gemm 0.19.0

Playground for matrix multiplication algorithms
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

use aligned_vec::{avec, AVec};
use diol::prelude::*;
use gemm::*;
use num_traits::One;

#[derive(Copy, Clone, Debug, PartialEq, Eq)]
enum Layout {
    Col,
    Row,
}

fn make_data<T: Copy + One>(
    layout: Layout,
    m: usize,
    n: usize,
    reg: usize,
) -> (isize, isize, AVec<T>) {
    let val = T::one();
    match layout {
        Layout::Col => (
            1,
            m.next_multiple_of(reg) as isize,
            avec![val; n * m.next_multiple_of(reg)],
        ),
        Layout::Row => (
            n.next_multiple_of(reg) as isize,
            1,
            avec![val; m * n.next_multiple_of(reg)],
        ),
    }
}

fn bench_gemm<T: One + Copy + 'static>(
    bencher: Bencher,
    list![par, dst, lhs, rhs, m, n, k]: List![
        Parallelism,
        Layout,
        Layout,
        Layout,
        usize,
        usize,
        usize
    ],
) {
    let reg = 64 / core::mem::size_of::<T>();

    let (dst_rs, dst_cs, mut dst) = make_data::<T>(dst, m, n, reg);
    let (lhs_rs, lhs_cs, mut lhs) = make_data::<T>(lhs, m, k, reg);
    let (rhs_rs, rhs_cs, mut rhs) = make_data::<T>(rhs, k, n, reg);

    lhs.fill(unsafe { core::mem::zeroed() });
    rhs.fill(unsafe { core::mem::zeroed() });
    dst.fill(unsafe { core::mem::zeroed() });

    bencher.bench(|| {
        unsafe {
            gemm(
                m,
                n,
                k,
                dst.as_mut_ptr(),
                dst_cs,
                dst_rs,
                true,
                lhs.as_ptr(),
                lhs_cs,
                lhs_rs,
                rhs.as_ptr(),
                rhs_cs,
                rhs_rs,
                T::one(),
                T::one(),
                false,
                false,
                false,
                par,
            )
        };
    })
}

fn args() -> Vec<List![Parallelism, Layout, Layout, Layout, usize, usize, usize]> {
    use itertools::Itertools;
    let pow2 = |i| 1usize << i;
    let halfway = |i| 3usize << (i - 1);
    itertools::iproduct!(
        [].into_iter()
            .chain((5..13).map(pow2).map(|n| (n, n, n)))
            .chain((5..13).map(halfway).map(|n| (n, n, n)))
            .chain((5..13).map(halfway).map(|n| (16, 16, n)))
            .sorted_unstable(),
        [Parallelism::Rayon(0), Parallelism::None],
        [Layout::Col, Layout::Row],
        [Layout::Col, Layout::Row],
        [Layout::Col, Layout::Row]
    )
    .map(|((m, n, k), par, dst, lhs, rhs)| list![par, dst, lhs, rhs, m, n, k])
    .collect()
}

fn main() -> std::io::Result<()> {
    let config = BenchConfig::from_args()?;

    gemm::set_wasm_simd128(true);

    let modifiers = [1];

    {
        let mut bench = Bench::new(&config);
        bench.register(bench_gemm::<f32>, args());

        for modifier in modifiers {
            gemm::set_threading_threshold(gemm::DEFAULT_THREADING_THRESHOLD / modifier);
            bench.run().unwrap();
        }
    }
    {
        let mut bench = Bench::new(&config);
        bench.register(bench_gemm::<f64>, args());
        for modifier in modifiers {
            gemm::set_threading_threshold(gemm::DEFAULT_THREADING_THRESHOLD / modifier);
            bench.run().unwrap();
        }
    }
    {
        let mut bench = Bench::new(&config);
        bench.register(bench_gemm::<c32>, args());
        for modifier in modifiers {
            gemm::set_threading_threshold(gemm::DEFAULT_THREADING_THRESHOLD / modifier);
            bench.run().unwrap();
        }
    }
    {
        let mut bench = Bench::new(&config);
        bench.register(bench_gemm::<c64>, args());
        for modifier in modifiers {
            gemm::set_threading_threshold(gemm::DEFAULT_THREADING_THRESHOLD / modifier);
            bench.run().unwrap();
        }
    }
    Ok(())
}