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
|
/*
sudo cpufreq-set -c 0 -g performance
mycl karatsuba.cpp -DMCL_USE_LLVM=1 ../lib/libmcl.a && ./a.out
*/
#include <stdio.h>
#include <mcl/fp.hpp>
#include <cybozu/xorshift.hpp>
#include "../src/proto.hpp"
#include "../src/low_func.hpp"
#ifdef MCL_USE_LLVM
#include "../src/low_func_llvm.hpp"
#endif
#include <cybozu/test.hpp>
#include <cybozu/benchmark.hpp>
typedef mcl::FpT<> Fp;
using namespace mcl::fp;
void dump(const Unit *x, size_t N)
{
for (size_t i = 0; i < N; i++) {
printf("%016llx ", (long long)x[N - 1 - i]);
}
printf("\n");
}
void gggKara(uint64_t *z, const uint64_t *x, const uint64_t *)
{
SqrPre<8, Gtag>::f(z, x);
}
void gggLLVM(uint64_t *z, const uint64_t *x, const uint64_t *y)
{
MulPre<8, Ltag>::f(z, x, y);
}
template<size_t N>
void benchKaratsuba()
{
cybozu::XorShift rg;
printf("N=%d\n", (int)N);
Unit z[N * 2];
rg.read(z, N);
CYBOZU_BENCH("g:mulPre ", (MulPreCore<N, Gtag>::f), z, z, z);
// CYBOZU_BENCH("g:mulKara", (MulPre<N, Gtag>::karatsuba), z, z, z);
CYBOZU_BENCH("g:sqrPre ", (SqrPreCore<N, Gtag>::f), z, z);
// CYBOZU_BENCH("g:sqrKara", (SqrPre<N, Gtag>::karatsuba), z, z);
#ifdef MCL_USE_LLVM
CYBOZU_BENCH("l:mulPre ", (MulPreCore<N, Ltag>::f), z, z, z);
CYBOZU_BENCH("l:sqrPre ", (SqrPreCore<N, Ltag>::f), z, z);
CYBOZU_BENCH("l:mulKara", (MulPre<N, Ltag>::karatsuba), z, z, z);
CYBOZU_BENCH("l:sqrKara", (SqrPre<N, Ltag>::karatsuba), z, z);
#endif
}
CYBOZU_TEST_AUTO(karatsuba)
{
benchKaratsuba<4>();
benchKaratsuba<6>();
benchKaratsuba<8>();
#if MCL_MAX_BIT_SIZE >= 640
benchKaratsuba<10>();
#endif
#if MCL_MAX_BIT_SIZE >= 768
benchKaratsuba<12>();
#endif
#if MCL_MAX_BIT_SIZE >= 896
benchKaratsuba<14>();
#endif
#if MCL_MAX_BIT_SIZE >= 1024
benchKaratsuba<16>();
#endif
}
|