cubecl-core 0.11.0-pre.3

CubeCL core create
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145

use crate as cubecl;
use alloc::vec;
use cubecl_ir::{
    NamedRewrite, Scope,
    dialect::{
        base::OperationPtrExt,
        math::{SaturatingSAddOp, SaturatingSSubOp, SaturatingUAddOp, SaturatingUSubOp},
    },
    interfaces::TypedExt,
    prelude::*,
};
use pliron::builtin::types::{IntegerType, Signedness};

use crate::prelude::*;

define_scalar!(Elem);
define_scalar!(ElemU);
define_size!(N);

pub type LowerSaturatingArithmeticPass = MatchRewritePass<LowerSaturatingArithmetic>;

#[op_interface]
trait SaturatingOp {
    verify_op_succ!();
    fn run_polyfill(&self, args: (&Scope, Value, Value)) -> Value;
}

#[op_interface_impl]
impl SaturatingOp for SaturatingSAddOp {
    fn run_polyfill(&self, args: (&Scope, Value, Value)) -> Value {
        run_polyfill(args, saturating_add_signed::expand::<Elem, ElemU, N>)
    }
}

#[op_interface_impl]
impl SaturatingOp for SaturatingUAddOp {
    fn run_polyfill(&self, args: (&Scope, Value, Value)) -> Value {
        run_polyfill(args, saturating_add_unsigned::expand::<Elem, N>)
    }
}

#[op_interface_impl]
impl SaturatingOp for SaturatingSSubOp {
    fn run_polyfill(&self, args: (&Scope, Value, Value)) -> Value {
        run_polyfill(args, saturating_sub_signed::expand::<Elem, ElemU, N>)
    }
}

#[op_interface_impl]
impl SaturatingOp for SaturatingUSubOp {
    fn run_polyfill(&self, args: (&Scope, Value, Value)) -> Value {
        run_polyfill(args, saturating_sub_unsigned::expand::<Elem, N>)
    }
}

/// Replaces saturating arithmetic with a performant polyfill
#[derive(new, Debug, Default, NamedRewrite)]
pub struct LowerSaturatingArithmetic;

impl MatchRewrite for LowerSaturatingArithmetic {
    fn r#match(&mut self, ctx: &Context, op: Ptr<Operation>) -> bool {
        op.impls::<dyn SaturatingOp>(ctx)
    }

    fn rewrite(
        &mut self,
        ctx: &mut Context,
        rewriter: &mut DialectConversionRewriter,
        op: Ptr<Operation>,
    ) -> Result<()> {
        let scope = Scope::from_context_and_inserter(ctx, rewriter);
        let lhs = op.deref(ctx).get_operand(0);
        let rhs = op.deref(ctx).get_operand(1);

        let dyn_op = op.dyn_op(ctx);
        let sat_op = op_cast::<dyn SaturatingOp>(&*dyn_op).unwrap();
        let value = sat_op.run_polyfill((&scope, lhs, rhs));

        rewriter.replace_operation_with_values(ctx, op, vec![value]);
        Ok(())
    }
}

fn run_polyfill<T: CubePrimitive>(
    (scope, lhs, rhs): (&Scope, Value, Value),
    mut polyfill: impl FnMut(&Scope, NativeExpand<T>, NativeExpand<T>) -> NativeExpand<T>,
) -> Value {
    let ctx = scope.ctx();
    let width = lhs.scalar_ty(ctx).size_bits(ctx);
    let unsigned_ty = IntegerType::get(ctx, width as u32, Signedness::Unsigned);
    scope.register_value_type::<Elem, N>(lhs);
    scope.register_value_type::<ElemU, ()>(unsigned_ty.to_handle());

    polyfill(scope, lhs.into(), rhs.into()).value(scope)
}

#[cube]
fn saturating_add_unsigned<U: Int, N: Size>(a: Vector<U, N>, b: Vector<U, N>) -> Vector<U, N> {
    let c = a.min(!b);
    c + b
}

#[cube]
fn saturating_sub_unsigned<U: Int, N: Size>(a: Vector<U, N>, b: Vector<U, N>) -> Vector<U, N> {
    let a = a.max(b);
    a - b
}

/// Don't ask me how this works
/// <https://locklessinc.com/articles/sat_arithmetic/>
#[cube]
fn saturating_add_signed<I: Int, U: Int, N: Size>(
    x: Vector<I, N>,
    y: Vector<I, N>,
) -> Vector<I, N> {
    let bit_width = I::size_bits();
    let shift = Vector::<U, N>::new(U::new(comptime![(bit_width - 1) as i64]));

    let ux = Vector::<U, N>::cast_from(x);
    let uy = Vector::<U, N>::cast_from(y);
    let res = ux + uy;
    let ux = (ux >> shift) + Vector::<U, N>::cast_from(I::max_value());
    let zero = Vector::new(I::new(0));
    let cond = Vector::<I, N>::cast_from((ux ^ uy) | !(uy ^ res)).greater_equal(&zero);
    select_many(cond, Vector::cast_from(ux), Vector::cast_from(res))
}

/// Don't ask me how this works
/// <https://locklessinc.com/articles/sat_arithmetic/>
#[cube]
fn saturating_sub_signed<I: Int, U: Int, N: Size>(
    x: Vector<I, N>,
    y: Vector<I, N>,
) -> Vector<I, N> {
    let bit_width = I::size_bits();
    let shift = Vector::<U, N>::new(U::new(comptime![(bit_width - 1) as i64]));

    let ux = Vector::<U, N>::cast_from(x);
    let uy = Vector::<U, N>::cast_from(y);
    let res = ux - uy;
    let ux = (ux >> shift) + Vector::<U, N>::cast_from(I::max_value());
    let zero = Vector::new(I::new(0));
    let cond = Vector::<I, N>::cast_from((ux ^ uy) & (ux ^ res)).less_than(&zero);
    select_many(cond, Vector::cast_from(ux), Vector::cast_from(res))
}