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
|
// Copyright ©2021 The Gonum Authors. All rights reserved.
// Use of this source code is governed by a BSD-style
// license that can be found in the LICENSE file.
package testlapack
import (
"fmt"
"math"
"testing"
"golang.org/x/exp/rand"
"gonum.org/v1/gonum/lapack"
)
func DlangbBenchmark(b *testing.B, impl Dlangber) {
var result float64
rnd := rand.New(rand.NewSource(1))
for _, bm := range []struct {
n, k int
}{
{n: 1000, k: 0},
{n: 1000, k: 1},
{n: 1000, k: 2},
{n: 1000, k: 5},
{n: 1000, k: 8},
{n: 1000, k: 10},
{n: 1000, k: 20},
{n: 1000, k: 30},
{n: 10000, k: 0},
{n: 10000, k: 1},
{n: 10000, k: 2},
{n: 10000, k: 5},
{n: 10000, k: 8},
{n: 10000, k: 10},
{n: 10000, k: 30},
{n: 10000, k: 60},
{n: 10000, k: 100},
} {
n := bm.n
k := bm.k
lda := 2*k + 1
aCopy := make([]float64, n*lda)
for i := range aCopy {
aCopy[i] = 1 - 2*rnd.Float64()
}
a := make([]float64, len(aCopy))
for _, norm := range []lapack.MatrixNorm{lapack.MaxAbs, lapack.MaxRowSum, lapack.MaxColumnSum} {
name := fmt.Sprintf("%v_N=%v_K=%v", normToString(norm), n, k)
b.Run(name, func(b *testing.B) {
for i := 0; i < b.N; i++ {
result = impl.Dlangb(norm, n, n, k, k, a, lda)
}
})
}
// Frobenius norm is benchmarked separately because its execution time
// depends on the element magnitude.
norm := lapack.Frobenius
for _, scale := range []string{"Small", "Medium", "Big"} {
name := fmt.Sprintf("%v_N=%v_K=%v_%v", normToString(norm), n, k, scale)
var scl float64
switch scale {
default:
scl = 1
case "Small":
scl = smlnum
case "Big":
scl = bignum
}
// Scale some elements so that the matrix contains a mix of small
// and medium, all medium, or big and medium values.
copy(a, aCopy)
for i := range a {
if i%2 == 0 {
a[i] *= scl
}
}
b.Run(name, func(b *testing.B) {
for i := 0; i < b.N; i++ {
result = impl.Dlangb(norm, n, n, k, k, a, lda)
}
})
}
}
if math.IsNaN(result) {
b.Error("unexpected NaN result")
}
}
|
