats)
index = list(index)
for i in range(len(repeats)):
if repeats[i] != 1:
if old_size[i] == 1:
index[i] = sympy.Integer(0)
else:
index[i] = ModularIndexing(index[i], 1, old_size[i])
return x_loader(index)
old_size_product = V.graph.sizevars.size_hint(sympy_product(old_size))
if old_size_product > 0:
# maybe realize the input
x.mark_reuse(
V.graph.sizevars.size_hint(sympy_product(new_size)) // old_size_product
)
x_loader = x.make_loader()
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=inner_fn,
ranges=list(new_size),
)
@register_lowering(aten._unsafe_view, type_promotion_kind=None)
@register_lowering(aten.view, type_promotion_kind=None)
@register_lowering(aten.reshape, type_promotion_kind=None)
def view(x, sizes):
assert isinstance(x, TensorBox)
assert isinstance(sizes, (list, tuple))
return TensorBox(View.create(x.data, sizes))
@register_lowering(aten.permute, type_promotion_kind=None)
def permute(x, dims):
assert isinstance(x, TensorBox)
assert isinstance(dims, (list, tuple))
return TensorBox(PermuteView.create(x.data, tuple(dims)))
@register_lowering(aten.slice, type_promotion_kind=None)
def slice_(x, dim=0, start=0, end=2**63, step=1):
assert isinstance(x, TensorBox)
dim = _validate_dim(x, dim, 0)
dim_size = x.get_size()[dim]
return TensorBox(ir.SliceView.create(x.data, dim, start, end, step))
@register_lowering(aten.as_strided, type_promotion_kind=None)
def as_strided(x, size, stride, storage_offset=None):
if isinstance(x, TensorBox) and isinstance(x.data, ir.BaseView):
# as_strided ignores views
x = x.data.unwrap_view()
x.realize()
if not ir.is_storage_and_layout(x):
raise NotImplementedError(f"unrealized as_strided({x}, ...)")
storage, old_layout = ir.as_storage_and_layout(x)
new_layout = ir.FixedLayout(
old_layout.device,
old_layout.dtype,
[sympy.expand(s) for s in size],
[sympy.expand(s) for s in stride],
sympy.expand(storage_offset or 0),
)
return TensorBox(ir.ReinterpretView(storage, new_layout))
@register_lowering(aten.as_strided_, type_promotion_kind=None)
def as_strided_(x, size, stride, storage_offset=None):
assert isinstance(x, TensorBox)
x.data = as_strided(x, size, stride, storage_offset).data
return x
@register_lowering(aten.as_strided_copy, type_promotion_kind=None)
def as_strided_copy(x, size, stride, storage_offset=None):
result = as_strided(x, size, stride, storage_offset)
return clone(result)
def pointwise_cat(inputs, dim=0):
# (inclusive, exclusive)
inputs_ranges: List[Tuple[sympy.Expr, sympy.Expr]] = []
prev_end = 0
for inp in inputs:
inputs_ranges.append((prev_end, prev_end + inp.get_size()[dim])) # type: ignore[arg-type]
prev_end = inputs_ranges[-1][-1] # type: ignore[assignment]
inputs_loaders = [inp.make_loader() for inp in inputs]
def inner_fn(idx):
idx_dim = ops.index_expr(idx[dim], torch.int64)
masks = []
masked_loads = []
for i in range(len(inputs)):
start = (
ops.constant(0, torch.int64)
if i == 0
else ops.index_expr(inputs_ranges[i][0], torch.int64)
)
end = ops.index_expr(inputs_ranges[i][1], torch.int64)
start_cond = ops.ge(idx_dim, start)
end_cond = ops.lt(idx_dim, end)
if i == 0:
mask = end_cond
elif i == len(inputs) - 1:
mask = start_cond
else:
mask = ops.and_(start_cond, end_cond)
masks.append(mask)
idx_load = list(idx)
# if we're concatting [4], [2]
# when we index the second tensor for 5 we want to index 5 - 4
idx_load[dim] -= inputs_ranges[i][0]
masked_loads.append(
ops.masked(
mask,
lambda: inputs_loaders[i](idx_load),
0.0, # this value should be unused
),
)
next_val = masked_loads[-1]
for i in range((len(inputs)) - 2, -1, -1):
next_val = ops.where(
masks[i],
masked_loads[i],
next_val,
)
return next_val
new_size = list(inputs[0].get_size())
new_size[dim] = inputs_ranges[-1][-1]
return Pointwise.create(
device=inputs[0].get_device(),
dtype=inputs[0].get_dtype(),
inner_fn=inner_fn,
ranges=new_size,
)
@register_lowering(quantized_decomposed.quantize_per_channel, type_promotion_kind=None)
def quantized_decomposed_quantize_per_channel(
input: TensorBox,
scales: TensorBox,
zero_points: TensorBox,
axis: int,
quant_min: int,
quant_max: int,
dtype: torch.dtype,
) -> TensorBox:
assert len(scales.get_size()) == 1, "expect scales 1 dim"
assert len(zero_points.get_size()) == 1, "expect zero_points 1 dim"
if input.get_dtype() == torch.bfloat16:
input = to_dtype(input, torch.float32)
assert (
input.get_dtype() == torch.float32
), f"Expecting input to have dtype torch.float32, but got dtype: {input.get_dtype()}"
assert axis < len(
input.get_size()
), f"Expecting axis to be < {len(input.get_size())}"
input_loader = input.make_loader()
scales_loader = scales.make_loader()
zero_points_loader = zero_points.make_loader()
def inner_fn(idx):
channel_idx = (idx[axis],)
input = input_loader(idx)
scale = scales_loader(channel_idx)
zero_point = zero_points_loader(channel_idx)
qmin, qmax = _create_constants(quant_min, quant_max, dtype=torch.float32)
if scales.dtype != torch.float32:
scale = ops.to_dtype(scale, torch.float32)
if zero_points.dtype != torch.int32:
zero_point = ops.to_dtype(zero_point, torch.int32)
inv_scale = ops.reciprocal(scale)
val = ops.round(input * inv_scale) + zero_point
clamped = ops.maximum(qmin, ops.minimum(qmax, val))
return ops.to_dtype(clamped, dtype)
return Pointwise.create(
device=input.get_device(),
dtype=dtype,
inner_fn=inner_fn,
ranges=input.get_size(),
)
@register_lowering(
quantized_decomposed.dequantize_per_channel, type_promotion_kind=None
)
def quantized_decomposed_dequantize_per_channel(
input: TensorBox,
scales: TensorBox,
zero_points: TensorBox,
axis: int,
quant_min: int,
quant_max: int,
dtype: torch.dtype,
) -> TensorBox:
assert len(scales.get_size()) == 1, "expect scales 1 dim"
assert len(zero_points.get_size()) == 1, "expect zero_points 1 dim"
assert (
input.get_dtype() == dtype
), f"Expecting input to have dtype {dtype}, but got dtype: {input.get_dtype()}"
assert axis < len(
input.get_size()
), f"Expecting axis to be < {len(input.get_size())}"
input_loader = input.make_loader()
scales_loader = scales.make_loader()
zero_points_loader = zero_points.make_loader()
def inner_fn(idx):
channel_idx = (idx[axis],)
input = input_loader(idx)
scale = scales_loader(channel_idx)
zero_point = zero_points_loader(channel_idx)
if scales.dtype != torch.float32:
scale = ops.to_dtype(scale, torch.float32)
if zero_points.dtype != torch.float32:
zero_point = ops.to_dtype(zero_point, torch.float32)
val = ops.sub(ops.to_dtype(input, torch.float32), zero_point) * scale
return val
return Pointwise.create(
device=input.get_device(),
dtype=torch.float32,
inner_fn=inner_fn,
ranges=input.get_size(),
)
@register_lowering(aten.cat)
def cat(inputs, dim=0):
if all(input.get_dtype() in [torch.int8, torch.uint8] for input in inputs):
# TODO <leslie> Remove this fallback when we support vectorization
# code gen with uint8 data type directly.
for input in inputs:
input.realize()
if all(len(input.get_size()) == 4 for input in inputs):
inputs, _ = require_channels_last(aten.cat, *inputs)
return fallback_handler(aten.cat.default)(inputs, dim)
if len(inputs) == 1:
return clone(inputs[0])
dim = _validate_dim(inputs[0], dim, 0)
dtype = get_promoted_dtype(
*inputs, type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.DEFAULT
)
inputs = [to_dtype(inp, dtype) for inp in inputs]
def unwrap_tensor(x: Union[TensorBox, ir.StorageBox]) -> ir.IRNode:
if isinstance(x, TensorBox):
if isinstance(x.data, ir.BaseView):
return x.data.unwrap_view()
else:
return x.data
if isinstance(x, ir.StorageBox):
return x.data
return x
def should_lower_cat_input(x) -> bool:
# Unrealized inputs will not be storage and layouts, and we dont want to realize
# them in case we want to fuse
if ir.is_storage_and_layout(x):
storage, _ = ir.as_storage_and_layout(x, freeze=False)
return not ir.ConcatKernel.can_realize_into_without_copy(storage)
if isinstance(x, (TensorBox, ir.StorageBox)):
return should_lower_cat_input(unwrap_tensor(x))
if isinstance(x, ir.Pointwise):
return True
return False
def is_reduction(t):
return isinstance(t, ir.ComputedBuffer) and isinstance(t.data, ir.Reduction)
def can_fuse_reduction(t):
if isinstance(t, (TensorBox, ir.StorageBox)):
return can_fuse_reduction(unwrap_tensor(t))
return (
is_reduction(t)
or isinstance(t, ir.Pointwise)
and any(
can_fuse_reduction(V.graph.get_buffer(read))
for read in t.get_read_names()
)
)
# fusing reducutions into computed concat buffer can cause regressions.
fusable_reduction = any(can_fuse_reduction(t) for t in inputs)
# TODO: We observed negative performance impact of pointwise_cat optimization on CPU so disabled it.
# We will revisit this later after enabling vectorization on index_expr.
if inputs[0].get_device().type == "cpu" or fusable_reduction:
return TensorBox(ir.ConcatKernel.create(inputs, dim))
def op_count(x):
if isinstance(x, (TensorBox, ir.StorageBox)):
return op_count(unwrap_tensor(x))
# this will correspond to a direct memory read
if not isinstance(x, ir.Pointwise):
return 0
count = x.inner_fn_opcount()
for read in x.get_read_names():
count += op_count(V.graph.get_buffer(read))
return count
# as of inputs increase, possibility for register spilling also increases
# past a certain threshold of inputs we only fuse if the if the input kernels
# are simple
# not sure if we want to expose to users via config since logic may change in future
MAX_COMPLEX_POINTWISE_CAT = 8
MAX_SIMPLE_OP_COUNT = 2
if len(inputs) <= MAX_COMPLEX_POINTWISE_CAT or (
(len(inputs) <= config.max_pointwise_cat_inputs)
and all(op_count(t) <= MAX_SIMPLE_OP_COUNT for t in inputs)
):
pointwise_uses = all(is_pointwise_use(use) for use in V.current_node.users)
all_pointwise_inputs = all(should_lower_cat_input(inp) for inp in inputs)
any_pointwise_inputs = any(should_lower_cat_input(inp) for inp in inputs)
if all_pointwise_inputs or (any_pointwise_inputs and pointwise_uses):
return pointwise_cat(inputs, dim)
return TensorBox(ir.ConcatKernel.create(inputs, dim))
@register_lowering(aten.diagonal, type_promotion_kind=None)
def diagonal(input, offset: int = 0, dim1: int = 0, dim2: int = 1):
original_shape = input.get_size()
num_dims = len(original_shape)
dim1 = canonicalize_dim(idx=dim1, rank=num_dims)
dim2 = canonicalize_dim(idx=dim2, rank=num_dims)
check(
dim1 != dim2, lambda: f"diagonal dimensions cannot be identical {dim1}, {dim2}"
)
offset_negative = V.graph.sizevars.evaluate_expr(sympy.Lt(offset, 0))
if offset_negative:
diag_size = max(min(original_shape[dim1] + offset, original_shape[dim2]), 0)
else:
diag_size = max(min(original_shape[dim1], original_shape[dim2] - offset), 0)
base_idx = (0, 0)
if offset_negative:
base_idx = (-offset, 0)
else:
base_idx = (0, offset)
sizes = [s for i, s in enumerate(original_shape) if i not in (dim1, dim2)]
sizes.append(diag_size)
def reindexer(idx):
diag_idx = idx[-1]
original_idx = [0] * len(original_shape)
cur_dim = 0
for d in range(num_dims):
if d == dim1:
original_idx[d] = diag_idx + base_idx[0]
elif d == dim2:
original_idx[d] = diag_idx + base_idx[1]
else:
original_idx[d] = idx[cur_dim]
cur_dim += 1
assert cur_dim == len(original_shape) - 2
return original_idx
return TensorBox(ir.GenericView.create(input, sizes, reindexer))
@register_lowering(aten.diagonal_copy, type_promotion_kind=None)
def diagonal_copy(input, offset: int = 0, dim1: int = 0, dim2: int = 1):
return clone(diagonal(input, offset, dim1, dim2))
@register_lowering(aten.diagonal_scatter, type_promotion_kind=None)
def diagonal_scatter(input, src, offset: int = 0, dim1: int = 0, dim2: int = 1):
output = clone(input)
target = diagonal(output, offset, dim1, dim2)
mutate_to(target, src)
return output
@register_lowering(aten.select, type_promotion_kind=None)
def select(x, dim, idx):
idx = View.handle_negative_index(idx, x.get_size()[dim])
return squeeze(slice_(x, dim, idx, idx + 1), dim)
@register_lowering(aten.split, type_promotion_kind=None)
def split(x, sizes, dim=0):
dim = _validate_dim(x, dim, 0)
x_size = V.graph.sizevars.evaluate_static_shape(x.get_size()[dim])
if isinstance(sizes, sympy.Expr):
# TODO: We don't have to guard on sizes per se, but the number
# of splits must stay constant
sizes = V.graph.sizevars.evaluate_static_shape(sizes)
if isinstance(sizes, (int, sympy.Integer)):
sizes = [sizes] * ((x_size + sizes - 1) // sizes)
result = []
start = 0
for size in sizes:
end = start + size
result.append(slice_(x, dim, start, end))
start = end
return result
@register_lowering(aten.split_with_sizes, type_promotion_kind=None)
def split_with_sizes(x, sizes, dim=0):
return split(x, sizes, dim)
@register_lowering(aten.unbind, type_promotion_kind=None)
def unbind(x, dim=0):
dim = _validate_dim(x, dim, 0)
x_size = V.graph.sizevars.evaluate_static_shape(x.get_size()[dim])
result = []
for i in range(x_size):
result.append(select(x, dim, i))
return result
@register_lowering(aten.unfold, type_promotion_kind=None)
def unfold(x, dimension, size, step):
sizes = x.get_size()
ndim = len(sizes)
dim = canonicalize_dim(ndim, dimension)
if ndim == 0:
return slice_(unsqueeze(x, 0), end=size)
dim_size = sizes[dim]
sizevars = V.graph.sizevars
sizevars.guard_leq(size, dim_size)
sizevars.guard_lt(0, step) # type: ignore[arg-type]
new_dim_size = FloorDiv(dim_size - size, step) + 1
if sizevars.size_hint(dim_size) > 0:
x.mark_reuse(sizevars.size_hint(CeilDiv(new_dim_size * size, dim_size)))
out_size = [*sizes[:dim], new_dim_size, *sizes[dim + 1 :], size]
def reindexer(idx):
dim_idx = idx[-1] + idx[dim] * step
return (*idx[:dim], dim_idx, *idx[dim + 1 : -1])
return TensorBox(ir.GenericView.create(x, out_size, reindexer))
@register_lowering(aten.unsqueeze, type_promotion_kind=None)
def unsqueeze(x, dim):
dim = _validate_dim(x, dim, 1)
new_shape = list(x.get_size())
new_shape.insert(dim, sympy.Integer(1))
return view(x, new_shape)
@register_lowering(aten.unsqueeze_, type_promotion_kind=None)
def unsqueeze_(x, dim):
val = unsqueeze(x, dim)
assert isinstance(x, TensorBox)
assert isinstance(val, TensorBox)
x.data = val.data
return x
def _validate_dim(x, dim, offset=0):
assert isinstance(dim, int)
ndim = len(x.get_size())
if dim < 0:
dim += ndim + offset
assert 0 <= dim < ndim + offset
return dim
@register_lowering(aten.glu)
def glu(x, dim=-1):
dim = _validate_dim(x, dim, 0)
# TODO: don't guard on static shape here
new_len = V.graph.sizevars.evaluate_static_shape(x.get_size()[dim]) // 2
a = slice_(x, dim, 0, new_len)
b = slice_(x, dim, new_len, new_len * 2)
return mul(a, sigmoid(b))
def register_onednn_fusion_ops():
if torch._C._has_mkldnn:
cpu_needs_realized_inputs = [
torch.ops.mkldnn._convolution_pointwise,
torch.ops.mkldnn._convolution_pointwise_,
torch.ops.mkldnn._convolution_transpose_pointwise,
torch.ops.mkldnn._linear_pointwise,
aten.mkldnn_rnn_layer.default,
torch.ops.onednn.qconv2d_pointwise,
]
@register_lowering(torch.ops.mkldnn._convolution_pointwise)
def convolution_unary(
x: TensorBox,
weight: TensorBox,
bias: TensorBox,
padding,
stride,
dilation,
groups,
attr,
scalars,
algorithm,
):
return TensorBox.create(
ir.ConvolutionUnary.create(
x,
weight,
bias,
padding,
stride,
dilation,
groups,
attr,
scalars,
algorithm,
)
)
@register_lowering(torch.ops.mkldnn._convolution_pointwise.binary)
def convolution_binary(
x: TensorBox,
other: TensorBox,
weight: TensorBox,
bias: TensorBox,
padding,
stride,
dilation,
groups,
binary_attr,
binary_alpha,
unary_attr,
unary_scalars,
unary_algorithm,
):
return TensorBox.create(
ir.ConvolutionBinary.create(
x,
other,
weight,
bias,
padding,
stride,
dilation,
groups,
binary_attr,
binary_alpha,
unary_attr,
unary_scalars,
unary_algorithm,
)
)
@register_lowering(torch.ops.mkldnn._convolution_pointwise_.binary)
def convolution_binary_inplace(
x: TensorBox,
other: TensorBox,
weight: TensorBox,
bias: TensorBox,
padding,
stride,
dilation,
groups,
binary_attr,
binary_alpha,
unary_attr,
unary_scalars,
unary_algorithm,
):
return TensorBox.create(
ir.ConvolutionBinaryInplace.create(
x,
other,
weight,
bias,
padding,
stride,
dilation,
groups,
binary_attr,
binary_alpha,
unary_attr,
unary_scalars,
unary_algorithm,
)
)
@register_lowering(torch.ops.mkldnn._linear_pointwise)
def linear_unary(
x: TensorBox, w: TensorBox, b: TensorBox, attr, scalars, algorithm
):
return TensorBox.create(
ir.LinearUnary.create(x, w, b, attr, scalars, algorithm)
)
@register_lowering(torch.ops.mkldnn._linear_pointwise.binary)
def linear_binary(x: TensorBox, y: TensorBox, w: TensorBox, b: TensorBox, attr):
return TensorBox.create(ir.LinearBinary.create(x, y, w, b, attr))
@register_lowering(torch.ops.mkldnn._convolution_transpose_pointwise)
def convolution_transpose_unary(
x: TensorBox,
weight: TensorBox,
bias: TensorBox,
padding,
output_padding,
stride,
dilation,
groups,
attr,
scalars,
algorithm,
):
return TensorBox.create(
ir.ConvolutionTransposeUnary.create(
x,
weight,
bias,
padding,
output_padding,
stride,
dilation,
groups,
attr,
scalars,
algorithm,
)
)
@register_lowering(aten.mkldnn_rnn_layer.default)
def mkldnn_rnn_layer(
x: TensorBox,
w0: TensorBox,
w1: TensorBox,
w2: TensorBox,
w3: TensorBox,
hx: TensorBox,
cx: TensorBox,
reverse: bool,
batch_sizes: List[int],
mode: int,
hidden_size: int,
num_layers: int,
has_biases: bool,
bidirectional: bool,
batch_first: bool,
train: bool,
):
return pytree.tree_map(
TensorBox.create,
ir.MkldnnRnnLayer.create(
x,
w0,
w1,
w2,
w3,
hx,
cx,
reverse,
batch_sizes,
mode,
hidden_size,
num_layers,
has_biases,
bidirectional,
batch_first,
train,
),
)
@register_lowering(torch.ops.onednn.qconv2d_pointwise, type_promotion_kind=None)
def qconvolution_unary(
x: TensorBox,
x_scale,
x_zp,
packed_weight: TensorBox,
w_scale: TensorBox,
w_zp: TensorBox,
bias: TensorBox,
stride,
padding,
dilation,
groups,
o_inv_scale,
o_zero_point,
output_dtype,
attr,
scalars,
algorithm,
):
return TensorBox.create(
ir.QConvPointWisePT2E.create(
x,
x_scale,
x_zp,
packed_weight,
w_scale,
w_zp,
bias,
stride,
padding,
dilation,
groups,
o_inv_scale,
o_zero_point,
output_dtype,
attr,
scalars,
algorithm,
)
)
@register_lowering(
torch.ops.onednn.qconv2d_pointwise.binary, type_promotion_kind=None
)
def qconvolution_binary(
x: TensorBox,
x_scale,
x_zp,
accum: TensorBox,
accum_scale,
accum_zp,
packed_weight: TensorBox,
w_scale: TensorBox,
w_zp: TensorBox,
bias: TensorBox,
stride,
padding,
dilation,
groups,
o_inv_scale,
o_zero_point,
output_dtype,
binary_attr,
alpha,
unary_attr,
unary_scalars,
unary_algorithmm,
):
if (
binary_attr == "sum"
and output_dtype in [torch.float32, torch.bfloat16]
and accum.get_dtype() in [torch.float32, torch.bfloat16]
and accum.get_dtype() != output_dtype
):
# For int8-mixed-bf16 quantization and inplace add,
# there is case when accum dtype is float32 but output dtype is bfloat16.
# Since the accum will be inplaced changed with post op sum,
# we will do accum dtype convertion here.
accum = to_dtype(accum, output_dtype)
return TensorBox.create(
ir.QConvPointWiseBinaryPT2E.create(
x,
x_scale,
x_zp,
accum,
accum_scale,
accum_zp,
packed_weight,
w_scale,
w_zp,
bias,
stride,
padding,
dilation,
groups,
o_inv_scale,
o_zero_point,
output_dtype,
binary_attr,
alpha,
unary_attr,
unary_scalars,
unary_algorithmm,
)
)
@register_lowering(torch.ops.onednn.qlinear_pointwise, type_promotion_kind=None)
def qlinear_unary(
x: TensorBox,
x_scale,
x_zp,
packed_weight: TensorBox,
w_scale: TensorBox,
w_zp: TensorBox,
bias: TensorBox,
o_inv_scale,
o_zero_point,
output_dtype,
attr,
scalars,
algorithm,
):
return TensorBox.create(
ir.QLinearPointwisePT2E.create(
x,
x_scale,
x_zp,
packed_weight,
w_scale,
w_zp,
bias,
o_inv_scale,
o_zero_point,
output_dtype,
attr,
scalars,
algorithm,
)
)
if torch._C.has_mkl:
cpu_needs_realized_inputs.append(torch.ops.mkl._mkl_linear)
@register_lowering(torch.ops.mkl._mkl_linear)
def mkl_packed_linear(
x: TensorBox,
packed_w: TensorBox,
orig_w: TensorBox,
b: TensorBox,
batch_size,
):
result = TensorBox.create(
ir.MKLPackedLinear.create(x, packed_w, orig_w, batch_size)
)
if b is not None:
result = add(result, b)
return result
add_needs_realized_inputs(cpu_needs_realized_inputs)
else:
pass
register_onednn_fusion_ops()
def fallback_handler(kernel, add_to_fallback_set=True):
if add_to_fallback_set:
fallbacks.add(kernel)
def handler(*args, **kwargs):
return pytree.tree_map(
TensorBox.create, ir.FallbackKernel.create(kernel, *args, **kwargs)
)
return handler
@functools.lru_cache(None)
def _warn_complex_not_supported():
warnings.warn(
"Torchinductor does not support code generation for complex operators. Performance may be worse than eager."
)
# There are some types (CPU) which we accept as input but not as
# output.
def unsupported_input_tensor(t: torch._subclasses.FakeTensor, parent=None):
"Do not support reading or writing to this tensor"
if t.is_complex():
# Complex views are supported with IR ComplexView
if parent and parent.target in (
torch.ops.aten.view.dtype,
torch.ops.prims.convert_element_type.default,
):
return False
_warn_complex_not_supported()
return True
return False
def unsupported_output_tensor(t: torch._subclasses.FakeTensor, parent=None):
"Do not support writing tensor but can read from it"
if unsupported_input_tensor(t, parent):
return True
return t.is_cpu and config.disable_cpp_codegen
def fallback_node_due_to_unsupported_type(node: torch.fx.Node, allow_cpu_inputs=True):
# Custom fallback lowering
if node.target is aten.view_as_complex.default:
return False
# We should be able to remove this special case once `disable_cpp_codegen` is killed.
if node.target is aten.lift_fresh_copy.default:
return False
def check_skip_condition(node, parent, is_output):
if not isinstance(node, torch.fx.Node):
return False
if "val" not in node.meta:
return False
for meta in pytree.tree_leaves(node.meta["val"]):
if not isinstance(meta, torch._subclasses.FakeTensor):
continue
if is_output:
if unsupported_output_tensor(meta, parent):
return True
else:
if unsupported_input_tensor(meta, parent):
return True
return False
# only skip codegen if there is a cpu output, not input
for arg in pytree.arg_tree_leaves(*node.args, **node.kwargs):
if check_skip_condition(arg, node, is_output=False):
return True
return check_skip_condition(node, node, is_output=True)
def make_fallback(op, layout_constraint=None, warn=True):
assert op not in decompositions, f"both a fallback and a decomp for same op: {op}"
if (
warn
and bool(os.getenv("CI"))
and get_decompositions([op])
# if fallback_random, we allow not decomposing random
and not (
config.fallback_random
and op in torch._decomp.decompositions_for_rng.extra_random_decomps
)
):
# Note: 'warn' is holdover from when this was a warning, but for ops that previously
# set warn=False we do not want a CI error.
# Ignore the 'suppress errors' configs in CI, as this particular warning happens on startup anyway and is not
# likely to be triggered preferentially on one CI config over another.
if torch._dynamo.config.suppress_errors:
torch._dynamo.config.suppress_errors = False
log.warning(
"A make_fallback error occurred in suppress_errors config,"
" and suppress_errors is being disabled to surface it."
)
raise AssertionError(
f"make_fallback({op}): a decomposition exists, we should switch to it."
" To fix this error, either add a decomposition to core_aten_decompositions (preferred)"
" or inductor_decompositions, and delete the corresponding `make_fallback` line."
" Get help from the inductor team if unsure, don't pick arbitrarily to unblock yourself.",
)
def register_fallback(op_overload):
add_needs_realized_inputs(op_overload)
if layout_constraint is not None:
add_layout_constraint(op_overload, layout_constraint)
return register_lowering(op_overload, type_promotion_kind=None)(
fallback_handler(op_overload)
)
if isinstance(op, torch._ops.OpOverloadPacket):
for ol in op.overloads():
op_overload = getattr(op, ol)
register_fallback(op_overload)
elif isinstance(op, (torch._ops.OpOverload, torch._ops.HigherOrderOperator)):
register_fallback(op)
else:
raise RuntimeError(f"Unsupported fallback {op} with type {type(op)}")
def philox_rand_offset(shape):
"""
TorchInductor offset calculation differs from PyTorch eager offset
calculation for random ops (tl.rand vs torch.rand). In future, we should
strive for same impl for tl.rand and torch.rand.
"""
numel = 1
for s in shape:
numel = numel * s
return tensor(numel, dtype=torch.int64)
@register_lowering(torch.ops.rngprims.philox_rand, type_promotion_kind=None)
def philox_rand(size, seed, offset, stride, device, dtype):
# stride arg is optional and will be used in future for distributed random
# ops. Currently, its unused.
random_pos = ir.FixedLayout(
device,
dtype,
size,
ir.FlexibleLayout.contiguous_strides(size),
).make_indexer()
seed_loader = seed.make_loader()
offset_loader = offset.make_loader()
def inner_fn(index):
# Both seed and offset in the philox_rand op are tensors.
# torch seed and offsets are of type int64, but tl.rand accepts int32
seed_index_expr = ops.to_dtype(seed_loader([]), torch.int32)
offset_index_expr = ops.to_dtype(offset_loader([]), torch.int32)
# Get the offset'd position
rand_index_expr = ops.add(
ops.index_expr(random_pos(index), torch.int32), offset_index_expr
)
result = ops.rand(
seed_index_expr,
rand_index_expr,
)
return ops.to_dtype(result, dtype)
random_values_node = Pointwise.create(
device=device,
dtype=dtype,
inner_fn=inner_fn,
ranges=list(size),
)
offset_node = philox_rand_offset(size)
return random_values_node, offset_node
@register_lowering(aten.native_dropout, type_promotion_kind=None)
def native_dropout(x, p, train):
if config.fallback_random:
return pytree.tree_map(
TensorBox.create,
ir.FallbackKernel.create(aten.native_dropout.default, x, p, train),
)
else:
raise AssertionError("should be handled in replace_random.py")
@register_lowering(aten.bernoulli_, type_promotion_kind=None)
def bernoulli_(x, *args):
assert config.fallback_random or x.get_device() == torch.device(
"cpu"
), "this should be handled in decomps unless config.fallback_random or the device is CPU"
x.realize()
ir.InplaceBernoulliFallback(x, *args)
return x
@register_lowering(aten.bernoulli.p, type_promotion_kind=None)
def bernoulli_p(x, *args):
assert config.fallback_random or x.get_device() == torch.device(
"cpu"
), "this should be handled in decomps unless config.fallback_random or the device is CPU"
return bernoulli_(clone(x), *args)
# This shouldn't be called in general
@register_lowering(aten._foobar)
def _foobar(_):
raise AssertionError()
@functools.lru_cache(1)
def _warn_triton_random(salt):
log.info("using triton random, expect difference from eager")
def warn_triton_random():
# only warn once per graph
_warn_triton_random(V.graph.creation_time)
fallback_rand_default = fallback_handler(aten.rand.default)
fallback_rand_generator = fallback_handler(aten.rand.generator)
fallback_randn_default = fallback_handler(aten.randn.default)
fallback_randn_generator = fallback_handler(aten.randn.generator)
make_fallback(aten.randint)
@register_lowering(aten.rand)
def rand(*args, **kwargs):
if kwargs.get("generator", None) is not None:
return fallback_rand_generator(*args, **kwargs)
elif config.fallback_random:
kwargs.pop("generator", None)
return fallback_rand_default(*args, **kwargs)
raise AssertionError("should have been handled in replace_random.py")
@register_lowering(aten.randn)
def randn(*args, **kwargs):
if kwargs.get("generator", None) is not None:
return fallback_randn_generator(*args, **kwargs)
elif config.fallback_random:
kwargs.pop("generator", None)
return fallback_randn_default(*args, **kwargs)
raise AssertionError("should have been handled in replace_random.py")
@register_lowering(inductor_prims.force_stride_order, type_promotion_kind=None)
def inductor_force_stride_order(input_tensor, stride):
stride_order = ir.get_stride_order(stride)
return ir.ExternKernel.require_stride_order(input_tensor, stride_order)
@register_lowering(inductor_prims.seed, type_promotion_kind=None)
def inductor_seed(device: torch.device):
raise AssertionError("should be handled in fuse_seed_creation_pass()")
@register_lowering(inductor_prims.seeds, type_promotion_kind=None)
def inductor_seeds(count, device):
warn_triton_random()
return TensorBox.create(ir.RandomSeeds(count, decode_device(device)))
@register_lowering(inductor_prims.lookup_seed, type_promotion_kind=None)
def inductor_lookup_seed(seeds, index):
def inner_fn(_):
return ops.load_seed(seeds.get_name(), index)
return Pointwise.create(
device=seeds.get_device(),
dtype=seeds.get_dtype(),
inner_fn=inner_fn,
ranges=[],
)
@register_lowering(inductor_prims.random, type_promotion_kind=None)
def inductor_random(size: List[int], seed: TensorBox, mode: str, *, offset: int = 0):
assert not config.fallback_random
assert mode in ("rand", "randn")
size = [*size]
dtype = torch.float32
device = seed.get_device()
random_pos = ir.FixedLayout(
device, dtype, size, ir.FlexibleLayout.contiguous_strides(size), offset=offset
).make_indexer()
seed_loader = seed.make_loader()
def inner_fn(index):
return getattr(ops, mode)(
seed_loader([]),
ops.index_expr(random_pos(index), torch.int32),
)
result = Pointwise.create(
device=device,
dtype=dtype,
inner_fn=inner_fn,
ranges=[*size],
)
result.realize()
return result
@register_lowering(inductor_prims.randint, type_promotion_kind=None)
def inductor_randint(
low: int, high: int, size: List[int], seed: TensorBox, *, offset: int = 0
):
assert not config.fallback_random
size = [*size]
dtype = torch.int64
device = seed.get_device()
random_pos = ir.FixedLayout(
device, dtype, size, ir.FlexibleLayout.contiguous_strides(size), offset=offset
).make_indexer()
seed_loader = seed.make_loader()
def inner_fn(index):
return ops.randint64(
seed_loader([]),
ops.index_expr(random_pos(index), torch.int32),
low,
high,
)
return Pointwise.create(
device=device,
dtype=dtype,
inner_fn=inner_fn,
ranges=[*size],
)
@register_lowering(aten.bucketize, type_promotion_kind=None)
def bucketize(
input: TensorBox,
boundaries: TensorBox,
*,
out_int32: bool = False,
right: bool = False,
):
assert len(boundaries.get_size()) == 1
if not (is_triton(input) and is_triton(boundaries)):
return fallback_handler(aten.bucketize.Tensor, add_to_fallback_set=False)(
input, boundaries, out_int32=out_int32, right=right
)
# The entire boundaries tensor needs to be used by ops.bucketize, so we
# need to realize it into global memory; or in other words, we can't
# guarantee that boundaries.get_name() (used below) will exist unless
# we call boundaries.realize().
boundaries.realize()
boundaries_size = boundaries.get_size()[0]
boundaries_loader = boundaries.make_loader()
device = input.get_device()
input_loader = input.make_loader()
index_dtype = torch.int32 if out_int32 else torch.int64
def inner_fn(index):
val = input_loader(index)
indices = ops.bucketize(
val,
boundaries.get_name(),
boundaries_size,
index_dtype,
right,
)
return indices
return Pointwise.create(
device=device,
dtype=index_dtype,
inner_fn=inner_fn,
ranges=input.get_size(),
)
def require_dense(_, *args, **kwargs):
args, kwargs = pytree.tree_map_only(
ir.IRNode, ir.ExternKernel.require_stride1, (args, kwargs)
)
return args, kwargs
def require_contiguous(_, *args, **kwargs):
args, kwargs = pytree.tree_map_only(
ir.IRNode, ir.ExternKernel.require_contiguous, (args, kwargs)
)
return args, kwargs
def require_channels_last(_, *args, **kwargs):
args, kwargs = pytree.tree_map_only(
ir.IRNode, ir.ExternKernel.require_channels_last, (args, kwargs)
)
return args, kwargs
def constrain_to_fx_strides(fx_node, *args, **kwargs):
def apply_constraint(arg, fx_arg):
if isinstance(arg, ir.IRNode):
stride_order = ir.get_stride_order(fx_arg.meta["val"].stride())
return ir.ExternKernel.require_stride_order(arg, stride_order)
return arg
args = tuple(
apply_constraint(arg, fx_arg) for arg, fx_arg in zip(args, fx_node.args)
)
kwargs = {k: apply_constraint(v, fx_node.kwargs[k]) for k, v in kwargs.items()}
return args, kwargs
# TODO(jansel): we should implement decomps or lowerings for these
# https://github.com/pytorch/torchdynamo/issues/327
FALLBACK_ALLOW_LIST = {
"torchvision::roi_align",
}
def sdpa_constraint(fx_node, *args, **kwargs):
# sdpa requires dense last dimension]
def apply_constraint(arg, fx_arg):
if not isinstance(arg, ir.IRNode):
return arg
meta_val = fx_arg.meta["val"]
if not meta_val.is_cuda:
return arg
stride_order = ir.get_stride_order(meta_val.stride())
if stride_order and stride_order[-1] != 0:
# contiguous stride order
stride_order = list(reversed(range(len(arg.get_size()))))
# This is the minimum alignment required by SDPA kernels for attention_bias.
# This value can be found in pytorch/aten/src/ATen/native/transformers/attention.cpp preprocess_mask
ALIGNMENT = 8
assert isinstance(arg, TensorBox)
if len(arg.get_size()) not in (3, 4):
return arg
def is_aligned_realized_tensor(x):
aligned_strides = all(
(V.graph.sizevars.size_hint(x.get_stride()[i]) % ALIGNMENT) == 0
for i in range(len(x.get_stride()) - 1)
)
return (
V.graph.sizevars.size_hint(x.get_stride()[-1])
) == 1 and aligned_strides
try:
arg.get_stride()
if is_aligned_realized_tensor(arg):
return arg
except AttributeError:
pass
def is_aligned(x):
return (V.graph.sizevars.size_hint(x.get_size()[-1]) % ALIGNMENT) == 0
if isinstance(arg.data, ir.BaseView):
if not is_aligned(arg):
if is_aligned(arg.unwrap_view()):
return arg
return ir.ExternKernel.require_stride_order(arg, stride_order)
args = tuple(
apply_constraint(arg, fx_arg) for arg, fx_arg in zip(args, fx_node.args)
)
kwargs = {k: apply_constraint(v, fx_node.kwargs[k]) for k, v in kwargs.items()}
return args, kwargs
# WIP
make_fallback(aten.index_reduce) # @pearu
make_fallback(aten._adaptive_avg_pool3d) # @isuruf
make_fallback(aten.adaptive_max_pool3d) # @isuruf
make_fallback(aten.avg_pool3d) # @isuruf
make_fallback(aten.fractional_max_pool3d) # @isuruf
make_fallback(aten.max_pool3d_with_indices) # @isuruf (can this one be implemented?)
make_fallback(aten.cummax) # @isuruf
make_fallback(aten.cummin) # @isuruf
# 1) Easy
make_fallback(aten.uniform, warn=False)
make_fallback(aten.exponential.default, warn=False) # (fails accuracy on test_torch.py)
make_fallback(aten._pdist_forward) # Has decomp. Needs benchmarks
make_fallback(aten.soft_margin_loss_backward, warn=False) # py_impl?
make_fallback(aten.searchsorted) # bucketized is implemented (see eager impl)
# 1.5) Easy or Impossible
make_fallback(aten._cdist_forward) # p=2 should be feasible
make_fallback(aten._cdist_backward)
# See resize_storage_bytes
make_fallback(aten.resize)
make_fallback(aten.resize_)
make_fallback(aten.resize_as)
make_fallback(aten.resize_as_)
# 2) Medium
make_fallback(aten.max_unpool2d)
make_fallback(aten.max_unpool3d)
make_fallback(aten._trilinear)
# 3) Difficult
# Scans
# See the discussion at
# https://dev-discuss.pytorch.org/t/pytorch-sparse-gnn-compiler-rfc/1644/19
make_fallback(aten.segment_reduce.default)
make_fallback(aten._segment_reduce_backward.default)
# Histogram (need to implement Histogram IR)
make_fallback(aten.histc)
make_fallback(aten.histogram.bin_ct)
make_fallback(aten._histogramdd_bin_edges.default)
make_fallback(aten._histogramdd_from_bin_cts.default)
# Need templated kernel
make_fallback(aten.addbmm)
make_fallback(aten.addmv, warn=False)
make_fallback(aten._addmm_activation, warn=False)
# Need templated kernel. Probably impossible to write efficiently
make_fallback(aten.convolution_backward, constrain_to_fx_strides)
make_fallback(aten._cudnn_rnn, require_dense)
make_fallback(aten._cudnn_rnn_backward, require_contiguous)
# Haven't checked but sound difficult / impossible
make_fallback(aten._embedding_bag, require_contiguous)
make_fallback(aten._embedding_bag_forward_only, require_contiguous)
make_fallback(aten._embedding_bag_dense_backward)
make_fallback(aten._embedding_bag_per_sample_weights_backward)
make_fallback(aten._embedding_bag_per_sample_weights_backward)
make_fallback(aten._fused_moving_avg_obs_fq_helper)
make_fallback(aten._fused_moving_avg_obs_fq_helper_functional)
# 4) Backwards (try py_impl'ing them) when fwd is written as a decomp
make_fallback(aten.avg_pool3d_backward)
make_fallback(aten.max_pool3d_with_indices_backward)
make_fallback(aten._adaptive_avg_pool2d_backward, require_dense)
make_fallback(aten._adaptive_avg_pool3d_backward)
make_fallback(aten.adaptive_max_pool2d_backward)
make_fallback(aten.adaptive_max_pool3d_backward)
make_fallback(aten.fractional_max_pool2d_backward)
make_fallback(aten.fractional_max_pool3d_backward)
make_fallback(aten.replication_pad1d_backward)
make_fallback(aten.replication_pad2d_backward)
make_fallback(aten.upsample_linear1d_backward)
make_fallback(aten.upsample_bicubic2d_backward, require_contiguous)
make_fallback(aten.upsample_trilinear3d_backward)
make_fallback(aten.grid_sampler_2d_backward, require_dense)
make_fallback(aten._pdist_backward)
# 5) Impossible (missing triton/CPU features)
# Sorting / Sorting-like
make_fallback(aten.sort)
make_fallback(aten.sort.stable)
make_fallback(aten.kthvalue)
make_fallback(aten.topk)
make_fallback(aten.mode)
make_fallback(aten.median)
make_fallback(aten.nanmedian)
make_fallback(aten.randperm)
# Linalg
make_fallback(aten._linalg_det)
make_fallback(aten.linalg_householder_product)
make_fallback(aten.linalg_inv_ex)
make_fallback(aten.linalg_ldl_factor_ex)
make_fallback(aten.linalg_ldl_solve)
make_fallback(aten.linalg_lu)
make_fallback(aten.linalg_lu_factor_ex)
make_fallback(aten.linalg_lu_solve)
make_fallback(aten.linalg_matrix_exp)
make_fallback(aten.linalg_qr)
make_fallback(aten._linalg_slogdet)
make_fallback(aten._linalg_solve_ex)
make_fallback(aten.linalg_solve_triangular)
make_fallback(aten._linalg_svd)
make_fallback(aten.lu_unpack)
make_fallback(aten.ormqr)
make_fallback(aten._linalg_check_errors)
make_fallback(aten.linalg_pinv.atol_rtol_tensor)
make_fallback(aten._linalg_eigh)
make_fallback(aten.triangular_solve)
make_fallback(aten.linalg_cholesky_ex)
make_fallback(aten.cholesky_inverse)
make_fallback(aten.cholesky_solve)
make_fallback(aten.geqrf)
make_fallback(aten._fft_r2c) # needs complex as well
# Data dependent (are these necessary?)
make_fallback(aten.nonzero.default)
# Misc
make_fallback(aten.gcd.default, warn=False)
make_fallback(aten._thnn_fused_lstm_cell, require_dense)
make_fallback(torch._prims.rng_prims.run_and_save_rng_state)
make_fallback(torch._prims.rng_prims.run_with_rng_state)
# Implmented / Half implemented
# Scans. Implemented for CUDA, missing CPU
make_fallback(aten.masked_scatter)
make_fallback(aten.masked_scatter_backward)
# Complex number support
make_fallback(aten.view_as_complex, require_contiguous)
make_fallback(aten.angle) # needs complex
# Needs efficentzerotensor
make_fallback(aten._efficientzerotensor)
# Needs Sparse
make_fallback(aten._sparse_coo_tensor_with_dims_and_tensors)
make_fallback(aten.to_sparse)
make_fallback(aten._to_sparse)
# Needs dimname support
make_fallback(aten.zeros.names)
# 6) Pattern-matched
make_fallback(
aten._scaled_dot_product_efficient_attention.default,
sdpa_constraint,
warn=False,
)
make_fallback(
aten._scaled_dot_product_efficient_attention_backward.default,
sdpa_constraint,
warn=False,
)
make_fallback(
aten._scaled_dot_product_flash_attention.default,
sdpa_constraint,
warn=False,
)
make_fallback(
aten._scaled_dot_product_flash_attention_backward.default,
sdpa_constraint,
warn=False,
)
make_fallback(
aten._scaled_dot_product_flash_attention_for_cpu.default,
sdpa_constraint,
warn=False,
)
make_fallback(
aten._scaled_dot_product_flash_attention_for_cpu_backward.default,
sdpa_constraint,
warn=False,
)
make_fallback(aten._flash_attention_forward.default, sdpa_constraint)
make_fallback(aten._flash_attention_backward.default, sdpa_constraint)
make_fallback(aten._efficient_attention_forward.default, sdpa_constraint)
make_fallback(aten._efficient_attention_backward.default, sdpa_constraint)
make_fallback(aten._scaled_mm.default, constrain_to_fx_strides)
# Register with type_promotion_kind None.
# For example, fp16.copy_(fp32) should **not** promote the first input's dtype.
@register_lowering(aten.copy, type_promotion_kind=None)
def copy(self, src, non_blocking=False):
x = src
if self.get_device() != src.get_device():
x = to_device(x, self.get_device())
if self.get_dtype() != src.get_dtype():
x = to_dtype(x, self.get_dtype())
if self.get_size() != src.get_size():
out = expand(x, self.get_size())
return clone(out)
return clone(x)
@register_lowering(aten.clone)
def clone(x, *, memory_format=None):
# TODO(jansel): memory format
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=x.make_loader(),
ranges=list(x.get_size()),
)
def clone_preserve_reinterpret_view(x):
reinterpret_view_layouts = []
if isinstance(x, TensorBox) and isinstance(x.data, ir.ReinterpretView):
x = x.data # unwrap TensorBox
while isinstance(x, ir.ReinterpretView):
reinterpret_view_layouts.append(x.get_layout())
x = x.data
x = TensorBox(x)
x = clone(x)
if reinterpret_view_layouts:
x = x.data # unwrap TensorBox
for layout in reinterpret_view_layouts[::-1]:
x = ir.ReinterpretView(x, layout)
x = TensorBox(x)
return x
if hasattr(aten, "lift_fresh_copy"):
register_lowering(aten.lift_fresh_copy)(clone)
@register_lowering(prims.iota)
def iota(
length,
*,
start,
step,
dtype,
device,
requires_grad,
):
def fn(index):
return ops.index_expr(step * index[0] + start, dtype=dtype)
return Pointwise.create(
device=decode_device(device),
dtype=dtype,
inner_fn=fn,
ranges=[length],
)
@register_lowering(aten.select_scatter, type_promotion_kind=None)
def select_scatter(x, src, dim: int, index: int):
assert x.get_dtype() == src.get_dtype()
x_loader = x.make_loader()
dim = _validate_dim(x, dim, 0)
if V.graph.sizevars.evaluate_expr(sympy.Lt(index, 0)):
index = index + x.get_size()[dim]
V.graph.sizevars.guard_leq(0, index) # type: ignore[arg-type]
V.graph.sizevars.guard_lt(index, x.get_size()[dim]) # type: ignore[arg-type]
src = expand(unsqueeze(src, dim), x.get_size())
src_loader = src.make_loader()
def inner_fn(idx):
return ops.where(
ops.eq(
ops.index_expr(idx[dim], torch.int32),
ops.index_expr(index, torch.int32),
),
src_loader(idx),
x_loader(idx),
)
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=inner_fn,
ranges=list(x.get_size()),
)
@register_lowering(aten.slice_scatter, type_promotion_kind=None)
def slice_scatter(x, src, dim=0, start=None, end=None, step=1):
assert x.get_dtype() == src.get_dtype()
x_loader = x.make_loader()
dim = _validate_dim(x, dim, 0)
dim_size = x.get_size()[dim]
start, end = ir.SliceView.normalize_start_end(x, dim, start, end)
src_size = list(x.get_size())
src_size[dim] = FloorDiv(end - start + (step - 1), step)
src = expand(src, src_size)
src_loader = src.make_loader()
def inner_fn(idx):
if start == 0 and end == dim_size and step == 1:
# selecting every element is the same as just src.clone()
return src_loader(idx)
idx_dim = ops.index_expr(idx[dim], torch.int64)
src_idx = list(idx)
src_idx[dim] = FloorDiv(idx[dim] - start, step)
mask = []
if start != 0:
mask.append(
ops.ge(
idx_dim,
ops.index_expr(sympy.expand(start), torch.int64),
)
)
if end != dim_size:
mask.append(
ops.lt(
idx_dim,
ops.index_expr(sympy.expand(end), torch.int64),
)
)
if step != 1:
mask.append(
ops.eq(
ops.index_expr(
ModularIndexing(idx[dim] - start, 1, step), torch.int64
),
ops.constant(0, torch.torch.int64),
)
)
assert mask
mask = functools.reduce(ops.and_, mask)
src_val = ops.masked(
mask,
lambda: src_loader(src_idx),
0 if is_integer_type(x) else 0.0,
)
return ops.where(
mask,
src_val,
x_loader(idx),
)
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=inner_fn,
ranges=list(x.get_size()),
)
def _unwrap(x):
if isinstance(x, (list, tuple)) and len(x) > 0:
return _unwrap(x[0])
return x
@register_lowering([torch.tensor, aten.scalar_tensor])
def tensor(data, *, dtype=None, device=None, layout=None, pin_memory=False):
assert_nyi(layout in (None, torch.strided), f"layout={layout}")
assert_nyi(not pin_memory, "pin_memory")
if isinstance(_unwrap(data), int):
dtype = dtype or torch.int64
else:
dtype = dtype or torch.get_default_dtype()
ranges: List[sympy.Expr] = []
if isinstance(data, sympy.Expr):
def inner_fn(index):
return ops.index_expr(data, dtype)
elif isinstance(data, (float, int)):
def inner_fn(index):
return ops.constant(data, dtype)
elif len(data) == 0 or isinstance(data[0], (float, int)) and len(data) <= 8:
# inline small tensors
ranges.append(sympy.Integer(len(data)))
def inner_fn(index):
def binary_search(start, end):
assert start < end
if end - start == 1:
return ops.constant(data[start], dtype)
mid = (end - start) // 2 + start
return ops.where(
ops.lt(
ops.index_expr(index[0], torch.int64),
ops.constant(mid, torch.int64),
),
binary_search(start, mid),
binary_search(mid, end),
)
if len(data) == 0:
return ops.constant(0, dtype)
return binary_search(0, len(data))
else:
return V.graph.add_tensor_constant(
torch.tensor(data, dtype=dtype, device=device)
)
return Pointwise.create(
device=decode_device(device),
dtype=dtype,
inner_fn=inner_fn,
ranges=ranges,
)
@register_lowering(torch.as_tensor)
def as_tensor(data, dtype=None, device=None):
if isinstance(data, TensorBox):
if dtype is not None:
data = to_dtype(data, dtype)
if device is not None:
data = to_device(data, device)
return data
return tensor(data, dtype=dtype, device=device)
@register_lowering(torch.LongTensor)
def long_tensor(data):
return tensor(data, dtype=torch.int64)
@register_lowering(aten._local_scalar_dense)
def _local_scalar_dense(data):
# This is interesting! Most lowerings return tensors, so you can just
# return the buffer you allocated and it will get used (or not used, if
# it's dead.) But _local_scalar_dense (aka item) returns an int,
# not a Tensor, so you would have a type mismatch if you return a buffer;
# we are obligated to return a sympy expression instead. However,
# we need to actually codegen the .item() call somehow. We do this
# by registering a faux buffer for the DynamicScalar IR node, which is
# solely responsible for generating this .item(). The buffer is
# not used for anything (notice we discard it); at codegen time,
# the "buffer" just gets assigned None.
sym = V.graph.current_node.meta["val"].node.expr
buffer = ir.DynamicScalar(sym, data)
buffer.name = V.graph.register_buffer(buffer)
return sym
@register_lowering(aten._assert_scalar)
def _assert_scalar(data, msg):
buffer = ir.AssertScalar(data, msg)
# This buffer isn't used by anyone (it returns None), so we must explicitly register it
buffer.name = V.graph.register_buffer(buffer)
return buffer
def _full(fill_value, device, dtype, size):
value = fill_value
if not isinstance(fill_value, (int, float)) and hasattr(value, "value"):
value = value.value
if isinstance(value, (int, float)):
def inner_fn(index):
return ops.constant(value, dtype)
elif isinstance(value, sympy.Expr):
def inner_fn(index):
return ops.index_expr(value, dtype)
else:
assert len(value.get_size()) == 0
value_loader = value.make_loader()
def inner_fn(index):
return value_loader([])
return Pointwise.create(
device=device,
dtype=dtype,
inner_fn=inner_fn,
ranges=list(size),
)
@register_lowering(aten.full_like, type_promotion_kind=None)
def full_like(x, fill_value, **kwargs):
return create_tensor_like(tensor_constructor(fill_value))(x, **kwargs)
def tensor_constructor(fill_value):
# torch.zeros, torch.ones, etc
def inner(
*size,
names=None,
dtype=None,
device=None,
layout=None,
pin_memory=False,
memory_format=None,
):
assert_nyi(names is None, "named tensors")
assert_nyi(layout in (None, torch.strided), f"layout={layout}")
assert_nyi(not pin_memory, "pin_memory")
device = decode_device(device)
dtype = dtype or torch.get_default_dtype()
if len(size) == 1 and isinstance(size[0], (list, tuple, torch.Size)):
size = tuple(size[0])
# See https://github.com/pytorch/pytorch/issues/118102
# All sizes at lowering time should be sympy.Symbol, not SymInt!
for s in size:
assert not isinstance(s, torch.SymInt)
size = [sympy.expand(s) for s in size]
return _full(fill_value, device, dtype, size)
return inner
@register_lowering([torch.empty, aten.empty])
def empty(
*size,
names=None,
dtype=None,
layout=None,
device=None,
pin_memory=None,
memory_format=None,
):
assert_nyi(names is None, "named tensors")
device = decode_device(device)
if len(size) == 1 and isinstance(size[0], (list, tuple, torch.Size)):
size = tuple(size[0])
return empty_strided(
size, None, dtype=dtype, layout=layout, device=device, pin_memory=pin_memory
)
def create_tensor_like(creation_fn):
"""
Shim to convert X_like(...) into X(...). For example zeros_like() into zeros().
"""
def _constant_like(
x, *, dtype=None, device=None, layout=None, pin_memory=False, memory_format=None
):
assert_nyi(not pin_memory, "pin_memory")
assert_nyi(layout in (None, torch.strided), f"layout={layout}")
if dtype is None:
dtype = x.get_dtype()
else:
dtype = decode_dtype(dtype)
device = device or x.get_device()
size = list(x.get_size())
return creation_fn(
size, dtype=dtype, device=device, layout=layout, pin_memory=pin_memory
)
return _constant_like
def constant_like(fill_value):
return create_tensor_like(tensor_constructor(fill_value))
empty_like = register_lowering(aten.empty_like)(create_tensor_like(empty))
ones_like = create_tensor_like(tensor_constructor(1))
zeros_like = create_tensor_like(tensor_constructor(0))
def new_constant(fill_value):
def _new_constant(
x, size, *, dtype=None, layout=None, device=None, pin_memory=None
):
assert isinstance(size, (list, tuple))
assert_nyi(not pin_memory, "pin_memory")
assert_nyi(layout in (None, torch.strided), f"layout={layout}")
dtype = decode_dtype(dtype) or x.get_dtype()
device = device or x.get_device()
size = [sympy.Integer(s) for s in size]
return _full(fill_value, device, dtype, size)
return _new_constant
@register_lowering(aten.new_empty)
def new_empty(x, size, *, dtype=None, layout=None, device=None, pin_memory=None):
if dtype is None:
dtype = x.get_dtype()
if device is None:
device = x.get_device()
return empty_strided(
size, None, dtype=dtype, layout=layout, device=device, pin_memory=pin_memory
)
@register_lowering(aten.empty_strided)
def empty_strided(
size, stride, *, dtype=None, layout=None, device=None, pin_memory=None
):
assert isinstance(size, (list, tuple))
assert isinstance(stride, (list, tuple, type(None)))
assert_nyi(not pin_memory, "pin_memory")
assert_nyi(layout in (None, torch.strided), f"layout={layout}")
dtype = decode_dtype(dtype) or torch.get_default_dtype()
device = device or torch.tensor(0.0).device
pointwise = _full(fill_value=0, device=device, dtype=dtype, size=size)
pointwise.realize()
buffer = pointwise.data.data
# explicitly set ranges to zeros in order to make a NopKernelSchedulerNode
buffer.data.ranges = [0] * len(size)
assert isinstance(buffer, ir.ComputedBuffer)
size = [sympy.expand(s) for s in size]
stride = (
[sympy.expand(s) for s in stride]
if stride
else ir.FlexibleLayout.contiguous_strides(size)
)
buffer.layout = ir.FixedLayout(
device=device,
dtype=dtype,
size=size,
stride=stride,
)
return pointwise
@register_lowering(aten.new_empty_strided)
def new_empty_strided(
x, size, stride, *, dtype=None, layout=None, device=None, pin_memory=None
):
if dtype is None:
dtype = x.get_dtype()
if device is None:
device = x.get_device()
return empty_strided(
size, stride, dtype=dtype, layout=layout, device=device, pin_memory=pin_memory
)
@register_lowering(prims.copy_strided.default)
def copy_strided(x, stride):
stride = [V.graph.sizevars.size_hint(s) for s in stride]
stride_order = sorted(range(len(stride)), key=stride.__getitem__)
return ir.ExternKernel.require_stride_order(x, stride_order)
@register_lowering([torch.full, aten.full])
def full(size, fill_value, **kwargs):
assert kwargs.get("dtype") is not None, "dtype should be handled by decomposition"
return tensor_constructor(fill_value)(size, **kwargs)
@register_lowering(aten.gather, type_promotion_kind=None)
def gather(x, dim, index, sparse_grad=False):
# sparse_grad doesn't affect forward computation,
# and backward tracing is taken care of by AOT Autograd
assert isinstance(x, TensorBox)
assert index.get_dtype() == torch.int64
size = x.get_size()
offset = len(size) == 0
dim = _validate_dim(x, dim, offset)
x_loader = x.make_loader()
index_loader = index.make_loader()
def fn(idx):
idx = list(idx)
if len(idx) != 0:
idx[dim] = ops.indirect_indexing(index_loader(idx), size[dim])
return x_loader(idx)
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=fn,
ranges=index.get_size(),
)
@register_lowering(aten.embedding, type_promotion_kind=None)
def embedding(weight, indices, padding_idx=-1, scale_grad_by_freq=False, sparse=False):
assert not sparse
assert isinstance(weight, TensorBox)
assert isinstance(indices, TensorBox)
assert "int" in str(indices.get_dtype())
weight_loader = weight.make_loader()
indices_loader = indices.make_loader()
indices_ndim = len(indices.get_size())
weight_size = weight.get_size()
new_size = [*indices.get_size(), *weight_size[1:]]
def fn(idx):
assert len(idx) == len(new_size), f"{idx} != {new_size}"
var_index = indices_loader(idx[:indices_ndim])
weight_idx = [ops.indirect_indexing(var_index, weight_size[0])] + [
*idx[indices_ndim:]
]
return weight_loader(weight_idx)
return Pointwise.create(
device=weight.get_device(),
dtype=weight.get_dtype(),
inner_fn=fn,
ranges=new_size,
)
def check_and_broadcast_indices(indices, device):
assert all(
i.get_dtype() in (torch.int64, torch.int32, torch.bool, torch.uint8)
for i in indices
if i is not None
), f"indices must be int64, byte or bool. Got {[i.get_dtype() for i in indices if i is not None]}"
if any(
i.get_dtype() in (torch.bool, torch.uint8) for i in indices if i is not None
):
raise NotImplementedError("Fallback for bool indices")
valid_idxs = [i for i, x in enumerate(indices) if isinstance(x, TensorBox)]
assert len(valid_idxs) > 0, "requires at least 1 non-None index"
new_indices = [None] * len(indices)
for i, x in zip(valid_idxs, broadcast_tensors(*[indices[i] for i in valid_idxs])):
# Eager allows indices to be CPU tensor when running on CUDA
# FIXME: Calling to_device(x, device) should work but
# test_advancedindex_mixed_cpu_devices still fails
if x.get_device() != device:
raise NotImplementedError("Fallback when indices is on a different device")
new_indices[i] = x
return new_indices, valid_idxs
def index_output_size_and_inner_fn(
x_size,
indices,
tensor_indices,
tensor_size,
indices_loaders,
indexed_size,
x_loader,
check,
):
# Note that behavior of indexing differs when there are non consecutive
# tensors. In this case, the tensor index is pulled to the beginning.
#
# Suppose a = torch.arange(3 * 4 * 5 * 6 * 7).view(3, 4, 5, 6, 7)
# x = torch.tensor[1,2]
# Then, a[:,x,:,x,:] will have shape 2,3,5,7 as due to x,:,x then 2 will
# be pulled to the front.
non_consecutive_tensors = False
for previous, current in zip(tensor_indices, tensor_indices[1:]):
if current - previous != 1:
non_consecutive_tensors = True
output_size = [x_size[i] for i, val in enumerate(indices) if val is None]
output_size = [*output_size, *x_size[len(output_size) + len(tensor_indices) :]]
first_tensor_index = tensor_indices[0]
if non_consecutive_tensors:
output_size = tensor_size + output_size
else:
output_size = (
output_size[:first_tensor_index]
+ tensor_size
+ output_size[first_tensor_index:]
)
def fn(idx):
assert len(idx) == len(output_size)
assert len(indices_loaders) == len(indexed_size)
rank = len(tensor_size)
new_index = []
first_tensor_index = tensor_indices[0]
start_offset = 0 if non_consecutive_tensors else first_tensor_index
next_idx = 0
for i in range(tensor_indices[-1] + 1):
if i == start_offset:
next_idx += rank
if indices[i] is None:
assert next_idx < len(idx)
new_index.append(idx[next_idx])
next_idx += 1
else:
loader = indices_loaders[i]
assert loader is not None
size = indexed_size[i]
new_index.append(
ops.indirect_indexing(
loader(idx[start_offset : start_offset + rank]),
size,
check=check,
)
)
new_index = [
*new_index,
*idx[next_idx:],
]
return new_index if x_loader is None else x_loader(new_index)
return output_size, fn
def index_impl(x, indices, check):
assert isinstance(indices, (list, tuple))
x_loader = x.make_loader()
indices, tensor_indices = check_and_broadcast_indices(indices, x.get_device())
assert len(tensor_indices) > 0, "Must have at least one valid idx"
indices_loaders = [i.make_loader() if i is not None else None for i in indices]
# no guards on output size, all the guards are set in broadcast_tensors
# We can use the first one since they are all required to be the same size
tensor_size = list(indices[tensor_indices[0]].get_size())
x_size = x.get_size()
indexed_size = [x_size[i] for i in range(len(indices)) if indices[i] is not None]
if 0 in indexed_size and 0 not in tensor_size:
raise IndexError("index is out of bounds for dimension with size 0")
indexed_size = [x_size[i] for i in range(len(indices))]
output_size, inner_fn = index_output_size_and_inner_fn(
x_size,
indices,
tensor_indices,
tensor_size,
indices_loaders,
indexed_size,
x_loader,
check=check,
)
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=inner_fn,
ranges=output_size,
)
@register_lowering(aten.index, type_promotion_kind=None)
def index(x, indices):
try:
return index_impl(x, indices, check=True)
except NotImplementedError:
# Fallback to ATen for boolean indexing
x.realize()
return fallback_handler(aten.index.Tensor, add_to_fallback_set=False)(
x, indices
)
@register_lowering(aten._unsafe_index, type_promotion_kind=None)
def _unsafe_index(x, indices):
return index_impl(x, indices, check=False)
# All the indexing decompositions are written in terms of index, index_put, and index_put_
# We cannot have this lowering as a decomposition as it introduces
# mutation in the graph, which is bad for Aot Autograd. Aot Autograd runs dead
# code elimination and common subexpression elimination optimizations, which
# assume graphs to be side-effect free. More details at
# https://github.com/pytorch/torchdynamo/issues/1235
# and
# https://github.com/pytorch/torchdynamo/issues/1863
@register_lowering(aten.index_put)
def index_put(x, indices, values, accumulate=False):
return index_put_(clone(x), indices, values, accumulate)
@register_lowering(aten._unsafe_index_put)
def _unsafe_index_put(x, indices, values, accumulate=False):
return index_put_impl_(clone(x), indices, values, accumulate, check=False)
def index_put_as_masked_fill(self, indices, value, accumulate):
if value.get_device() != self.get_device():
value = to_device(value, self.get_device())
if accumulate:
value = add(self, value)
return mutate_to(self, where(indices[0], value, self))
def index_put_fallback(self, indices, values, accumulate):
deterministic = torch.are_deterministic_algorithms_enabled()
if is_triton(values) and (accumulate or deterministic):
msg = (
"index put with accumulate."
if not deterministic
else "deterministic index put."
)
if stack_trace := V.graph.current_node.meta.get("stack_trace", None):
msg = f"{msg} Found from : \n {stack_trace}"
V.graph.disable_cudagraphs_reason = msg
ir.IndexPutFallback(V.graph.current_node.target, self, indices, values, accumulate)
return self
@register_lowering(aten.index_put_, type_promotion_kind=None)
def index_put_(self, indices, values, accumulate=False):
return index_put_impl_(self, indices, values, accumulate, check=True)
@register_lowering(inductor_prims._unsafe_index_put_, type_promotion_kind=None)
def _unsafe_index_put_(self, indices, values, accumulate=False):
return index_put_impl_(self, indices, values, accumulate, check=False)
def needs_fallback_due_to_atomic_add_limitations(dtype):
# tl.atomic_add does NOT support the following types
return dtype in {torch.int64, torch.bool, torch.bfloat16}
def index_put_impl_(self, indices, values, accumulate, check):
# Dispatch to masked fill for single boolean index with single value
if (
values.get_numel() == 1
and len(indices) == 1
and indices[0].get_dtype() in {torch.bool, torch.uint8}
):
mask = indices[0]
for _ in range(len(mask.get_size()), len(self.get_size())):
mask = unsqueeze(mask, -1)
return index_put_as_masked_fill(self, [mask], values, accumulate)
# Fallback in torch deterministic mode
if torch.are_deterministic_algorithms_enabled():
return index_put_fallback(self, indices, values, accumulate)
# Fallback if there is a boolean index
for index in indices:
if index is not None and index.get_dtype() in {torch.bool, torch.uint8}:
return index_put_fallback(self, indices, values, accumulate)
x_size = self.get_size()
x_ndim = len(x_size)
if accumulate and needs_fallback_due_to_atomic_add_limitations(self.get_dtype()):
# self is an scalar Tensor
if x_ndim == 0:
self = view(self, [1])
self = index_put_fallback(self, indices, values, accumulate)
if x_ndim == 0:
self = view(self, [])
return self
values = to_dtype(values, self.get_dtype())
try:
# Note that code will only get here when dtype is uint32
indices, tensor_indices = check_and_broadcast_indices(
indices, self.get_device()
)
except NotImplementedError:
return index_put_fallback(self, indices, values, accumulate)
indices_loaders = [i.make_loader() if i is not None else None for i in indices]
assert isinstance(self, TensorBox)
self.realize()
# self is an scalar Tensor
if x_ndim == 0:
self = view(self, [1])
# We can use the first one since they are all required to be the same size
tensor_size = list(indices[tensor_indices[0]].get_size())
indexed_size = [x_size[i] for i in range(len(indices))]
expected_vals_size, inner_fn = index_output_size_and_inner_fn(
x_size,
indices,
tensor_indices,
tensor_size,
indices_loaders,
indexed_size,
None,
check=check,
)
values = expand(values, expected_vals_size)
# all guards are set above during broadcast_tensors and expand
scatter = ir.Scatter(
device=self.get_device(),
dtype=self.get_dtype(),
inner_fn=values.make_loader(),
ranges=expected_vals_size, # iter_ranges,
output_indexer=inner_fn,
scatter_mode="atomic_add" if accumulate else None,
)
buffer = ir.ComputedBuffer(
None,
ir.MutationLayout(self),
scatter,
)
buffer.name = V.graph.register_buffer(buffer)
if x_ndim == 0:
self = view(self, [])
return self
@register_lowering(
inductor_prims.masked_scatter_with_index, type_promotion_kind=None, broadcast=False
)
def masked_scatter_with_index(self, mask, source_idx, source):
self_flat, mask_flat, source_flat = (view(x, (-1,)) for x in (self, mask, source))
assert self.get_size() == mask.get_size()
assert mask.get_dtype() in {torch.bool, torch.uint8}
self_loader = self_flat.make_loader()
mask_loader = mask_flat.make_loader()
source_idx_loader = source_idx.make_loader()
source_loader = source_flat.make_loader()
source_numel = source.get_numel()
def inner_fn(idx):
self_val = self_loader(idx)
mask_val = ops.to_dtype(mask_loader(idx), torch.bool)
def load_source_val():
source_idx_val = source_idx_loader(idx)
i = ops.indirect_indexing(source_idx_val, source_numel)
return source_loader([i])
source_val = ops.masked(mask_val, load_source_val, 0)
return ops.where(mask_val, source_val, self_val)
result_flat = Pointwise.create(
device=self.get_device(),
dtype=self.get_dtype(),
inner_fn=inner_fn,
ranges=self_flat.get_size(),
)
return view(result_flat, self.get_size())
@register_lowering(aten.as_strided_scatter, type_promotion_kind=None)
def as_strided_scatter(self, src, size, stride, storage_offset=None):
output = clone(self)
output_view = as_strided(output, size, stride, storage_offset)
copy_(output_view, src)
return output
@register_lowering(aten.scatter, type_promotion_kind=None)
def scatter(x, dim: int, index, src, **kwargs):
return scatter_(clone(x), dim, index, src, **kwargs)
def scatter_fallback(
fn,
self,
dim: int,
index,
src,
*,
reduce: Optional[str] = None,
include_self: bool = True,
):
reduce_ty = "add" if fn == "aten.scatter_" else "sum"
if (
reduce not in {None, reduce_ty}
or (
isinstance(src, TensorBox)
and src.get_device().type == torch.device("cuda").type
and needs_fallback_due_to_atomic_add_limitations(src.get_dtype())
)
or (
fn == "aten.scatter_reduce_"
and reduce == "sum"
and isinstance(src, TensorBox)
and src.get_device() == torch.device("cpu")
and config.cpp.fallback_scatter_reduce_sum
and (config.cpp.dynamic_threads or parallel_num_threads() != 1)
)
or (reduce == reduce_ty and self.get_dtype() in {torch.bool, torch.int64})
or torch.are_deterministic_algorithms_enabled()
):
ir.ScatterFallback(
V.graph.current_node.target,
fn,
self,
dim,
index,
src,
reduce=reduce,
include_self=include_self,
)
return self
return None
@register_lowering(aten.scatter_, type_promotion_kind=None)
def scatter_(self, dim: int, index, src, *, reduce: Optional[str] = None):
assert reduce in {None, "add", "multiply"}
fallback_result = scatter_fallback(
"aten.scatter_", self, dim, index, src, reduce=reduce
)
if fallback_result:
return fallback_result
if reduce == "add":
reduce = "sum"
elif reduce == "multiply":
reduce = "prod"
return scatter_reduce_(self, dim, index, src, reduce)
@register_lowering(aten.scatter_add, type_promotion_kind=None)
def scatter_add(x, dim: int, index, src):
return scatter_add_(clone(x), dim, index, src)
@register_lowering(aten.scatter_add_, type_promotion_kind=None)
def scatter_add_(x, dim: int, index, src):
return scatter_reduce_(x, dim, index, src, "sum")
@register_lowering(aten.scatter_reduce, type_promotion_kind=None)
def scatter_reduce(x, dim: int, index, src, reduction_type, **kwargs):
return scatter_reduce_(clone(x), dim, index, src, reduction_type, **kwargs)
@register_lowering(aten.scatter_reduce_, type_promotion_kind=None)
def scatter_reduce_(self, dim: int, index, src, reduce, *, include_self: bool = True):
assert reduce in {None, "sum", "prod", "mean", "amax", "amin"}
fallback_result = scatter_fallback(
"aten.scatter_reduce_",
self,
dim,
index,
src,
reduce=reduce,
include_self=include_self,
)
if fallback_result:
return fallback_result
assert isinstance(self, TensorBox)
assert "int" in str(index.get_dtype())
ndim = len(self.get_size())
if ndim == 0:
self = view(self, [1])
if isinstance(src, TensorBox) and len(src.get_size()) == 0:
src = view(src, [1])
if isinstance(index, TensorBox) and len(index.get_size()) == 0:
index = view(index, [1])
dim = _validate_dim(self, dim)
self.realize()
index_loader = index.make_loader()
src_loader = src.make_loader() if isinstance(src, TensorBox) else None
def output_indexer(idx):
# self is captured from the end of the function, so it may have 0 dim
shape = self.get_size()
ndim = len(shape)
indirect_idx = list(idx)
indirect_idx[dim] = ops.indirect_indexing(
index_loader(idx), 1 if ndim == 0 else shape[dim]
)
return indirect_idx
def fn(idx):
if src_loader:
return src_loader(idx)
else:
# src is a scalar
return ops.constant(src, self.get_dtype())
def backend_reduce_str(reduce):
if reduce == "sum":
return "atomic_add"
else:
# TODO: Need to support more reduction type
assert reduce is None
return None
if not include_self:
# zero out the corresponding elements first
zero_out = ir.Scatter(
device=self.get_device(),
dtype=self.get_dtype(),
inner_fn=lambda index: ops.constant(0, self.get_dtype()),
ranges=index.get_size(),
output_indexer=output_indexer,
scatter_mode=None,
)
buffer = ir.ComputedBuffer(
None,
ir.MutationLayout(self),
zero_out,
)
buffer.name = V.graph.register_buffer(buffer)
# self[index[i][j][k]][j][k] += src[i][j][k] # if dim == 0
# self[i][index[i][j][k]][k] += src[i][j][k] # if dim == 1
# self[i][j][index[i][j][k]] += src[i][j][k] # if dim == 2
scatter = ir.Scatter(
device=self.get_device(),
dtype=self.get_dtype(),
inner_fn=fn,
ranges=index.get_size(),
output_indexer=output_indexer,
scatter_mode=backend_reduce_str(reduce),
)
buffer = ir.ComputedBuffer(
None,
ir.MutationLayout(self),
scatter,
)
buffer.name = V.graph.register_buffer(buffer)
if ndim == 0:
self = view(self, [])
return self
def upsample_nearestnd(
x,
output_size,
scales_x: Tuple[Optional[float], ...],
n: int = 2,
exact: bool = False,
):
x.realize_hint() # elements are reused
x_loader = x.make_loader()
i_sizes = x.get_size()[-n:]
batch = x.get_size()[:-n]
i_sizes = [V.graph.sizevars.evaluate_static_shape(i) for i in i_sizes]
assert len(scales_x) == n
o_sizes = output_size
inv_scales = [i / o for i, o in zip(i_sizes, o_sizes)]
for i, scale in enumerate(scales_x):
if scale is not None:
inv_scales[i] = 1.0 / scale
def scale_fn(x, scale, size):
# Nearest Exact: input_index = round(scale * (output_index + 0.5) - 0.5)
# = floor(scale * (output_index + 0.5))
# Nearest: input_index = floor(scale * output_index)
x = ops.index_expr(x, torch.float32)
if exact:
x = ops.add(x, ops.constant(0.5, torch.float32))
x = ops.mul(x, ops.constant(scale, torch.float32))
x = ops.to_dtype(x, torch.int32)
return ops.indirect_indexing(x, size, check=False)
def fn(idx):
x = idx[-n:]
b = idx[:-n]
return x_loader(
[*b, *[scale_fn(i, s, size) for i, s, size in zip(x, inv_scales, i_sizes)]]
)
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=fn,
ranges=[*batch, *o_sizes],
)
@register_lowering(aten.upsample_nearest1d.default)
def upsample_nearest1d(x, output_size, scales: Optional[float] = None):
return upsample_nearestnd(x, output_size, (scales,), n=1)
@register_lowering(aten._upsample_nearest_exact1d.default)
def _upsample_nearest_exact1d(x, output_size, scales: Optional[float] = None):
return upsample_nearestnd(x, output_size, (scales,), n=1, exact=True)
@register_lowering(aten.upsample_nearest2d.default)
def upsample_nearest2d(
x, output_size, scales_h: Optional[float] = None, scales_w: Optional[float] = None
):
return upsample_nearestnd(x, output_size, (scales_h, scales_w), n=2)
@register_lowering(aten._upsample_nearest_exact2d.default)
def _upsample_nearest_exact2d(
x, output_size, scales_h: Optional[float] = None, scales_w: Optional[float] = None
):
return upsample_nearestnd(x, output_size, (scales_h, scales_w), n=2, exact=True)
@register_lowering(aten.upsample_nearest3d.default)
def upsample_nearest3d(
x,
output_size,
scales_d: Optional[float] = None,
scales_h: Optional[float] = None,
scales_w: Optional[float] = None,
):
return upsample_nearestnd(x, output_size, (scales_d, scales_h, scales_w), n=3)
@register_lowering(aten._upsample_nearest_exact3d.default)
def _upsample_nearest_exact3d(
x,
output_size,
scales_d: Optional[float] = None,
scales_h: Optional[float] = None,
scales_w: Optional[float] = None,
):
return upsample_nearestnd(
x, output_size, (scales_d, scales_h, scales_w), n=3, exact=True
)
def _create_constants(*args, dtype):
return tuple(ops.constant(a, dtype) for a in args)
@register_lowering(aten.upsample_bicubic2d.default)
def upsample_bicubic2d_default(
x,
output_size,
align_corners: bool,
scales_h: Optional[float] = None,
scales_w: Optional[float] = None,
):
x.realize_hint()
x_loader = x.make_loader()
N, C, iH, iW = x.get_size()
oH, oW = output_size
iH = V.graph.sizevars.evaluate_static_shape(iH)
iW = V.graph.sizevars.evaluate_static_shape(iW)
def get_int_dtype(maxval):
if maxval > torch.iinfo(torch.int32).max:
return torch.int64
return torch.int32
def compute_scale(in_size, out_size, align_corners, scale=None):
if align_corners:
return (in_size - 1) / (out_size - 1) if out_size > 1 else 0
else:
return 1 / scale if scale is not None and scale > 0 else in_size / out_size
def compute_source_index(scale, dst_index, align_corners):
dst_index_ie = ops.index_expr(dst_index, torch.float32)
scale = ops.constant(scale, torch.float32)
if align_corners:
return ops.mul(scale, dst_index_ie)
else:
half = ops.constant(0.5, torch.float32)
return scale * (dst_index_ie + half) - half
def cubic_convolution1(x, A):
_Ap2, _Ap3, _1 = _create_constants(A + 2, A + 3, 1, dtype=torch.float32)
return (_Ap2 * x - _Ap3) * x * x + _1
def cubic_convolution2(x, A):
_A, _4A, _5A, _8A = _create_constants(
A, 4 * A, 5 * A, 8 * A, dtype=torch.float32
)
return ((_A * x - _5A) * x + _8A) * x - _4A
def get_cubic_upsample_coefficients(t):
A = -0.75
_1 = ops.constant(1.0, torch.float32)
c0 = cubic_convolution2(ops.add(t, _1), A)
c1 = cubic_convolution1(t, A)
x2 = ops.sub(_1, t)
c2 = cubic_convolution1(x2, A)
c3 = cubic_convolution2(ops.add(x2, _1), A)
return (c0, c1, c2, c3)
def cubic_interp1d(xs, t):
cs = get_cubic_upsample_coefficients(t)
# dot product between xs and cs
return xs[0] * cs[0] + xs[1] * cs[1] + xs[2] * cs[2] + xs[3] * cs[3]
height_scale = compute_scale(iH, oH, align_corners, scales_h)
width_scale = compute_scale(iW, oW, align_corners, scales_h)
def clamp(v, min, max):
return ops.maximum(min, ops.minimum(max, v))
def fn(idx):
n, c, oy, ox = idx
real_x = compute_source_index(width_scale, ox, align_corners)
in_x = ops.floor(real_x)
t_x = ops.sub(real_x, in_x)
real_y = compute_source_index(height_scale, oy, align_corners)
in_y = ops.floor(real_y)
t_y = ops.sub(real_y, in_y)
def load_bounded(fy, fx):
# TODO(Lezcano) Here we may not need to set-up a device_size
_0 = ops.constant(0, torch.int32)
iHm1 = ops.constant(iH - 1, torch.int32)
iWm1 = ops.constant(iW - 1, torch.int32)
iy = ops.indirect_indexing(clamp(fy, _0, iHm1), iH, check=False)
ix = ops.indirect_indexing(clamp(fx, _0, iWm1), iW, check=False)
return x_loader([n, c, iy, ix])
iy = ops.to_dtype(in_y, get_int_dtype(iH + 1))
ix = ops.to_dtype(in_x, get_int_dtype(iW + 1))
iys_ofs = tuple(ops.add(iy, ofs) for ofs in (-1, 0, 1, 2))
ixs_ofs = tuple(ops.add(ix, ofs) for ofs in (-1, 0, 1, 2))
def get_x_interp(y):
coeffs_x = tuple(load_bounded(y, x) for x in ixs_ofs)
return cubic_interp1d(coeffs_x, t_x)
coeffs_y = tuple(get_x_interp(y) for y in iys_ofs)
return cubic_interp1d(coeffs_y, t_y)
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=fn,
ranges=[N, C, sympy.Integer(oH), sympy.Integer(oW)],
)
@register_lowering(aten.reflection_pad1d_backward)
@register_lowering(aten.reflection_pad2d_backward)
@register_lowering(aten.reflection_pad3d_backward)
def _reflection_padnd_backward(grad_output, x, padding):
dim = len(padding) // 2
dhw = [h - 1 for h in x.get_size()[-dim:]]
grad_loader = grad_output.make_loader()
padding_left = [padding[2 * (dim - 1 - i)] for i in range(dim)]
padding_right = [padding[2 * (dim - 1 - i) + 1] for i in range(dim)]
def fn(idx):
b = idx[:-dim]
xyz = idx[-dim:]
def load_from_output(x):
return grad_loader([*b, *x])
def index_range_condition(index_range):
i, lb, ub = index_range
i = ops.index_expr(i, torch.int32)
lb = ops.index_expr(lb, torch.int64)
ub = ops.index_expr(ub, torch.int64)
return ops.and_(ops.ge(i, lb), ops.le(i, ub))
# Areas after reflection:
#
# top-left | top | top-right
# -----------------------------------------
# left | center | right
# -----------------------------------------
# bottom-left | bottom | bottom-right
#
# The center area is the original matrix. Other areas are reflections.
center = [xyz[i] + padding_left[i] for i in range(dim)]
left_reflect = [padding_left[i] - xyz[i] for i in range(dim)]
right_reflect = [2 * dhw[i] + padding_left[i] - xyz[i] for i in range(dim)]
# Accumulate gradients from different areas
# If some of the padding is negative, center load is not always valid
range_c = [
(center[i], 0, dhw[i] + padding_left[i] + padding_right[i])
for i in range(dim)
]
cond = functools.reduce(
ops.and_, [index_range_condition(range_c[i]) for i in range(dim)]
)
grad = ops.masked(cond, lambda: load_from_output(center), 0.0)
def accumulate(grad, out, index_ranges):
# If the upper bound is less than the lower bound, we can get rid of one accumulation.
# This happens when the padding size is zero.
for i in range(dim):
upper_less_than_lower = index_ranges[i][2] < index_ranges[i][1]
if isinstance(upper_less_than_lower, bool) and upper_less_than_lower:
return grad
cond = functools.reduce(
ops.and_,
[index_range_condition(index_range) for index_range in index_ranges],
)
g = ops.masked(cond, lambda: load_from_output(out), 0.0)
return ops.add(grad, g)
for area in itertools.product(*[[-1, 0, 1] for _ in range(dim)]):
if area == tuple([0] * dim):
# center, this is already done.
continue
outs = []
index_ranges = []
for i in range(dim):
if area[i] == 0:
out = center[i]
index_range = range_c[i]
elif area[i] == -1:
out = left_reflect[i]
index_range = (xyz[i], 1, padding_left[i])
elif area[i] == 1:
out = right_reflect[i]
index_range = (xyz[i], dhw[i] - padding_right[i], dhw[i] - 1)
outs.append(out) # type: ignore[possibly-undefined]
index_ranges.append(index_range) # type: ignore[possibly-undefined]
grad = accumulate(grad, outs, index_ranges)
return grad
return Pointwise.create(
device=grad_output.get_device(),
dtype=grad_output.get_dtype(),
inner_fn=fn,
ranges=list(x.get_size()),
)
@register_lowering(prims.rev.default)
def rev(x, dims):
# note - dims pre-canonicalized
x_loader = x.make_loader()
sizes = x.get_size()
def loader(idx):
idx = list(idx)
assert len(idx) == len(sizes)
for dim in dims:
idx[dim] = (sizes[dim] - 1) - idx[dim]
return x_loader(idx)
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=loader,
ranges=sizes,
)
@register_lowering(aten.constant_pad_nd, type_promotion_kind=None)
def constant_pad_nd(x, padding, fill_value=0):
assert (len(padding) % 2) == 0
if all(p == 0 for p in padding):
return clone(x)
sizes = x.get_size()
bounds = list(reversed(list(zip(padding[::2], padding[1::2]))))
n = len(sizes) - len(bounds)
# if padding is a complicated expression, hoist it
bounds_precomp: List[Tuple[sympy.Symbol, Any]] = []
for l, h in bounds:
bounds_precomp.append((V.graph.sizevars.lookup_precomputed_size(l), h)) # type: ignore[arg-type]
output_size = list(sizes[:n])
mask_sizes = []
for (low, high), size in zip(bounds, sizes[n:]):
mask_sizes.append(size)
output_size.append(sympy.expand(size + low + high))
assert len(output_size) == len(sizes)
fill_value = dtype_to_type(x.get_dtype())(fill_value)
def mask(index):
mask = []
for idx, (low, high), length in zip(index[n:], bounds, mask_sizes):
if low != 0:
mask.append(range_mask_low(idx, 0))
if high != 0:
mask.append(range_mask_high(idx, length))
mask = functools.reduce(ops.and_, mask)
return ops.masked(mask, lambda: x_loader(index), fill_value)
def offset_fn(index):
new_index = list(index[:n])
for idx, (low, high) in zip(index[n:], bounds_precomp):
new_index.append(idx - low)
assert len(new_index) == len(index)
return mask(new_index)
x_loader = x.make_loader()
return Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=offset_fn,
ranges=output_size,
)
def range_mask_low(i: sympy.Expr, low: Union[sympy.Expr, int]):
return ops.ge(
ops.index_expr(i, torch.int64),
ops.index_expr(sympy.Integer(low), torch.int64),
)
def range_mask_high(i: sympy.Expr, high: sympy.Expr):
return ops.lt(
ops.index_expr(i, torch.int64),
ops.index_expr(high, torch.int64),
)
def range_mask(i: sympy.Expr, high: sympy.Expr, low: sympy.Expr):
return ops.and_(
range_mask_low(i, low),
range_mask_high(i, high),
)
def constant_boundary_condition_2d(x, fill_value, padding=None, pad_fill_value=1.0):
*_, h, w = x.get_size()
x_loader = x.make_loader()
padding_h = padding[0] if padding else 0
padding_w = padding[1] if padding else 0
def load(index):
*prefix, ih, iw = index
mask = ops.and_(
range_mask(ih, h + padding_h, -padding_h),
range_mask(iw, w + padding_w, -padding_w),
)
return (
ops.masked(
mask,
lambda: constant_boundary_condition_2d(x, pad_fill_value)(
[*prefix, ih, iw]
),
fill_value,
)
if padding
else ops.masked(mask, lambda: x_loader([*prefix, ih, iw]), fill_value)
)
return load
def pooling_size(x, i, kernel_size, stride, padding, ceil_mode):
x_out = FloorDiv(
x + 2 * padding[i] - (kernel_size[i] - 1) + (stride[i] - 1), stride[i]
)
if ceil_mode:
x_alt = FloorDiv(
x + 2 * padding[i] - (kernel_size[i] - 1) + 2 * (stride[i] - 1), stride[i]
)
if V.graph.sizevars.size_hint((x_alt - 1) * stride[i] - x - padding[i]) >= 0:
# Sliding windows must start within the input or left padding
x_alt -= 1 # type: ignore[assignment]
V.graph.sizevars.guard_leq(0, x_alt * stride[i] - x - padding[i]) # type: ignore[arg-type]
if V.graph.sizevars.size_hint(x_out - x_alt) == 0:
# ceil mode is actually a no-op, lets guard on that
V.graph.sizevars.guard_equals(x_out, x_alt)
ceil_mode = False
else:
x_out = x_alt
return x_out, ceil_mode
fallback_max_pool2d_with_indices = fallback_handler(
aten.max_pool2d_with_indices.default,
add_to_fallback_set=False,
)
@register_lowering(aten.max_pool2d_with_indices, type_promotion_kind=None)
def max_pool2d_with_indices(
x, kernel_size, stride=None, padding=0, dilation=1, ceil_mode=False
):
if padding == 0:
padding = [0, 0]
if dilation == 1:
dilation = [1, 1]
if not stride:
stride = kernel_size
kernel_size = pad_listlike(kernel_size, 2)
stride = pad_listlike(stride, 2)
padding = pad_listlike(padding, 2)
dilation = pad_listlike(dilation, 2)
assert isinstance(x, TensorBox)
assert len(kernel_size) == 2
assert len(stride) == 2
assert len(padding) == 2
assert len(dilation) == 2
assert len(x.get_size()) in (3, 4)
x.realize_hint()
*batch, h, w = x.get_size()
h_out, ceil_mode1 = pooling_size(h, 0, kernel_size, stride, padding, ceil_mode)
w_out, ceil_mode2 = pooling_size(w, 1, kernel_size, stride, padding, ceil_mode)
if padding[0] or padding[1] or ceil_mode1 or ceil_mode2:
x_loader = constant_boundary_condition_2d(x, float("-inf"))
else:
x_loader = x.make_loader()
new_size = list(batch) + [h_out, w_out]
window_size = kernel_size[0] * kernel_size[1]
if window_size > 25 or any(d != 1 for d in dilation):
# Kernel size too big. Results in hard-to-optimize Triton code. Use fallback.
return fallback_max_pool2d_with_indices(
x, kernel_size, stride, padding, dilation, ceil_mode
)
def fn(idx, return_index):
*prefix, bh, bw = idx
maxval = None
maxindex = None
for ih, iw in itertools.product(range(kernel_size[0]), range(kernel_size[1])):
ih = bh * stride[0] + ih - padding[0]
iw = bw * stride[1] + iw - padding[1]
val = x_loader([*prefix, ih, iw])
if return_index:
index = ops.index_expr(ih * w + iw, torch.int64)
if maxindex is None:
maxindex = index
else:
maxindex = ops.where(ops.gt(val, maxval), index, maxindex)
if maxval is None:
maxval = val
else:
maxval = ops.maximum(val, maxval)
if return_index:
return maxindex
else:
return maxval
r1 = Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=functools.partial(fn, return_index=False),
ranges=new_size,
)
r2 = Pointwise.create(
device=x.get_device(),
dtype=torch.int64,
inner_fn=functools.partial(fn, return_index=True),
ranges=new_size,
)
# TODO(jansel): should we force these to be realized?
return r1, r2
fallback_max_pool2d_with_indices_backward = fallback_handler(
aten.max_pool2d_with_indices_backward.default,
add_to_fallback_set=False,
)
@register_lowering(aten.max_pool2d_with_indices_backward, type_promotion_kind=None)
def max_pool2d_with_indices_backward(
grad_output, x, kernel_size, stride, padding, dilation, ceil_mode, indices
):
if padding == 0:
padding = [0, 0]
if dilation == 1:
dilation = [1, 1]
if not stride:
stride = kernel_size
assert isinstance(x, TensorBox)
assert len(kernel_size) == 2
assert len(stride) == 2
assert len(padding) == 2
assert len(dilation) == 2
assert len(x.get_size()) in (3, 4)
# we will read this many times, so make sure it is computed
grad_output.realize_hint()
try:
gO_stride = grad_output.get_stride()
except AttributeError:
# some classes don't have `get_stride`
# TODO will need a better way of determining if inputs are channels-last
gO_stride = None
if isinstance(x, TensorBox) and isinstance(x.data.data, Pointwise): # type: ignore[attr-defined]
data = x.data.data # type: ignore[attr-defined]
x_buffer = ir.ComputedBuffer(
name=None,
layout=ir.FlexibleLayout(
device=data.get_device(),
dtype=data.get_dtype(),
size=data.get_size(),
),
data=data,
)
x_buffer.decide_layout()
x_stride = x_buffer.get_stride()
else:
try:
x_stride = x.get_stride()
except AttributeError:
x_stride = None
is_channels_last = (x_stride is not None and x_stride[1] == 1) or (
gO_stride is not None and gO_stride[1] == 1
)
autotune = (
config.coordinate_descent_tuning
or config.max_autotune
or config.max_autotune_pointwise
)
if any(d != 1 for d in dilation) or (is_channels_last and not autotune):
# don't codegen channels-last when autotune is not enabled, it's very slow
return fallback_max_pool2d_with_indices_backward(
grad_output, x, kernel_size, stride, padding, dilation, ceil_mode, indices
)
indices.realize_hint()
*batch, height, width = x.get_size()
*_, pooled_height, pooled_width = grad_output.get_size()
indices_loader = indices.make_loader()
grad_loader = grad_output.make_loader()
new_size = list(x.get_size())
h_window_size = max(
[
max(h // stride[0] - max(0, (h - kernel_size[0]) // stride[0]), 1)
for h in range(kernel_size[0] * 2)
]
)
w_window_size = max(
[
max(w // stride[1] - max(0, (w - kernel_size[1]) // stride[1]), 1)
for w in range(kernel_size[1] * 2)
]
)
window_size = h_window_size * w_window_size
if window_size > 25:
# Kernel size too big. Results in hard-to-optimize Triton code. Use fallback.
return fallback_max_pool2d_with_indices_backward(
grad_output, x, kernel_size, stride, padding, dilation, ceil_mode, indices
)
indices_size = indices.get_size()
def fn(idx):
*prefix, h, w = idx
index_test = ops.index_expr(h * width + w, torch.int32)
h = h + padding[0]
w = w + padding[1]
phstart = ops.index_expr(
FloorDiv(h - kernel_size[0] + stride[0], stride[0]), torch.int32
)
pwstart = ops.index_expr(
FloorDiv(w - kernel_size[1] + stride[1], stride[1]), torch.int32
)
phend = ops.index_expr(FloorDiv(h, stride[0]) + 1, torch.int32)
pwend = ops.index_expr(FloorDiv(w, stride[1]) + 1, torch.int32)
phstart = ops.maximum(phstart, ops.constant(0, torch.int32))
pwstart = ops.maximum(pwstart, ops.constant(0, torch.int32))
phend = ops.minimum(phend, ops.index_expr(pooled_height, torch.int32))
pwend = ops.minimum(pwend, ops.index_expr(pooled_width, torch.int32))
gradient = None
for ph_ in range(h_window_size):
for pw_ in range(w_window_size):
ph = ops.add(phstart, ops.constant(ph_, torch.int32))
pw = ops.add(pwstart, ops.constant(pw_, torch.int32))
grad_index = [
*prefix,
ops.indirect_indexing(
ops.minimum(ph, ops.sub(phend, ops.constant(1, torch.int32))),
indices_size[-2],
check=False,
),
ops.indirect_indexing(
ops.minimum(pw, ops.sub(pwend, ops.constant(1, torch.int32))),
indices_size[-1],
check=False,
),
]
index_actual = indices_loader(grad_index)
grad_part = grad_loader(grad_index)
check = ops.eq(index_actual, index_test)
if gradient is None:
# don't need mask for 0, 0
gradient = ops.where(
check, grad_part, ops.constant(0.0, torch.float32)
)
else:
mask = ops.and_(
ops.and_(
ops.lt(ph, phend),
ops.lt(pw, pwend),
),
check,
)
gradient = ops.where(mask, ops.add(gradient, grad_part), gradient)
assert gradient is not None
return gradient
return Pointwise.create(
device=grad_output.get_device(),
dtype=grad_output.get_dtype(),
inner_fn=fn,
ranges=new_size,
)
def pad_adaptive_loader(x, pad_val=0.0):
*_, h, w = x.get_size()
x_loader = x.make_loader()
def load(prefix, increments, start_indices, end_indices):
ih, iw = increments
h_start_index, w_start_index = start_indices
h_end_index, w_end_index = end_indices
mask = ops.and_(
ops.lt(
ops.index_expr(h_start_index + ih, torch.int64),
ops.index_expr(h_end_index, torch.int64),
),
ops.lt(
ops.index_expr(w_start_index + iw, torch.int64),
ops.index_expr(w_end_index, torch.int64),
),
)
return ops.masked(
mask,
lambda: x_loader([*prefix, h_start_index + ih, w_start_index + iw]),
pad_val,
)
return load
def _adaptive_pooling_idx_sum(kernel_maxes, start_index_fns, end_index_fns):
h_start_index_fn, w_start_index_fn = start_index_fns
h_end_index_fn, w_end_index_fn = end_index_fns
def fn_sum(idx, loader):
*prefix, bh, bw = idx
h_start_index = h_start_index_fn(bh)
h_end_index = h_end_index_fn(bh)
w_start_index = w_start_index_fn(bw)
w_end_index = w_end_index_fn(bw)
total = None
for ih, iw in itertools.product(range(kernel_maxes[0]), range(kernel_maxes[1])):
val = loader(
prefix,
[ih, iw],
[h_start_index, w_start_index],
[h_end_index, w_end_index],
)
if total is None:
total = val
else:
total = ops.add(val, total)
return total
return fn_sum
fallback_adaptive_avg_pool2d = fallback_handler(
aten._adaptive_avg_pool2d.default, add_to_fallback_set=False
)
@register_lowering(aten._adaptive_avg_pool2d)
def _adaptive_avg_pool2d(x, output_size):
assert isinstance(x, TensorBox)
assert len(output_size) == 2
x.realize_hint()
*batch, h_in, w_in = x.get_size()
h_in = V.graph.sizevars.evaluate_static_shape(h_in)
w_in = V.graph.sizevars.evaluate_static_shape(w_in)
h_out, w_out = output_size
# no-op if the same input and output
if h_in == h_out and w_in == w_out:
return clone(x)
if h_out == 0 or w_out == 0:
o_size = [*batch, h_out, w_out]
return empty(o_size, dtype=x.get_dtype(), device=x.get_device())
if h_in % h_out == 0 and w_in % w_out == 0:
kernel_size = [h_in // h_out, w_in // w_out]
return avg_pool2d(x, kernel_size)
h_kernel_max = ceildiv((h_in + h_out - 1), h_out)
w_kernel_max = ceildiv((w_in + w_out - 1), w_out)
new_size = list(batch) + [h_out, w_out]
dtype = x.get_dtype()
def start_index(index, out_dim, inp_dim):
return FloorDiv((index * inp_dim), out_dim)
def end_index(index, out_dim, inp_dim):
return FloorDiv((index + 1) * inp_dim + out_dim - 1, out_dim)
h_start_index = functools.partial(start_index, out_dim=h_out, inp_dim=h_in)
h_end_index = functools.partial(end_index, out_dim=h_out, inp_dim=h_in)
w_start_index = functools.partial(start_index, out_dim=w_out, inp_dim=w_in)
w_end_index = functools.partial(end_index, out_dim=w_out, inp_dim=w_in)
window_size = h_kernel_max * w_kernel_max
if window_size > 25:
# Kernel size too big. Results in hard-to-optimize Triton code. Use fallback.
return fallback_adaptive_avg_pool2d(x, output_size)
fn_sum = _adaptive_pooling_idx_sum(
[h_kernel_max, w_kernel_max],
[h_start_index, w_start_index],
[h_end_index, w_end_index],
)
ones_loader = pad_adaptive_loader(ones_like(x))
def fn(idx):
return ops.truediv(
fn_sum(idx, pad_adaptive_loader(x)), fn_sum(idx, ones_loader)
)
rv = Pointwise.create(
device=x.get_device(),
dtype=dtype,
inner_fn=fn,
ranges=new_size,
)
# TODO: should we force these to be realized?
return rv
def _adaptive_pooling_idx_max(kernel_maxes, in_sizes, out_sizes, return_index, loader):
# NOTE: There is some duplication between this and addaptive_avg_pool2d and max_pool2d
# Look into refactoring/deduplication after #116418 is merged.
h_in, w_in = in_sizes
h_out, w_out = out_sizes
def start_index(index, out_dim, inp_dim):
return FloorDiv((index * inp_dim), out_dim)
def end_index(index, out_dim, inp_dim):
return FloorDiv((index + 1) * inp_dim + out_dim - 1, out_dim)
h_start_index_fn = functools.partial(start_index, out_dim=h_out, inp_dim=h_in)
h_end_index_fn = functools.partial(end_index, out_dim=h_out, inp_dim=h_in)
w_start_index_fn = functools.partial(start_index, out_dim=w_out, inp_dim=w_in)
w_end_index_fn = functools.partial(end_index, out_dim=w_out, inp_dim=w_in)
def fn_max(idx):
*prefix, bh, bw = idx
h_start_index = h_start_index_fn(bh)
h_end_index = h_end_index_fn(bh)
w_start_index = w_start_index_fn(bw)
w_end_index = w_end_index_fn(bw)
maxval = None
maxindex = None
for ih, iw in itertools.product(range(kernel_maxes[0]), range(kernel_maxes[1])):
val = loader(
prefix,
[ih, iw],
[h_start_index, w_start_index],
[h_end_index, w_end_index],
)
index = ops.index_expr(
(h_start_index + ih) * w_in + w_start_index + iw, torch.int64
)
if return_index:
if maxindex is None:
maxindex = index
else:
maxindex = ops.where(ops.gt(val, maxval), index, maxindex)
if maxval is None:
maxval = val
else:
maxval = ops.maximum(val, maxval)
if return_index:
return maxindex
else:
return maxval
return fn_max
fallback_adaptive_max_pool2d = fallback_handler(
aten.adaptive_max_pool2d.default, add_to_fallback_set=False
)
@register_lowering(aten.adaptive_max_pool2d)
def adaptive_max_pool2d(x, output_size):
assert isinstance(x, TensorBox)
assert len(output_size) == 2
x.realize_hint()
*batch, h_in, w_in = x.get_size()
h_in = V.graph.sizevars.evaluate_static_shape(h_in)
w_in = V.graph.sizevars.evaluate_static_shape(w_in)
h_out, w_out = output_size
if h_out == 0 or w_out == 0:
o_size = [*batch, h_out, w_out]
return empty(o_size, dtype=x.get_dtype(), device=x.get_device()), empty(
o_size, dtype=torch.int64, device=x.get_device()
)
if h_in % h_out == 0 and w_in % w_out == 0:
kernel_size = [h_in // h_out, w_in // w_out]
return max_pool2d_with_indices(x, kernel_size)
h_kernel_max = ceildiv((h_in + h_out - 1), h_out)
w_kernel_max = ceildiv((w_in + w_out - 1), w_out)
new_size = list(batch) + [h_out, w_out]
dtype = x.get_dtype()
window_size = h_kernel_max * w_kernel_max
if window_size > 25:
# Kernel size too big. Results in hard-to-optimize Triton code. Use fallback.
return fallback_adaptive_max_pool2d(x, output_size)
inner_func_max_val = _adaptive_pooling_idx_max(
kernel_maxes=[h_kernel_max, w_kernel_max],
in_sizes=[h_in, w_in],
out_sizes=[h_out, w_out],
return_index=False,
loader=pad_adaptive_loader(x, float("-inf")),
)
inner_func_max_idx = _adaptive_pooling_idx_max(
kernel_maxes=[h_kernel_max, w_kernel_max],
in_sizes=[h_in, w_in],
out_sizes=[h_out, w_out],
return_index=True,
loader=pad_adaptive_loader(x, float("-inf")),
)
rv = Pointwise.create(
device=x.get_device(),
dtype=dtype,
inner_fn=inner_func_max_val,
ranges=new_size,
)
ri = Pointwise.create(
device=x.get_device(),
dtype=torch.int64,
inner_fn=inner_func_max_idx,
ranges=new_size,
)
return rv, ri
fallback_fractional_max_pool2d = fallback_handler(
aten.fractional_max_pool2d.default, add_to_fallback_set=False
)
def _fractional_pooling_offsets(samples, in_sz, out_sz, kernel_sz, dim):
out_sz = out_sz[dim]
in_sz = in_sz[dim]
kernel_sz = kernel_sz[dim]
alpha = (in_sz - kernel_sz) / (out_sz - 1)
samples_loader = samples.make_loader()
def load(prefix, i):
sample = samples_loader([*prefix, dim])
i_expr = ops.index_expr(i, samples.get_dtype())
alpha_expr = ops.index_expr(alpha, samples.get_dtype())
seq_i = ops.floor((i_expr + sample) * alpha_expr) - ops.floor(
sample * alpha_expr
)
seq_i = ops.to_dtype(seq_i, torch.int64)
mask = ops.lt(
i_expr,
ops.index_expr(out_sz - 1, torch.int64),
)
return ops.where(mask, seq_i, ops.index_expr(in_sz - kernel_sz, torch.int64))
return load
@register_lowering(aten.fractional_max_pool2d)
def fractional_max_pool2d(x, kernel_size, output_size, random_samples):
x.realize_hint()
*batch, inp_h, inp_w = x.get_size()
kernel_h, kernel_w = kernel_size
h_out, w_out = output_size
if kernel_h * kernel_w >= 25:
return fallback_fractional_max_pool2d(
x, kernel_size, output_size, random_samples
)
gen_offsets_for_dim = functools.partial(
_fractional_pooling_offsets,
samples=random_samples,
in_sz=[inp_h, inp_w],
out_sz=output_size,
kernel_sz=kernel_size,
)
h_index_fn = gen_offsets_for_dim(dim=0)
w_index_fn = gen_offsets_for_dim(dim=1)
x_loader = x.make_loader()
def fn(idx, return_index):
*prefix, bh, bw = idx
h_start_index = ops.indirect_indexing(h_index_fn(prefix, bh), inp_h)
w_start_index = ops.indirect_indexing(w_index_fn(prefix, bw), inp_w)
maxval = None
maxindex = None
for ih, iw in itertools.product(range(kernel_size[0]), range(kernel_size[1])):
val = x_loader([*prefix, h_start_index + ih, w_start_index + iw])
if return_index:
index = ops.index_expr(
(h_start_index + ih) * inp_w + w_start_index + iw, torch.int64
)
if maxindex is None:
maxindex = index
else:
maxindex = ops.where(
ops.or_(ops.gt(val, maxval), ops.isnan(val)), index, maxindex
)
if maxval is None:
maxval = val
else:
maxval = ops.maximum(val, maxval)
if return_index:
return maxindex
else:
return maxval
new_size = list(batch) + [h_out, w_out]
rv = Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=functools.partial(fn, return_index=False),
ranges=new_size,
)
ri = Pointwise.create(
device=x.get_device(),
dtype=torch.int64,
inner_fn=functools.partial(fn, return_index=True),
ranges=new_size,
)
return rv, ri
@register_lowering(aten.upsample_nearest2d_backward.default)
def upsample_nearest2d_backward(
x, output_size=None, input_size=None, scales_h=None, scales_w=None
):
x.realize_hint()
*batch, inp_h, inp_w = x.get_size()
inp_h = V.graph.sizevars.evaluate_static_shape(inp_h)
inp_w = V.graph.sizevars.evaluate_static_shape(inp_w)
*batch, out_h, out_w = input_size
if inp_h % out_h == 0 and inp_w % out_w == 0:
return avg_pool2d(x, [inp_h // out_h, inp_w // out_w], divisor_override=1)
h_kernel_max = ceildiv(inp_h, out_h)
w_kernel_max = ceildiv(inp_w, out_w)
def start_index(index, out_dim, inp_dim):
return CeilDiv(index * inp_dim, out_dim)
def end_index(index, out_dim, inp_dim):
return start_index((index + 1), out_dim, inp_dim)
h_start_index = functools.partial(start_index, out_dim=out_h, inp_dim=inp_h)
h_end_index = functools.partial(end_index, out_dim=out_h, inp_dim=inp_h)
w_start_index = functools.partial(start_index, out_dim=out_w, inp_dim=inp_w)
w_end_index = functools.partial(end_index, out_dim=out_w, inp_dim=inp_w)
fn_sum = _adaptive_pooling_idx_sum(
[h_kernel_max, w_kernel_max],
[h_start_index, w_start_index],
[h_end_index, w_end_index],
)
def fn(idx):
return fn_sum(idx, pad_adaptive_loader(x))
rv = Pointwise.create(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=fn,
ranges=list(input_size),
)
return rv
fallback_avg_pool2d = fallback_handler(
aten.avg_pool2d.default, add_to_fallback_set=False
)
@register_lowering(aten.avg_pool2d, type_promotion_kind=None)
def avg_pool2d(
x,
kernel_size,
stride=(),
padding=0,
ceil_mode=False,
count_include_pad=True,
divisor_override=None,
):
if not stride:
stride = kernel_size
if not padding:
padding = [0, 0]
kernel_size = pad_listlike(kernel_size, 2)
stride = pad_listlike(stride, 2)
padding = pad_listlike(padding, 2)
assert isinstance(x, TensorBox)
assert len(kernel_size) == 2
assert len(stride) == 2
assert len(padding) == 2
assert len(x.get_size()) in (3, 4)
x.realize_hint()
*batch, h, w = x.get_size()
h_out, ceil_mode1 = pooling_size(h, 0, kernel_size, stride, padding, ceil_mode)
w_out, ceil_mode2 = pooling_size(w, 1, kernel_size, stride, padding, ceil_mode)
if padding[0] or padding[1] or ceil_mode1 or ceil_mode2:
x_loader = constant_boundary_condition_2d(x, 0.0)
had_padding = True
else:
x_loader = x.make_loader()
had_padding = False
new_size = list(batch) + [h_out, w_out]
dtype = x.get_dtype()
window_size = kernel_size[0] * kernel_size[1]
if window_size > 25:
# Kernel size too big. Results in hard-to-optimize Triton code. Use fallback.
return fallback_avg_pool2d(
x,
kernel_size,
stride,
padding,
ceil_mode,
count_include_pad,
divisor_override,
)
def fn_sum(idx, loader):
*prefix, bh, bw = idx
total = None
for ih, iw in itertools.product(range(kernel_size[0]), range(kernel_size[1])):
ih = bh * stride[0] + ih - padding[0]
iw = bw * stride[1] + iw - padding[1]
val = loader([*prefix, ih, iw])
if total is None:
total = val
else:
total = ops.add(val, total)
return total
if not had_padding or divisor_override:
if divisor_override:
scale = 1 / divisor_override
else:
scale = 1.0 / (kernel_size[0] * kernel_size[1])
def fn(idx):
return ops.mul(fn_sum(idx, x_loader), ops.constant(scale, dtype))
else:
ones_loader = constant_boundary_condition_2d(
ones_like(x), 0.0, padding if count_include_pad else None
)
def fn(idx):
# TODO(jansel): optimize to do `int(x<h)` rather than `x<h?1:0`
return ops.truediv(fn_sum(idx, x_loader), fn_sum(idx, ones_loader))
rv = Pointwise.create(
device=x.get_device(),
dtype=dtype,
inner_fn=fn,
ranges=new_size,
)
# TODO(jansel): should we force these to be realized?
return rv
fallback_avg_pool2d_backward = fallback_handler(
aten.avg_pool2d_backward.default, add_to_fallback_set=False
)
@register_lowering(aten.avg_pool2d_backward, type_promotion_kind=None)
def avg_pool2d_backward(
grad_output,
x,
kernel_size,
stride,
padding,
ceil_mode,
count_include_pad,
divisor_override=None,
):
assert divisor_override is None or divisor_override != 0, "divisor must be not zero"
if not stride:
stride = kernel_size
if not padding:
padding = [0, 0]
assert isinstance(grad_output, TensorBox)
assert isinstance(x, TensorBox)
assert len(kernel_size) == 2
assert len(stride) == 2
assert len(padding) == 2
assert len(x.get_size()) in (3, 4)
grad_output.realize_hint() # we will read this many times, so make sure it is computed
*batch, height, width = x.get_size()
h_out, ceil_mode1 = pooling_size(height, 0, kernel_size, stride, padding, ceil_mode)
w_out, ceil_mode2 = pooling_size(width, 1, kernel_size, stride, padding, ceil_mode)
grad_loader = grad_output.make_loader()
had_padding = padding[0] or padding[1] or ceil_mode1 or ceil_mode2
*_, pooled_height, pooled_width = grad_output.get_size()
new_size = list(x.get_size())
dtype = x.get_dtype()
h_window_size = max(
[
max(h // stride[0] - max(0, (h - kernel_size[0]) // stride[0]), 1)
for h in range(kernel_size[0] * 2)
]
)
w_window_size = max(
[
max(w // stride[1] - max(0, (w - kernel_size[1]) // stride[1]), 1)
for w in range(kernel_size[1] * 2)
]
)
window_size = h_window_size * w_window_size
if window_size > 25:
# Kernel size too big. Results in hard-to-optimize Triton code. Use fallback.
return fallback_avg_pool2d_backward(
grad_output,
x,
kernel_size,
stride,
padding,
ceil_mode,
count_include_pad,
divisor_override,
)
def compute_pool_size_without_padding(ph, pw):
"""
This computes the scaling factor that we will divide an element
by when `count_include_pad=False`
"""
stride_h = ops.constant(stride[0], torch.int32)
stride_w = ops.constant(stride[1], torch.int32)
pad_h = ops.constant(padding[0], torch.int32)
pad_w = ops.constant(padding[1], torch.int32)
kernel_h = ops.constant(kernel_size[0], torch.int32)
kernel_w = ops.constant(kernel_size[1], torch.int32)
hstart = ops.sub(ops.mul(ph, stride_h), pad_h)
wstart = ops.sub(ops.mul(pw, stride_w), pad_w)
hend = ops.minimum(
ops.add(hstart, kernel_h),
ops.add(ops.index_expr(height, torch.int32), pad_h),
)
wend = ops.minimum(
ops.add(wstart, kernel_w),
ops.add(ops.index_expr(width, torch.int32), pad_w),
)
hstart = ops.maximum(hstart, ops.constant(0, torch.int32))
wstart = ops.maximum(wstart, ops.constant(0, torch.int32))
hend = ops.minimum(hend, ops.index_expr(height, torch.int32))
wend = ops.minimum(wend, ops.index_expr(width, torch.int32))
divide_factor = ops.mul(ops.sub(hend, hstart), ops.sub(wend, wstart))
return divide_factor
def fn(idx):
*prefix, h, w = idx
h = h + padding[0]
w = w + padding[1]
phstart = ops.index_expr(
FloorDiv(h - kernel_size[0] + stride[0], stride[0]), torch.int32
)
pwstart = ops.index_expr(
FloorDiv(w - kernel_size[1] + stride[1], stride[1]), torch.int32
)
phend = ops.index_expr(FloorDiv(h, stride[0]) + 1, torch.int32)
pwend = ops.index_expr(FloorDiv(w, stride[1]) + 1, torch.int32)
phstart = ops.maximum(phstart, ops.constant(0, torch.int32))
pwstart = ops.maximum(pwstart, ops.constant(0, torch.int32))
phend = ops.minimum(phend, ops.index_expr(pooled_height, torch.int32))
pwend = ops.minimum(pwend, ops.index_expr(pooled_width, torch.int32))
gradient = None
for ph_ in range(h_window_size):
for pw_ in range(w_window_size):
ph = ops.add(phstart, ops.constant(ph_, torch.int32))
pw = ops.add(pwstart, ops.constant(pw_, torch.int32))
if divisor_override is not None:
scale = divisor_override
elif count_include_pad or not had_padding:
scale = kernel_size[0] * kernel_size[1]
else:
scale = compute_pool_size_without_padding(ph, pw)
part = ops.truediv(
grad_loader(
[
*prefix,
ops.indirect_indexing(
ops.minimum(
ph, ops.sub(phend, ops.constant(1, torch.int32))
),
pooled_height,
check=False,
),
ops.indirect_indexing(
ops.minimum(
pw, ops.sub(pwend, ops.constant(1, torch.int32))
),
pooled_width,
check=False,
),
]
),
scale,
)
mask = ops.and_(
ops.lt(ph, phend),
ops.lt(pw, pwend),
)
if gradient is None:
gradient = ops.where(mask, part, ops.constant(0.0, torch.float32))
else:
gradient = ops.where(mask, ops.add(gradient, part), gradient)
assert gradient is not None
return gradient
rv = Pointwise.create(
device=grad_output.get_device(),
dtype=dtype,
inner_fn=fn,
ranges=new_size,
)
return rv
def _validate_reduction_axis(x, axis):
size = x.get_size()
if isinstance(axis, int):
axis = [axis]
elif not axis:
axis = range(len(size))
if len(size) == 0:
assert tuple(axis) in [(), (0,), (-1,)], f"invalid axis: {axis}"
return []
axis = list(axis)
for i in range(len(axis)):
if axis[i] < 0:
axis[i] += len(size) if len(size) else 1
assert 0 <= axis[i] < len(size) or (len(size) == 0 and axis[i] == 0)
assert len(set(axis)) == len(axis), "reduction axis not unique"
return axis
def _make_reduction_inner(x, *, axis, keepdims, dtype, override_return_dtype):
if dtype is not None:
x = to_dtype(x, dtype)
size = x.get_size()
axis = set(_validate_reduction_axis(x, axis))
kept_sizes = []
kept_idx = []
reduced_sizes = []
reduced_idx = []
for i in range(len(size)):
if i in axis:
reduced_idx.append(i)
reduced_sizes.append(size[i])
else:
kept_idx.append(i)
kept_sizes.append(size[i])
def loader(index, reduction_index):
assert len(reduction_index) == len(reduced_idx)
if keepdims:
assert len(index) == len(size)
index = [index[i] for i in kept_idx]
assert len(index) == len(kept_idx)
new_index = [None] * (len(index) + len(reduction_index))
for idx, var in itertools.chain(
zip(kept_idx, index), zip(reduced_idx, reduction_index)
):
new_index[idx] = var
return inner_loader(new_index)
if keepdims:
new_size = list(size)
for i in reduced_idx:
new_size[i] = sympy.Integer(1)
else:
new_size = kept_sizes
inner_loader = x.make_loader()
return dict(
device=x.get_device(),
dst_dtype=override_return_dtype or x.get_dtype(),
src_dtype=x.get_dtype(),
inner_fn=loader,
ranges=new_size,
reduction_ranges=reduced_sizes,
)
def make_reduction(reduction_type: str, override_return_dtype=None):
def inner(x, axis=None, keepdims=False, *, dtype=None):
kwargs = _make_reduction_inner(
x,
axis=axis,
keepdims=keepdims,
dtype=dtype,
override_return_dtype=override_return_dtype,
)
result = Reduction.create(reduction_type=reduction_type, input_node=x, **kwargs)
if isinstance(
result.data.data, Reduction
): # Only realize if reduction isn't unrolled
result.realize()
return result
return inner
def _make_scan_inner(x, *, axis, dtype):
if dtype is not None:
x = to_dtype(x, dtype)
size = x.get_size()
axis = _validate_dim(x, axis)
return dict(
device=x.get_device(),
dtype=x.get_dtype(),
inner_fn=x.make_loader(),
size=x.get_size(),
axis=axis,
)
@register_lowering(aten.mean)
def mean(x, axis=None, keepdim=False, *, dtype=None):
if dtype is not None:
x = to_dtype(x, dtype)
size = x.get_size()
axis = _validate_reduction_axis(x, axis)
# compute in higher-precision until end of mean lowering
output_dtype = x.get_dtype()
if output_dtype in (torch.float16, torch.bfloat16):
x = to_dtype(x, torch.float)
sum_result = sum_(x, axis, keepdim)
denom = sympy_product(size[i] for i in axis)
denom = ir.IndexingConstant(denom, x.get_dtype(), x.get_device())
denom = ExpandView.create(denom, list(sum_result.get_size()))
return to_dtype(div(sum_result, denom), output_dtype)
def var_mean_sum_(x, axis, correction, keepdim, return_mean):
if correction is None:
correction = 1
size = x.get_size()
axis = _validate_reduction_axis(x, axis)
x_mean = mean(x, axis, keepdim=True)
if return_mean:
x_mean.realize()
diffs = square(sub(x, x_mean))
sum_result = sum_(diffs, axis, keepdim)
denom = sympy_product(size[i] for i in axis)
if correction:
denom = sympy.Max(denom - correction, 0)
denom = ir.IndexingConstant(denom, x.get_dtype(), x.get_device())
denom = ExpandView.create(denom, list(sum_result.get_size()))
x_var = div(sum_result, denom)
if not return_mean:
return (x_var,)
x_mean = x_mean if keepdim else squeeze(x_mean, axis)
return x_var, x_mean
def use_two_step_variance(x, axis, keepdim):
# Instead of unrolling welford, just unroll the simpler two-step var
axis = _validate_reduction_axis(x, axis)
kwargs = _make_reduction_inner(
x, axis=axis, keepdims=keepdim, dtype=None, override_return_dtype=None
)
ranges = kwargs["ranges"]
reduction_numel = sympy_product(kwargs["reduction_ranges"])
return (
isinstance(reduction_numel, sympy.Integer)
and int(reduction_numel) < config.unroll_reductions_threshold
and sympy_product(ranges) != 1
)
def var_mean_welford_(x, axis, *, correction, keepdim, return_mean):
if correction is None:
correction = 1
kwargs = _make_reduction_inner(
x, axis=axis, keepdims=keepdim, dtype=None, override_return_dtype=None
)
loader = kwargs.pop("inner_fn")
kwargs.pop("dst_dtype")
kwargs.pop("src_dtype")
mean, m2, _ = ir.WelfordReduction.create(
inner_fns=(loader,),
reduction_type="welford_reduce",
dtype=x.get_dtype(),
**kwargs,
)
m2.realize()
dtype = x.get_dtype()
size = x.get_size()
axis = _validate_reduction_axis(x, axis)
rnumel = sympy_product(size[i] for i in axis)
def get_constant_or_index_expr(x, dtype):
if isinstance(x, sympy.Expr) and not x.is_number:
return ops.to_dtype(ops.index_expr(x, torch.int64), dtype)
return ops.constant(x, dtype)
def scale_fn(data):
c = get_constant_or_index_expr(correction, dtype)
N = get_constant_or_index_expr(rnumel, dtype)
zero = ops.constant(0, dtype)
return data / ops.maximum(zero, N - c)
var = make_pointwise(scale_fn)(m2)
if return_mean:
mean.realize()
return var, mean
return (var,)
def var_mean_helper_(x, *, axis, correction, keepdim, return_mean):
out_dtype = x.get_dtype()
compute_dtype = get_computation_dtype(out_dtype)
x = to_dtype(x, compute_dtype, copy=False)
kwargs = dict(
x=x,
axis=axis,
correction=correction,
keepdim=keepdim,
return_mean=return_mean,
)
output = (
var_mean_sum_(**kwargs)
if use_two_step_variance(x, axis=axis, keepdim=keepdim)
else var_mean_welford_(**kwargs)
)
output = tuple(to_dtype(x, out_dtype, copy=False) for x in output)
return output[0] if not return_mean else output
@register_lowering([aten.var, prims.var])
def var_(x, axis=None, *, correction=None, keepdim=False):
return var_mean_helper_(
x, axis=axis, correction=correction, keepdim=keepdim, return_mean=False
)
@register_lowering(aten.var_mean)
def var_mean(x, axis=None, *, correction=None, keepdim=False):
return var_mean_helper_(
x, axis=axis, correction=correction, keepdim=keepdim, return_mean=True
)
def pow_recursive(x, y, dtype):
if y < 0:
return pow_recursive(ops.reciprocal(x), -y, dtype)
if y == 0:
return ops.constant(1, dtype)
if y == 1:
return x
result = pow_recursive(x, y // 2, dtype)
result = ops.mul(result, result)
if (y % 2) == 1:
result = ops.mul(result, x)
return result
@make_pointwise
def pow_native(a, b):
return ops.pow(a, b)
fallback_pow_tensor_tensor = fallback_handler(
aten.pow.Tensor_Tensor, add_to_fallback_set=False
)
fallback_pow_scalar = fallback_handler(aten.pow.Scalar, add_to_fallback_set=False)
fallback_pow_tensor_scalar = fallback_handler(
aten.pow.Tensor_Scalar, add_to_fallback_set=False
)
@register_lowering(aten.pow, broadcast=True)
def pow(a, b):
if isinstance(b, float) and b == int(b):
return pow(a, int(b))
elif isinstance(b, float) and b == 0.5:
return sqrt(a)
elif isinstance(b, int) and b == 1:
return clone(a)
# Type promotion ensures all tensor arguments have the same type
dtype = next(x.get_dtype() for x in (a, b) if isinstance(x, ir.TensorBox))
is_integer_pow = is_integer_dtype(dtype)
# Optimize away small fixed powers, or for integers avoid falling back to ATen
embed_exponent = isinstance(b, int) and (
-32 < b < 32 or (is_integer_pow and b >= 0)
)
if embed_exponent:
loader = a.make_loader()
def fn(idx):
return pow_recursive(loader(idx), b, a.get_dtype())
return Pointwise.create(
device=a.get_device(),
dtype=a.get_dtype(),
inner_fn=fn,
ranges=a.get_size(),
)
if isinstance(a, Number):
if a == 1:
return full_like(b, 1)
if a == 2 and is_float_dtype(b.get_dtype()):
return exp2(b)
if is_integer_pow:
# ops.pow doesn't work for integers
if isinstance(a, Number):
return fallback_pow_scalar(a, b)
elif isinstance(b, Number):
return fallback_pow_tensor_scalar(a, b)
else:
return fallback_pow_tensor_tensor(a, b)
return pow_native(a, b)
def mutate_to(changed, val, unsafe_alias=False):
if isinstance(changed, TensorBox):
changed_data = changed.data
else:
changed_data = changed
if isinstance(val, TensorBox):
val = val.data
if not isinstance(val, ir.StorageBox):
# introduce a copy to handle views
val = Pointwise.create(
device=changed.get_device(),
dtype=changed.get_dtype(),
inner_fn=val.make_loader(),
ranges=changed.get_size(),
).data
assert isinstance(val, ir.StorageBox)
if isinstance(changed_data, ir.StorageBox) and not (
changed_data.is_input_buffer() or isinstance(changed_data.data, ir.NopKernel)
):
# Fast path, just swing the data pointer
val.realize()
changed_data.data = val.data
return changed
ir.MutationLayout.realize_into(val, changed_data, unsafe_alias=unsafe_alias)
return changed
@register_lowering(aten.fill_)
def fill_(x, fill_value):
return mutate_to(x, full_like(x, fill_value))
@register_lowering(aten.copy_, type_promotion_kind=None)
def copy_(dst, src, non_blocking=False):
src = to_device(src, dst.get_device())
src = to_dtype(src, dst.get_dtype())
src = expand(src, dst.get_size())
return mutate_to(dst, src)
@make_pointwise
def floordiv(a, b):
return ops.floordiv(a, b)
@make_pointwise
def truncdiv(a, b):
return ops.truncdiv(a, b)
@register_lowering(aten.div, broadcast=True)
def div_mode(a, b, rounding_mode=None):
both_integer = is_integer_type(a) and is_integer_type(b)
both_boolean = is_boolean_type(a) and is_boolean_type(b)
# floordiv and truncdiv need special handling for integer tensors on Triton,
# see the discussion at https://github.com/openai/triton/issues/605
if rounding_mode == "floor":
assert not both_boolean, "floordiv operands can not be boolean at the same time"
return floordiv(a, b) if both_integer else floor(div(a, b))
if rounding_mode == "trunc":
assert not both_boolean, "truncdiv operands can not be boolean at the same time"
return truncdiv(a, b) if both_integer else trunc(div(a, b))
return div(a, b)
@register_lowering([aten.mul], broadcast=True)
def mul(a, b):
both_bool = is_boolean_type(a) and is_boolean_type(b)
if both_bool:
return logical_and(a, b)
else:
fn = ops_wrapper(aten.mul.__name__)
return make_pointwise(fn)(a, b)
# NOTE: prims.div maps to a / b in C, so performs truncation division on
# integer inputs and true division for floating and complex inputs.
@register_lowering([prims.div], broadcast=True)
def div_prim(a, b):
is_integral = all(is_boolean_type(x) or is_integer_type(x) for x in [a, b])
if is_integral:
return truncdiv(a, b)
def fn(*args):
return ops.truediv(*args)
return make_pointwise(fn)(a, b)
@register_lowering(
[aten.true_divide, aten.div.Tensor],
broadcast=True,
type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.INT_TO_FLOAT,
)
def div(a, b):
a, b = promote_constants(
(a, b), type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.INT_TO_FLOAT
)
return div_prim(a, b)
@register_lowering([aten.fmod, prims.fmod], broadcast=True)
def fmod(a, b):
is_integral = is_boolean_type(a) or is_integer_type(a)
if is_integral:
def fn(a, b):
return ops.mod(a, b)
else:
def fn(a, b):
return ops.fmod(a, b)
return make_pointwise(fn)(a, b)
@register_lowering(aten.rsqrt)
def rsqrt(x):
dtype = x.get_dtype()
if is_integer_dtype(dtype) or is_boolean_dtype(dtype):
x = to_dtype(x, torch.get_default_dtype())
def _rsqrt(x):
return ops.rsqrt(x)
return make_pointwise(_rsqrt)(x)
@register_lowering([aten.sum, prims.sum])
def sum_(x, axis=None, keepdims=False, *, dtype=None):
if (
is_integer_dtype(x.get_dtype()) or is_boolean_dtype(x.get_dtype())
) and dtype is None:
dtype = torch.int64
fn = make_reduction("sum", override_return_dtype=dtype)
return fn(x, axis, keepdims, dtype=dtype)
fallback_cumsum = fallback_handler(aten.cumsum.default)
fallback_cumprod = fallback_handler(aten.cumprod.default)
fallback_logcumsumexp = fallback_handler(aten.logcumsumexp.default)
@register_lowering(aten.cumsum)
def cumsum(x, axis=None, dtype=None):
if (
is_integer_dtype(x.get_dtype()) or is_boolean_dtype(x.get_dtype())
) and dtype is None:
dtype = torch.int64
if len(x.get_size()) == 0:
assert axis in [0, -1]
dtype = dtype or x.get_dtype()
return to_dtype(x, dtype, copy=True)
kwargs = _make_scan_inner(x, axis=axis, dtype=dtype)
result = ir.Scan.create(**kwargs, combine_fn=ops.add, init=0)
if result is None:
return fallback_cumsum(x, dim=axis, dtype=dtype)
return result
@register_lowering(aten.cumprod)
def cumprod(x, axis=None, dtype=None):
if (
is_integer_dtype(x.get_dtype()) or is_boolean_dtype(x.get_dtype())
) and dtype is None:
dtype = torch.int64
if len(x.get_size()) == 0:
assert axis in [0, -1]
dtype = dtype or x.get_dtype()
return to_dtype(x, dtype, copy=True)
kwargs = _make_scan_inner(x, axis=axis, dtype=dtype)
result = ir.Scan.create(**kwargs, combine_fn=ops.mul, init=1)
if result is None:
return fallback_cumprod(x, dim=axis, dtype=dtype)
return result
@register_lowering(aten.logcumsumexp)
def logcumsumexp(x, dim):
def log_add_exp_helper(a, b):
min_v = ops.minimum(a, b)
max_v = ops.maximum(a, b)
mask = (min_v != max_v) | (~ops.isinf(min_v))
return ops.where(mask, ops.log1p(ops.exp(min_v - max_v)) + max_v, a)
dtype = x.get_dtype()
if len(x.get_size()) == 0:
assert dim in [0, -1]
return clone(x)
kwargs = _make_scan_inner(x, axis=dim, dtype=dtype)
result = ir.Scan.create(**kwargs, combine_fn=log_add_exp_helper, init=float("-inf"))
if result is None:
return fallback_logcumsumexp(x, dim=dim)
return result
@register_lowering(aten.prod)
def prod(x, axis=None, keepdims=False, *, dtype=None):
if (
is_integer_dtype(x.get_dtype()) or is_boolean_dtype(x.get_dtype())
) and dtype is None:
dtype = torch.int64
fn = make_reduction("prod", override_return_dtype=dtype)
return fn(x, axis, keepdims, dtype=dtype)
@register_lowering(aten.any)
def reduce_any(x, dim=None, keepdim=False):
x = to_dtype(x, torch.bool)
return make_reduction("any")(x, axis=dim, keepdims=keepdim)
@register_lowering(aten.max, type_promotion_kind=None)
def reduce_max(x, dim=None, keepdim=False):
if dim is not None:
return (
reduce_amax(x, axis=dim, keepdims=keepdim),
reduce_argmax(x, axis=dim, keepdims=keepdim),
)
return reduce_amax(x, axis=None, keepdims=keepdim)
@register_lowering(aten.min, type_promotion_kind=None)
def reduce_min(x, dim=None, keepdim=False):
if dim is not None:
return (
reduce_amin(x, axis=dim, keepdims=keepdim),
reduce_argmin(x, axis=dim, keepdims=keepdim),
)
return reduce_amin(x, axis=None, keepdims=keepdim)
register_lowering(prims.xor_sum)(make_reduction("xor_sum"))
reduce_amax = register_lowering(aten.amax)(make_reduction("max"))
reduce_amin = register_lowering(aten.amin)(make_reduction("min"))
reduce_argmax = register_lowering(aten.argmax)(
make_reduction("argmax", override_return_dtype=torch.int64)
)
reduce_argmin = register_lowering(aten.argmin)(
make_reduction("argmin", override_return_dtype=torch.int64)
)
add = register_pointwise(
aten.add, allow_alpha=True, override_fn_when_input_bool="logical_or"
)
def register_pointwise_numeric(op, name=None, triton_fallback=None):
return register_pointwise(
op,
name=name,
type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.INT_TO_FLOAT,
triton_fallback=triton_fallback,
)
def register_pointwise_numeric_ldf64(op):
return register_pointwise(
op,
type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.INT_TO_FLOAT,
use_libdevice_for_f64=True,
)
exp = register_pointwise_numeric_ldf64(aten.exp)
exp2 = register_pointwise_numeric(aten.exp2)
expm1 = register_pointwise_numeric(aten.expm1)
relu = register_pointwise(aten.relu)
sigmoid = register_pointwise_numeric_ldf64(aten.sigmoid)
sqrt = register_pointwise_numeric_ldf64(aten.sqrt)
square = register_pointwise(aten.square)
sub = register_pointwise(aten.sub, allow_alpha=True)
register_pointwise_numeric_ldf64(aten.cos)
register_pointwise_numeric_ldf64(aten.sin)
abs = register_pointwise(aten.abs)
bitwise_and = register_pointwise(aten.bitwise_and)
bitwise_left_shift = register_pointwise(aten.bitwise_left_shift)
bitwise_not = register_pointwise(
aten.bitwise_not, override_fn_when_input_bool="logical_not"
)
bitwise_or = register_pointwise(aten.bitwise_or)
bitwise_right_shift = register_pointwise(aten.bitwise_right_shift)
bitwise_xor = register_pointwise(aten.bitwise_xor)
register_pointwise_numeric(aten.lgamma)
erf = register_pointwise_numeric(aten.erf)
register_lowering(
aten.special_erf, type_promotion_kind=ELEMENTWISE_TYPE_PROMOTION_KIND.INT_TO_FLOAT
)(erf)
register_pointwise_numeric(aten.log1p)
register_pointwise_numeric(aten.tan)
register_pointwise_numeric(aten.tanh)
register_pointwise_numeric_ldf64(aten.log)
logical_and = register_pointwise(
aten.logical_and,
type_promotion_kind=None,
convert_input_to_bool=True,
override_return_dtype=torch.bool,
)
logical_not = register_pointwise(
aten.logical_not,
type_promotion_kind=None,
convert_input_to_bool=True,
override_return_dtype=torch.bool,
)
logical_or = register_pointwise(
aten.logical_or,
type_promotion_kind=None,
convert_input_to_bool=True,
override_return_dtype=torch.bool,
)
logical_xor = register_pointwise(
aten.logical_xor,
type_promotion_kind=None,
convert_input_to_bool=True,
override_return_dtype=torch.bool,
)
maximum = register_pointwise(aten.maximum)
minimum = register_pointwise(aten.minimum)
register_lowering(aten.clamp_min)(maximum)
register_lowering(aten.clamp_max)(minimum)
neg = register_pointwise(aten.neg)
abs = register_pointwise(aten.abs)
reciprocal = register_pointwise_numeric(aten.reciprocal)
register_pointwise(aten.remainder)
sign = register_pointwise(aten.sign, override_fn_when_input_bool="identity")
register_pointwise(aten.ceil)
register_pointwise(aten.signbit, override_return_dtype=torch.bool)
register_lowering(aten._neg_view)(neg)
register_pointwise(aten.le, override_return_dtype=torch.bool)
register_pointwise(aten.lt, override_return_dtype=torch.bool)
register_pointwise(aten.ge, override_return_dtype=torch.bool)
gt = register_pointwise(aten.gt, override_return_dtype=torch.bool)
register_pointwise(aten.eq, override_return_dtype=torch.bool)
register_pointwise(aten.ne, override_return_dtype=torch.bool)
register_pointwise_numeric(aten.cosh)
register_pointwise_numeric(aten.sinh)
register_pointwise_numeric(aten.acos)
register_pointwise_numeric(aten.acosh)
register_pointwise_numeric(aten.asin)
register_pointwise_numeric(aten.asinh)
register_pointwise_numeric(aten.atan2)
register_pointwise_numeric(aten.atan)
register_pointwise_numeric(aten.atanh)
register_pointwise_numeric(aten.copysign)
register_pointwise_numeric(aten.erfc)
register_pointwise_numeric(aten.erfinv)
register_pointwise_numeric(aten.hypot)
register_pointwise_numeric(aten.log10)
register_pointwise_numeric(aten.nextafter)
from .codegen.common import pointwise_overrides_data
def _get_pointwise_overrides(ns, name):
data = pointwise_overrides_data[name]
op = getattr(ns, data.name, None)
if op is None:
return
def make_triton_fallback(op):
if data.triton is None:
return fallback_handler(op)
if isinstance(op, torch._ops.OpOverloadPacket):
for olname in op.overloads():
ol = getattr(op, olname)
yield ol, data.type_promotion_kind, make_triton_fallback(ol)
else:
yield op, data.type_promotion_kind, make_triton_fallback(op)
for name in pointwise_overrides_data:
for op, type_promotion_kind, triton_fallback in _get_pointwise_overrides(
aten, name
):
register_pointwise(
op,
name=name,
type_promotion_kind=type_promotion_kind,
triton_fallback=triton_fallback,
)
for op, type_promotion_kind, triton_fallback in _get_pointwise_overrides(
prims, name
):
register_pointwise(
op,
name=name,
type_promotion_kind=type_promotion_kind,
triton_fallback=triton_fallback,
)
foreach_add_list = register_foreach_pointwise(
aten._foreach_add.List, add, allow_alpha=True
)
foreach_add_scalar = register_foreach_pointwise(
aten._foreach_add.Scalar, add, allow_alpha=True
)
register_foreach_pointwise(aten._foreach_add.Tensor, add, allow_alpha=True)
foreach_mul_list = register_foreach_pointwise(aten._foreach_mul.List, mul)
foreach_mul_scalar = register_foreach_pointwise(aten._foreach_mul.Scalar, mul)
register_foreach_pointwise(aten._foreach_sub.List, sub)
register_foreach_pointwise(aten._foreach_sub.Scalar, sub)
register_foreach_pointwise(aten._foreach_neg.default, neg)
register_foreach_pointwise(aten._foreach_abs.default, abs)
register_foreach_pointwise(aten._foreach_pow.Scalar, pow)
register_foreach_pointwise(aten._foreach_pow.ScalarAndTensor, pow)
foreach_div_list = register_foreach_pointwise(aten._foreach_div.List, div)
foreach_div_scalar = register_foreach_pointwise(aten._foreach_div.Scalar, div)
register_foreach_pointwise(aten._foreach_sqrt, sqrt)
register_foreach_pointwise(aten._foreach_maximum.List, maximum)
register_foreach_pointwise(aten._foreach_maximum.Scalar, maximum)
register_foreach_pointwise(aten._foreach_minimum.List, minimum)
register_foreach_pointwise(aten._foreach_minimum.Scalar, minimum)
register_foreach_pointwise(aten._foreach_clamp_min.List, maximum)
register_foreach_pointwise(aten._foreach_clamp_min.Scalar, maximum)
register_foreach_pointwise(aten._foreach_clamp_max.List, minimum)
register_foreach_pointwise(aten._foreach_clamp_max.Scalar, minimum)
register_foreach_pointwise(aten._foreach_reciprocal, reciprocal)
register_foreach_pointwise(aten._foreach_sign, sign)
register_foreach_pointwise(aten._foreach_copy, copy)
# these are only encountered as outputs of the graph
# reinplacing epilogue copies improves compile time
# by removing extra buffers sent to the scheduler.
def register_foreach_inplace(aten_op, outplace_aten_op, outplace_op):
inplaceable_foreach_ops[outplace_aten_op] = aten_op
inplace_foreach_ops.add(aten_op)
def fn(*args, **kwargs):
results = outplace_op(*args, **kwargs)
mut_results = []
for arg, result in zip(args[0], results):
mut_results.append(mutate_to(arg, result, unsafe_alias=True))
return mut_results
_register_foreach_lowering(aten_op, fn)
register_foreach_inplace(
aten._foreach_add_.List, aten._foreach_add.List, foreach_add_list
)
register_foreach_inplace(
aten._foreach_add_.Scalar, aten._foreach_add.Scalar, foreach_add_scalar
)
register_foreach_inplace(
aten._foreach_mul_.List, aten._foreach_mul.List, foreach_mul_list
)
register_foreach_inplace(
aten._foreach_mul_.Scalar, aten._foreach_mul.Scalar, foreach_mul_scalar
)
register_foreach_inplace(
aten._foreach_div_.List, aten._foreach_div.List, foreach_div_list
)
register_foreach_inplace(
aten._foreach_div_.Scalar, aten._foreach_div.Scalar, foreach_div_scalar
)
def register_inplace(aten_op, outplace_op):
@register_lowering(aten_op, type_promotion_kind=None)
def fn(*args, **kwargs):
result = outplace_op(*args, **kwargs)
result = to_dtype(result, args[0].get_dtype())
return mutate_to(args[0], result)
return fn
register_inplace(aten.add_, add)
register_inplace(aten.bitwise_and_, bitwise_and)
register_inplace(aten.bitwise_left_shift_, bitwise_left_shift)
register_inplace(aten.bitwise_not_, bitwise_not)
register_inplace(aten.bitwise_or_, bitwise_or)
register_inplace(aten.bitwise_right_shift_, bitwise_right_shift)
register_inplace(aten.bitwise_xor_, bitwise_xor)
register_inplace(aten.mul_, mul)
register_inplace(aten.div_.Tensor, div)
register_inplace(aten.div_.Tensor_mode, div_mode)
register_inplace(aten.logical_and_, logical_and)
register_inplace(aten.logical_not_, logical_not)
register_inplace(aten.logical_or_, logical_or)
register_inplace(aten.logical_xor_, logical_xor)
register_inplace(aten.sub_, sub)
register_inplace(aten.relu_, relu)
register_inplace(aten.sigmoid_, sigmoid)
register_lowering(aten.__and__)(bitwise_and)
register_lowering(aten.__lshift__)(bitwise_left_shift)
register_lowering(aten.__or__)(bitwise_or)
register_lowering(aten.__rshift__)(bitwise_right_shift)
register_lowering(aten.__xor__)(bitwise_xor)
register_inplace(aten.__iand__, aten.__and__)
register_inplace(aten.__ilshift__, aten.__lshift__)
register_inplace(aten.__ior__, aten.__or__)
register_inplace(aten.__irshift__, aten.__rshift__)
register_inplace(aten.__ixor__, aten.__xor__)
@register_lowering(aten.sym_constrain_range)
def sym_constrain_range(a, min=None, max=None):
tracing_context = torch._guards.TracingContext.try_get()
assert (
tracing_context is None or a in tracing_context.fake_mode.shape_env.var_to_range
)
return a
@register_lowering(aten.sym_size.int)
def sym_size(a, dim):
val = V.graph.current_node.meta["val"]
# Note [Can val be an int?]
# ~~~~~~~~~~~~~~~~~~~~~~~~~
# In principle, someone could construct an FX graph where
# a call to size/stride has a val that is a plain int (not
# SymInt). However, we will maintain the invariant that
# this is not possible: if you are constructing an FX graph
# where there is a call to size/stride that returns an
# int, but you KNOW that int must always be a constant,
# then you do not need trace that call at all (and just
# constant propagate the integer as is.)
assert isinstance(val, torch.SymInt)
return val.node.expr
@register_lowering(aten.sym_stride.int)
def sym_stride(a, dim):
val = V.graph.current_node.meta["val"]
# See Note [Can val be an int?]
assert isinstance(val, torch.SymInt)
return val.node.expr
@register_lowering(aten.sym_numel)
def sym_numel(a):
return a.get_numel()
for method, func in magic_methods.items():
register_lowering(method_to_operator(method))(func)
@register_lowering(aten._foobar)
def foobar(self, *args, **kwargs):
raise NotImplementedError("Helpful for debugging")
@register_lowering(torch.ops._inductor_test.realize)
def _realize(x):
x.realize()
return clone(x)
@register_lowering(torch.ops.inductor.resize_storage_bytes_)
def resize_storage_bytes_(variable, new_size):
variable.realize()
ir.ResizeStorageBytes(variable, new_size)
return variable
from torch._higher_order_ops.auto_functionalize import auto_functionalized
make_fallback(auto_functionalized)
@register_lowering(triton_kernel_wrapper_mutation)
def triton_kernel_wrap_(*, kernel_idx, grid, kwargs):
ir.UserDefinedTritonKernel(kernel_idx=kernel_idx, grid=grid, kernel_args=kwargs)
return {key: val for key, val in kwargs.items() if isinstance(val, TensorBox)}
@register_lowering(triton_kernel_wrapper_functional)
def triton_kernel_wrap(*, kernel_idx, grid, kwargs, tensors_to_clone):
new_kwargs = {}
for name, value in kwargs.items():
if isinstance(value, ir.TensorBox):
x = value.data
has_non_rv_views = False
while isinstance(x, ir.BaseView):
if not isinstance(x, ir.ReinterpretView):
has_non_rv_views = True
break
x = x.data
if has_non_rv_views:
# we realize the inputs wrapped into any view which is not
# ReinterpretView to convert them into ReinterpretView during
# realization; all views being ReinterpretView is assumed by
# the downstream code (e.g., preserving ReinterpretView in
# cloning; layout should be available in mutation marking)
value = ir.TensorBox(ir.ExternKernel.realize_input(value))
if name in tensors_to_clone:
value = clone_preserve_reinterpret_view(value)
new_kwargs[name] = value
return triton_kernel_wrap_(kernel_idx=kernel_idx, grid=grid, kwargs=new_kwargs)
@register_lowering(torch.ops.higher_order.cond)
def cond(pred, true_fn, false_fn, operands):
if is_triton(pred) or any(map(is_triton, operands)):
msg = "control flow operator: torch.cond."
if stack_trace := V.graph.current_node.meta.get("stack_trace", None):
msg = f"{msg} Found from : \n {stack_trace}"
V.graph.disable_cudagraphs_reason = msg
result = ir.Conditional.create(pred, true_fn, false_fn, operands)
return list(map(TensorBox.create, result))
try:
import torch.distributed._functional_collectives
c10d_functional = torch.ops.c10d_functional
@register_lowering(c10d_functional.wait_tensor)
def wait(input):
return TensorBox.create(ir.Wait.create(input))
@register_lowering(c10d_functional.broadcast)
def broadcast(input, src, tag, ranks, group_size):
return ir.Broadcast.create(input, src, tag, ranks, group_size)
@register_lowering(c10d_functional.all_reduce)
def allreduce(input, reduce_op, tag, ranks, group_size):
return ir.AllReduce.create(input, reduce_op, tag, ranks, group_size)
@register_lowering(c10d_functional.all_gather_into_tensor)
def all_gather_into_tensor(shard, tag, ranks, group_size):
return TensorBox.create(
ir.AllGatherIntoTensor.create(
ir.ExternKernel.require_contiguous(shard), tag, ranks, group_size
)
)
@register_lowering(c10d_functional.reduce_scatter_tensor)
def reduce_scatter_tensor(input, reduce_op, tag, ranks, group_size):
return TensorBox.create(
ir.ReduceScatterTensor.create(input, reduce_op, tag, ranks, group_size)
)
@register_lowering(c10d_functional.all_reduce_coalesced)
def all_reduce_coalesced(input, reduce_op, tag, ranks, group_size):
return ir.AllReduceCoalesced.create(input, reduce_op, tag, ranks, group_size)
@register_lowering(c10d_functional.all_gather_into_tensor_coalesced)
def all_gather_into_tensor_coalesced(self, tag, ranks, group_size):
result = ir.AllGatherIntoTensorCoalesced.create(self, tag, ranks, group_size)
return list(map(TensorBox.create, result))
@register_lowering(c10d_functional.reduce_scatter_tensor_coalesced)
def reduce_scatter_tensor_coalesced(self, reduceOp, tag, ranks, group_size):
result = ir.ReduceScatterTensorCoalesced.create(
self, reduceOp, tag, ranks, group_size
)
return list(map(TensorBox.create, result))
@register_lowering(c10d_functional.all_to_all_single)
def all_to_all_single(
self, output_split_sizes, input_split_sizes, tag, ranks, group_size
):
return TensorBox.create(
ir.AllToAllSingle.create(
self, output_split_sizes, input_split_sizes, tag, ranks, group_size
)
)
_c10d_functional = torch.ops._c10d_functional
@register_lowering(_c10d_functional.all_reduce)
def _all_reduce(inp, reduce_op, group_name):
inp = clone(inp)
ir._CollectiveKernel.create_inplace(
_c10d_functional.all_reduce_.default, inp, reduce_op, group_name
)
return inp
@register_lowering(_c10d_functional.all_reduce_)
def _all_reduce_(inp, reduce_op, group_name):
ir._CollectiveKernel.create_inplace(
_c10d_functional.all_reduce_.default, inp, reduce_op, group_name
)
return inp
@register_lowering(_c10d_functional.all_reduce_coalesced)
def _all_reduce_coalesced(inputs, reduce_op, group_name):
inputs = [clone(inp) for inp in inputs]
ir._CollectiveKernel.create_inplace(
_c10d_functional.all_reduce_coalesced_.default,
inputs,
reduce_op,
group_name,
)
return inputs
@register_lowering(_c10d_functional.all_reduce_coalesced_)
def _all_reduce_coalesced_(inputs, reduce_op, group_name):
ir._CollectiveKernel.create_inplace(
_c10d_functional.all_reduce_coalesced_.default,
inputs,
reduce_op,
group_name,
)
return inputs
@register_lowering(_c10d_functional.all_gather_into_tensor)
def _all_gather_into_tensor(inp, group_size, group_name):
return ir.TensorBox.create(
ir._CollectiveKernel.create_out_of_place(
_c10d_functional.all_gather_into_tensor.default,
inp,
group_size,
group_name,
)
)
@register_lowering(_c10d_functional.all_gather_into_tensor_coalesced)
def _all_gather_into_tensor_coalesced(inputs, group_size, group_name):
return pytree.tree_map(
ir.TensorBox.create,
ir._CollectiveKernel.create_out_of_place(
_c10d_functional.all_gather_into_tensor_coalesced.default,
inputs,
group_size,
group_name,
),
)
@register_lowering(_c10d_functional.reduce_scatter_tensor)
def _reduce_scatter_tensor(inp, reduce_op, group_size, group_name):
return ir.TensorBox.create(
ir._CollectiveKernel.create_out_of_place(
_c10d_functional.reduce_scatter_tensor.default,
inp,
reduce_op,
group_size,
group_name,
)
)
@register_lowering(_c10d_functional.reduce_scatter_tensor_coalesced)
def _reduce_scatter_tensor_coalesced(inputs, reduce_op, group_size, group_name):
return pytree.tree_map(
ir.TensorBox.create,
ir._CollectiveKernel.create_out_of_place(
_c10d_functional.reduce_scatter_tensor_coalesced.default,
inputs,
reduce_op,
group_size,
group_name,
),
)
@register_lowering(_c10d_functional.all_to_all_single)
def _all_to_all_single(inp, output_split_sizes, input_split_sizes, group_name):
return ir.TensorBox.create(
ir._CollectiveKernel.create_out_of_place(
_c10d_functional.all_to_all_single.default,
inp,
output_split_sizes,
input_split_sizes,
group_name,
)
)
@register_lowering(_c10d_functional.broadcast)
def _broadcast(inp, src, group_name):
inp = clone(inp)
ir._CollectiveKernel.create_inplace(
_c10d_functional.broadcast_.default, inp, src, group_name
)
return inp
@register_lowering(_c10d_functional.broadcast_)
def _broadcast_(inp, src, group_name):
ir._CollectiveKernel.create_inplace(
_c10d_functional.broadcast_.default, inp, src, group_name
)
return inp
@register_lowering(_c10d_functional.wait_tensor)
def _wait_tensor(inp):
ir._WaitKernel.create_wait(_c10d_functional.wait_tensor.default, inp)
return inp
except ImportError:
log.info(
"Inductor support for distributed collectives depends on building torch.distributed"
)
# populate lowerings defined in kernel/*
from . import kernel
import_submodule(kernel)
from . import quantized_lowerings
quantized_lowerings.register_quantized_ops()