M-Labs/nac3
M-Labs
/
nac3
8
0
Fork 10
Code Issues 98 Pull Requests 20 Releases Wiki Activity

ndstrides: [10] Reimplement binops, unary ops, and cmpops. #520

Open
lyken wants to merge 4 commits from ndstrides-10-ops into ndstrides-9-subassign
pull from: ndstrides-10-ops
merge into: M-Labs:ndstrides-9-subassign
Conversation 1 Commits 4 Files Changed 3 +375 -170
3 changed files with 375 additions and 170 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

241
nac3core/src/codegen/expr.rs
View File

@ -8,7 +8,10 @@ use std::{
use inkwell::{
attributes::{Attribute, AttributeLoc},
types::{AnyType, BasicType, BasicTypeEnum},
values::{BasicValueEnum, CallSiteValue, FunctionValue, IntValue, PointerValue, StructValue},
values::{
BasicValue, BasicValueEnum, CallSiteValue, FunctionValue, IntValue, PointerValue,
StructValue,
},
AddressSpace, IntPredicate, OptimizationLevel,
};
use itertools::{chain, izip, Either, Itertools};
@ -20,8 +23,8 @@ use nac3parser::ast::{
use super::{
classes::{
ArrayLikeIndexer, ArrayLikeValue, ListType, ListValue, NDArrayValue, ProxyType, ProxyValue,
RangeValue, UntypedArrayLikeAccessor,
ArrayLikeIndexer, ArrayLikeValue, ListType, ListValue, ProxyType, ProxyValue, RangeValue,
UntypedArrayLikeAccessor,
},
concrete_type::{ConcreteFuncArg, ConcreteTypeEnum, ConcreteTypeStore},
gen_in_range_check, get_llvm_abi_type, get_llvm_type, get_va_count_arg_name,
@ -31,10 +34,13 @@ use super::{
call_int_umin, call_memcpy_generic,
},
macros::codegen_unreachable,
need_sret, numpy,
need_sret,
object::{
any::AnyObject,
ndarray::{indexing::util::gen_ndarray_subscript_ndindices, NDArrayObject},
ndarray::{
indexing::util::gen_ndarray_subscript_ndindices, NDArrayObject, NDArrayOut,
ScalarOrNDArray,
},
},
stmt::{
gen_for_callback_incrementing, gen_if_callback, gen_if_else_expr_callback, gen_raise,
@ -44,7 +50,7 @@ use super::{
};
use crate::{
symbol_resolver::{SymbolValue, ValueEnum},
toplevel::{helper::PrimDef, numpy::unpack_ndarray_var_tys, DefinitionId, TopLevelDef},
toplevel::{helper::PrimDef, DefinitionId, TopLevelDef},
typecheck::{
magic_methods::{Binop, BinopVariant, HasOpInfo},
typedef::{FunSignature, FuncArg, Type, TypeEnum, TypeVarId, Unifier, VarMap},
@ -1549,99 +1555,71 @@ pub fn gen_binop_expr_with_values<'ctx, G: CodeGenerator>(
} else if ty1.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id())
|| ty2.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id())
{
let llvm_usize = generator.get_size_type(ctx.ctx);
let left =
ScalarOrNDArray::split_object(generator, ctx, AnyObject { ty: ty1, value: left_val });
let right =
ScalarOrNDArray::split_object(generator, ctx, AnyObject { ty: ty2, value: right_val });
let is_ndarray1 = ty1.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id());
let is_ndarray2 = ty2.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id());
// Inhomogeneous binary operations are not supported.
assert!(ctx.unifier.unioned(left.get_dtype(), right.get_dtype()));
if is_ndarray1 && is_ndarray2 {
let (ndarray_dtype1, _) = unpack_ndarray_var_tys(&mut ctx.unifier, ty1);
let (ndarray_dtype2, _) = unpack_ndarray_var_tys(&mut ctx.unifier, ty2);
let common_dtype = left.get_dtype();
assert!(ctx.unifier.unioned(ndarray_dtype1, ndarray_dtype2));
let left_val =
NDArrayValue::from_ptr_val(left_val.into_pointer_value(), llvm_usize, None);
let right_val =
NDArrayValue::from_ptr_val(right_val.into_pointer_value(), llvm_usize, None);
let res = if op.base == Operator::MatMult {
// MatMult is the only binop which is not an elementwise op
numpy::ndarray_matmul_2d(
generator,
ctx,
ndarray_dtype1,
match op.variant {
BinopVariant::Normal => None,
BinopVariant::AugAssign => Some(left_val),
},
left_val,
right_val,
)?
let out = match op.variant {
BinopVariant::Normal => NDArrayOut::NewNDArray { dtype: common_dtype },
BinopVariant::AugAssign => {
// If this is an augmented assignment.
// `left` has to be an ndarray. If it were a scalar then NAC3 simply doesn't support it.
if let ScalarOrNDArray::NDArray(out_ndarray) = left {
NDArrayOut::WriteToNDArray { ndarray: out_ndarray }
} else {
numpy::ndarray_elementwise_binop_impl(
generator,
ctx,
ndarray_dtype1,
match op.variant {
BinopVariant::Normal => None,
BinopVariant::AugAssign => Some(left_val),
},
(left_val.as_base_value().into(), false),
(right_val.as_base_value().into(), false),
|generator, ctx, (lhs, rhs)| {
gen_binop_expr_with_values(
generator,
ctx,
(&Some(ndarray_dtype1), lhs),
op,
(&Some(ndarray_dtype2), rhs),
ctx.current_loc,
)?
.unwrap()
.to_basic_value_enum(
ctx,
generator,
ndarray_dtype1,
)
},
)?
panic!("left must be an ndarray")
}
}
};
Ok(Some(res.as_base_value().into()))
if op.base == Operator::MatMult {
// Handle matrix multiplication.
todo!()
} else {
let (ndarray_dtype, _) =
unpack_ndarray_var_tys(&mut ctx.unifier, if is_ndarray1 { ty1 } else { ty2 });
let ndarray_val = NDArrayValue::from_ptr_val(
if is_ndarray1 { left_val } else { right_val }.into_pointer_value(),
llvm_usize,
None,
);
let res = numpy::ndarray_elementwise_binop_impl(
// For other operations, they are all elementwise operations.
// There are only three cases:
// - LHS is a scalar, RHS is an ndarray.
// - LHS is an ndarray, RHS is a scalar.
// - LHS is an ndarray, RHS is an ndarray.
//
// For all cases, the scalar operand is promoted to an ndarray,
// the two are then broadcasted, and starmapped through.
let left = left.to_ndarray(generator, ctx);
let right = right.to_ndarray(generator, ctx);
let result = NDArrayObject::broadcast_starmap(
generator,
ctx,
ndarray_dtype,
match op.variant {
BinopVariant::Normal => None,
BinopVariant::AugAssign => Some(ndarray_val),
},
(left_val, !is_ndarray1),
(right_val, !is_ndarray2),
|generator, ctx, (lhs, rhs)| {
gen_binop_expr_with_values(
&[left, right],
out,
|generator, ctx, scalars| {
let left_value = scalars[0];
let right_value = scalars[1];
let result = gen_binop_expr_with_values(
generator,
ctx,
(&Some(ndarray_dtype), lhs),
(&Some(left.dtype), left_value),
op,
(&Some(ndarray_dtype), rhs),
(&Some(right.dtype), right_value),
ctx.current_loc,
)?
.unwrap()
.to_basic_value_enum(ctx, generator, ndarray_dtype)
},
)?;
.to_basic_value_enum(ctx, generator, common_dtype)?;
Ok(Some(res.as_base_value().into()))
Ok(result)
},
)
.unwrap();
Ok(Some(ValueEnum::Dynamic(result.instance.value.as_basic_value_enum())))
}
} else {
let left_ty_enum = ctx.unifier.get_ty_immutable(left_ty.unwrap());
@ -1799,14 +1777,12 @@ pub fn gen_unaryop_expr_with_values<'ctx, G: CodeGenerator>(
_ => val.into(),
}
} else if ty.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id()) {
let llvm_usize = generator.get_size_type(ctx.ctx);
let (ndarray_dtype, _) = unpack_ndarray_var_tys(&mut ctx.unifier, ty);
let val = NDArrayValue::from_ptr_val(val.into_pointer_value(), llvm_usize, None);
let ndarray = AnyObject { value: val, ty };
let ndarray = NDArrayObject::from_object(generator, ctx, ndarray);
// ndarray uses `~` rather than `not` to perform elementwise inversion, convert it before
// passing it to the elementwise codegen function
let op = if ndarray_dtype.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::Bool.id()) {
let op = if ndarray.dtype.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::Bool.id()) {
if op == ast::Unaryop::Invert {
ast::Unaryop::Not
} else {
@ -1820,20 +1796,18 @@ pub fn gen_unaryop_expr_with_values<'ctx, G: CodeGenerator>(
op
};
let res = numpy::ndarray_elementwise_unaryop_impl(
let mapped_ndarray = ndarray.map(
generator,
ctx,
ndarray_dtype,
None,
val,
|generator, ctx, val| {
gen_unaryop_expr_with_values(generator, ctx, op, (&Some(ndarray_dtype), val))?
NDArrayOut::NewNDArray { dtype: ndarray.dtype },
|generator, ctx, scalar| {
gen_unaryop_expr_with_values(generator, ctx, op, (&Some(ndarray.dtype), scalar))?
.unwrap()
.to_basic_value_enum(ctx, generator, ndarray_dtype)
.to_basic_value_enum(ctx, generator, ndarray.dtype)
},
)?;
res.as_base_value().into()
ValueEnum::Dynamic(mapped_ndarray.instance.value.as_basic_value_enum())
} else {
unimplemented!()
}))
@ -1876,39 +1850,33 @@ pub fn gen_cmpop_expr_with_values<'ctx, G: CodeGenerator>(
if left_ty.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id())
|| right_ty.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id())
{
let llvm_usize = generator.get_size_type(ctx.ctx);
let (Some(left_ty), lhs) = left else { codegen_unreachable!(ctx) };
let (Some(right_ty), rhs) = comparators[0] else { codegen_unreachable!(ctx) };
let (Some(left_ty), left) = left else { codegen_unreachable!(ctx) };
let (Some(right_ty), right) = comparators[0] else { codegen_unreachable!(ctx) };
let op = ops[0];
let is_ndarray1 =
left_ty.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id());
let is_ndarray2 =
right_ty.obj_id(&ctx.unifier).is_some_and(|id| id == PrimDef::NDArray.id());
let left = AnyObject { value: left, ty: left_ty };
let left =
ScalarOrNDArray::split_object(generator, ctx, left).to_ndarray(generator, ctx);
return if is_ndarray1 && is_ndarray2 {
let (ndarray_dtype1, _) = unpack_ndarray_var_tys(&mut ctx.unifier, left_ty);
let (ndarray_dtype2, _) = unpack_ndarray_var_tys(&mut ctx.unifier, right_ty);
let right = AnyObject { value: right, ty: right_ty };
let right =
ScalarOrNDArray::split_object(generator, ctx, right).to_ndarray(generator, ctx);
assert!(ctx.unifier.unioned(ndarray_dtype1, ndarray_dtype2));
let left_val =
NDArrayValue::from_ptr_val(lhs.into_pointer_value(), llvm_usize, None);
let res = numpy::ndarray_elementwise_binop_impl(
let result_ndarray = NDArrayObject::broadcast_starmap(
generator,
ctx,
ctx.primitives.bool,
None,
(left_val.as_base_value().into(), false),
(rhs, false),
|generator, ctx, (lhs, rhs)| {
&[left, right],
NDArrayOut::NewNDArray { dtype: ctx.primitives.bool },
|generator, ctx, scalars| {
let left_scalar = scalars[0];
let right_scalar = scalars[1];
let val = gen_cmpop_expr_with_values(
generator,
ctx,
(Some(ndarray_dtype1), lhs),
(Some(left.dtype), left_scalar),
&[op],
&[(Some(ndarray_dtype2), rhs)],
&[(Some(right.dtype), right_scalar)],
)?
.unwrap()
.to_basic_value_enum(
@ -1921,40 +1889,7 @@ pub fn gen_cmpop_expr_with_values<'ctx, G: CodeGenerator>(
},
)?;
Ok(Some(res.as_base_value().into()))
} else {
let (ndarray_dtype, _) = unpack_ndarray_var_tys(
&mut ctx.unifier,
if is_ndarray1 { left_ty } else { right_ty },
);
let res = numpy::ndarray_elementwise_binop_impl(
generator,
ctx,
ctx.primitives.bool,
None,
(lhs, !is_ndarray1),
(rhs, !is_ndarray2),
|generator, ctx, (lhs, rhs)| {
let val = gen_cmpop_expr_with_values(
generator,
ctx,
(Some(ndarray_dtype), lhs),
&[op],
&[(Some(ndarray_dtype), rhs)],
)?
.unwrap()
.to_basic_value_enum(
ctx,
generator,
ctx.primitives.bool,
)?;
Ok(generator.bool_to_i8(ctx, val.into_int_value()).into())
},
)?;
Ok(Some(res.as_base_value().into()))
};
return Ok(Some(result_ndarray.instance.value.into()));
}
}

219
nac3core/src/codegen/object/ndarray/map.rs Normal file
View File

@ -0,0 +1,219 @@
use inkwell::values::BasicValueEnum;
use itertools::Itertools;
use crate::{
codegen::{
object::ndarray::{AnyObject, NDArrayObject},
stmt::gen_for_callback,
CodeGenContext, CodeGenerator,
},
typecheck::typedef::Type,
};
use super::{nditer::NDIterHandle, NDArrayOut, ScalarOrNDArray};
impl<'ctx> NDArrayObject<'ctx> {
/// Generate LLVM IR to broadcast `ndarray`s together, and starmap through them with `mapping` elementwise.
///
/// `mapping` is an LLVM IR generator. The input of `mapping` is the list of elements when iterating through
/// the input `ndarrays` after broadcasting. The output of `mapping` is the result of the elementwise operation.
///
/// `out` specifies whether the result should be a new ndarray or to be written an existing ndarray.
pub fn broadcast_starmap<'a, G, MappingFn>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
ndarrays: &[Self],
out: NDArrayOut<'ctx>,
mapping: MappingFn,
) -> Result<Self, String>
where
G: CodeGenerator + ?Sized,
MappingFn: FnOnce(
&mut G,
&mut CodeGenContext<'ctx, 'a>,
&[BasicValueEnum<'ctx>],
) -> Result<BasicValueEnum<'ctx>, String>,
{
// Broadcast inputs
let broadcast_result = NDArrayObject::broadcast(generator, ctx, ndarrays);
let out_ndarray = match out {
NDArrayOut::NewNDArray { dtype } => {
// Create a new ndarray based on the broadcast shape.
let result_ndarray =
NDArrayObject::alloca(generator, ctx, dtype, broadcast_result.ndims);
result_ndarray.copy_shape_from_array(generator, ctx, broadcast_result.shape);
result_ndarray.create_data(generator, ctx);
result_ndarray
}
NDArrayOut::WriteToNDArray { ndarray: result_ndarray } => {
// Use an existing ndarray.
// Check that its shape is compatible with the broadcast shape.
result_ndarray.assert_can_be_written_by_out(
generator,
ctx,
broadcast_result.ndims,
broadcast_result.shape,
);
result_ndarray
}
};
// Map element-wise and store results into `mapped_ndarray`.
let nditer = NDIterHandle::new(generator, ctx, out_ndarray);
gen_for_callback(
generator,
ctx,
Some("broadcast_starmap"),
|generator, ctx| {
// Create NDIters for all broadcasted input ndarrays.
let other_nditers = broadcast_result
.ndarrays
.iter()
.map(|ndarray| NDIterHandle::new(generator, ctx, *ndarray))
.collect_vec();
Ok((nditer, other_nditers))
},
|generator, ctx, (out_nditer, _in_nditers)| {
// We can simply use `out_nditer`'s `has_element()`.
// `in_nditers`' `has_element()`s should return the same value.
Ok(out_nditer.has_element(generator, ctx).value)
},
|generator, ctx, _hooks, (out_nditer, in_nditers)| {
// Get all the scalars from the broadcasted input ndarrays, pass them to `mapping`,
// and write to `out_ndarray`.
let in_scalars = in_nditers
.iter()
.map(|nditer| nditer.get_scalar(generator, ctx).value)
.collect_vec();
let result = mapping(generator, ctx, &in_scalars)?;
let p = out_nditer.get_pointer(generator, ctx);
ctx.builder.build_store(p, result).unwrap();
Ok(())
},
|generator, ctx, (out_nditer, in_nditers)| {
// Advance all iterators
out_nditer.next(generator, ctx);
in_nditers.iter().for_each(|nditer| nditer.next(generator, ctx));
Ok(())
},
)?;
Ok(out_ndarray)
}
/// Map through this ndarray with an elementwise function.
pub fn map<'a, G, Mapping>(
&self,
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
out: NDArrayOut<'ctx>,
mapping: Mapping,
) -> Result<Self, String>
where
G: CodeGenerator + ?Sized,
Mapping: FnOnce(
&mut G,
&mut CodeGenContext<'ctx, 'a>,
BasicValueEnum<'ctx>,
) -> Result<BasicValueEnum<'ctx>, String>,
{
NDArrayObject::broadcast_starmap(
generator,
ctx,
&[*self],
out,
|generator, ctx, scalars| mapping(generator, ctx, scalars[0]),
)
}
}
impl<'ctx> ScalarOrNDArray<'ctx> {
/// Starmap through a list of inputs using `mapping`, where an input could be an ndarray, a scalar.
///
/// This function is very helpful when implementing NumPy functions that takes on either scalars or ndarrays or a mix of them
/// as their inputs and produces either an ndarray with broadcast, or a scalar if all its inputs are all scalars.
///
/// For example ,this function can be used to implement `np.add`, which has the following behaviors:
/// - `np.add(3, 4) = 7` # (scalar, scalar) -> scalar
/// - `np.add(3, np.array([4, 5, 6]))` # (scalar, ndarray) -> ndarray; the first `scalar` is converted into an ndarray and broadcasted.
/// - `np.add(np.array([[1], [2], [3]]), np.array([[4, 5, 6]]))` # (ndarray, ndarray) -> ndarray; there is broadcasting.
///
/// ## Details:
///
/// If `inputs` are all [`ScalarOrNDArray::Scalar`], the output will be a [`ScalarOrNDArray::Scalar`] with type `ret_dtype`.
///
/// Otherwise (if there are any [`ScalarOrNDArray::NDArray`] in `inputs`), all inputs will be 'as-ndarray'-ed into ndarrays,
/// then all inputs (now all ndarrays) will be passed to [`NDArrayObject::broadcasting_starmap`] and **create** a new ndarray
/// with dtype `ret_dtype`.
pub fn broadcasting_starmap<'a, G, MappingFn>(
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
inputs: &[ScalarOrNDArray<'ctx>],
ret_dtype: Type,
mapping: MappingFn,
) -> Result<ScalarOrNDArray<'ctx>, String>
where
G: CodeGenerator + ?Sized,
MappingFn: FnOnce(
&mut G,
&mut CodeGenContext<'ctx, 'a>,
&[BasicValueEnum<'ctx>],
) -> Result<BasicValueEnum<'ctx>, String>,
{
// Check if all inputs are Scalars
let all_scalars: Option<Vec<_>> = inputs.iter().map(AnyObject::try_from).try_collect().ok();
if let Some(scalars) = all_scalars {
let scalars = scalars.iter().map(|scalar| scalar.value).collect_vec();
let value = mapping(generator, ctx, &scalars)?;
Ok(ScalarOrNDArray::Scalar(AnyObject { ty: ret_dtype, value }))
} else {
// Promote all input to ndarrays and map through them.
let inputs = inputs.iter().map(|input| input.to_ndarray(generator, ctx)).collect_vec();
let ndarray = NDArrayObject::broadcast_starmap(
generator,
ctx,
&inputs,
NDArrayOut::NewNDArray { dtype: ret_dtype },
mapping,
)?;
Ok(ScalarOrNDArray::NDArray(ndarray))
}
}
/// Map through this [`ScalarOrNDArray`] with an elementwise function.
///
/// If this is a scalar, `mapping` will directly act on the scalar. This function will return a [`ScalarOrNDArray::Scalar`] of that result.
///
/// If this is an ndarray, `mapping` will be applied to the elements of the ndarray. A new ndarray of the results will be created and
/// returned as a [`ScalarOrNDArray::NDArray`].
pub fn map<'a, G, Mapping>(
&self,
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, 'a>,
ret_dtype: Type,
mapping: Mapping,
) -> Result<ScalarOrNDArray<'ctx>, String>
where
G: CodeGenerator + ?Sized,
Mapping: FnOnce(
&mut G,
&mut CodeGenContext<'ctx, 'a>,
BasicValueEnum<'ctx>,
) -> Result<BasicValueEnum<'ctx>, String>,
{
ScalarOrNDArray::broadcasting_starmap(
generator,
ctx,
&[*self],
ret_dtype,
|generator, ctx, scalars| mapping(generator, ctx, scalars[0]),
)
}
}

51
nac3core/src/codegen/object/ndarray/mod.rs
View File

@ -13,6 +13,7 @@ use crate::{
call_nac3_ndarray_get_pelement_by_indices, call_nac3_ndarray_is_c_contiguous,
call_nac3_ndarray_len, call_nac3_ndarray_nbytes,
call_nac3_ndarray_set_strides_by_shape, call_nac3_ndarray_size,
call_nac3_ndarray_util_assert_output_shape_same,
},
model::*,
CodeGenContext, CodeGenerator,
@ -25,6 +26,7 @@ pub mod array;
pub mod broadcast;
pub mod factory;
pub mod indexing;
pub mod map;
pub mod nditer;
pub mod shape_util;
pub mod view;
@ -499,6 +501,31 @@ impl<'ctx> NDArrayObject<'ctx> {
ndarray.instance.set(ctx, |f| f.data, data);
ndarray
}
/// Check if this `NDArray` can be used as an `out` ndarray for an operation.
///
/// Raise an exception if the shapes do not match.
pub fn assert_can_be_written_by_out<G: CodeGenerator + ?Sized>(
&self,
generator: &mut G,
ctx: &mut CodeGenContext<'ctx, '_>,
out_ndims: u64,
out_shape: Instance<'ctx, Ptr<Int<SizeT>>>,
) {
let ndarray_ndims = self.ndims_llvm(generator, ctx.ctx);
let ndarray_shape = self.instance.get(generator, ctx, |f| f.shape);
let output_ndims = Int(SizeT).const_int(generator, ctx.ctx, out_ndims, false);
let output_shape = out_shape;
call_nac3_ndarray_util_assert_output_shape_same(
generator,
ctx,
ndarray_ndims,
ndarray_shape,
output_ndims,
output_shape,
);
}
}
/// A convenience enum for implementing functions that acts on scalars or ndarrays or both.
@ -584,3 +611,27 @@ impl<'ctx> ScalarOrNDArray<'ctx> {
}
}
}
/// An helper enum specifying how a function should produce its output.
///
/// Many functions in NumPy has an optional `out` parameter (e.g., `matmul`). If `out` is specified
/// with an ndarray, the result of a function will be written to `out`. If `out` is not specified, a function will
/// create a new ndarray and store the result in it.
#[derive(Debug, Clone, Copy)]
pub enum NDArrayOut<'ctx> {
/// Tell a function should create a new ndarray with the expected element type `dtype`.
NewNDArray { dtype: Type },
/// Tell a function to write the result to `ndarray`.
WriteToNDArray { ndarray: NDArrayObject<'ctx> },
}
impl<'ctx> NDArrayOut<'ctx> {
/// Get the dtype of this output.
#[must_use]
pub fn get_dtype(&self) -> Type {
match self {
NDArrayOut::NewNDArray { dtype } => *dtype,
NDArrayOut::WriteToNDArray { ndarray } => ndarray.dtype,
}
}
}