fast wc -l

wcl.c

#include <stdio.h>
#include <stdlib.h>
#include <string.h>

#include <unistd.h>
#include <fcntl.h>
#include <sys/stat.h>
#include <sys/types.h>

#define BUF_SIZE 1048576

extern unsigned long memchrcount(void *mem, int c, size_t size);

int main(int argc, char **argv)
{
    unsigned long count = 0;
    ssize_t bytes_read;
    void *buf;
    int fd = open(argv[1], O_RDONLY);
    if (fd < 0)
    {
        perror("Cannot open file");
        return 1;
    }
    if (posix_memalign(&buf, 16, BUF_SIZE))
    {
        fputs("Cannot allocate memory\n", stderr);
        close(fd);
        return 1;
    }
    memset(buf, 0, BUF_SIZE);
    while ((bytes_read = read(fd, buf, BUF_SIZE)) > 0)
    {
        count += memchrcount(buf, '\n', BUF_SIZE);
    }
    close(fd);
    free(buf);
    printf("%lu\n", count);
    return 0;
}

wcl.s

.text

.global memchrcount
.type memchrcount, @function
memchrcount:
    mov %esi, %eax
    sal $8, %esi
    or %esi, %eax
    sal $8, %esi
    or %esi, %eax
    sal $8, %esi
    or %esi, %eax
    movd %eax, %xmm1
    movsldup %xmm1, %xmm1
    movddup %xmm1, %xmm1
    xor %eax, %eax
    xor %r8, %r8
    jmp .L2
.p2align 5
.L1:
    movaps (%rdi, %r8), %xmm0
    pcmpeqb %xmm1, %xmm0
    pmovmskb %xmm0, %ecx
    popcnt %ecx, %ecx
    add %rcx, %rax
    add $16, %r8
.L2:
    cmp %r8, %rdx
    jne .L1
    ret

make -B memchrcount.o
as -o memchrcount.o memchrcount.s
CXXFLAGS=memchrcount.o make -B wcl
cc memchrcount.o -O3 -march=native -I/home/lvv/p wcl.c -o wcl
./wcl /tmp/l
Illegal instruction
uname -a
Linux ahp 3.0.0 #9 SMP PREEMPT Tue Jul 26 10:23:31 EEST 2011 x86_64 Intel(R) Core(TM)2 Duo CPU T7500 @ 2.20GHz GenuineIntel GNU/Linux