2402d,ldc一秒700万

以下可一秒800万.

module strategy;

int run(byte[] a, byte[] b) {
    int sum = 0;

    for(size_t i; i < a.length; i += 64) {
        static foreach(j; 0 .. 64) {
            sum += cast(short)a[i + j] * cast(short)b[i + j];
        }
    }

    return sum;
}
1
2
3
4
5
6
7
8
9
10
11
12
13

而如下,用ldc可700万:

module strategy;

int run(byte[] a, byte[] b) {
    assert(a.length == b.length);

    int sum = 0;

    foreach(i; 0 .. a.length) {
        sum += cast(short)a[i] * cast(short)b[i];
    }

    return sum;
}
1
2
3
4
5
6
7
8
9
10
11
12
13