CS:APP - 기계어, 부동소수점

float float_mov(float v1, float *src, float *dst) {
    float v2 = *src;
    *dst = v1;
    return v2;
}

float_mov:
    movss   (%rdi), %xmm1
    movss   %xmm0, (%rsi)
    movaps  %xmm1, %xmm0
    ret

float_mov:
    vmovss  (%rdi), %xmm1    ; Read v2 from src
    vmovss  %xmm0, (%rsi)    ; Write v1 to dst
    vmovaps %xmm1, %xmm0
    ret

double fcvt(int i, float *fp, double *dp, long *lp) {
    float  f = *fp;
    double d = *dp;
    long   l = *lp;

    *lp = (long)    d;
    *fp = (float)   i;
    *dp = (double)  l;
    return (double) f;
}

fcvt:
    vmovss      (%rsi), %xmm0          ; Get f = *fp
    movq        (%rcx), %rax           ; Get l = *lp
    vcvttsd2siq (%rdx), %r8            ; Get d = *dp and convert to long
    movq        %r8, (%rcx)            ; Store at lp
    vxorps      %xmm1, %xmm1, %xmm1
    vcvtsi2ss   %edi, %xmm1, %xmm1     ; Convert i to float
    vmovss      %xmm1, (%rsi)          ; Store at fp
    vxorpd      %xmm1, %xmm1, %xmm1
    vcvtsi2sdq  %rax, %xmm1, %xmm1     ; Convert l to double
    vmovsd      %xmm1, (%rdx)          ; Store at dp
    vcvtss2sd   %xmm0, %xmm0, %xmm0    ; Convert f to double
    ret

double funct(double a, float x, double b, int i) {
    return a*x - b/i;
}

funct:
    vcvtss2sd   %xmm1, %xmm1, %xmm1    ; Convert x to double
    vmulsd      %xmm0, %xmm1, %xmm0    ; Multiply a by x
    vxorpd      %xmm1, %xmm1, %xmm1
    vcvtsi2sd   %edi, %xmm1, %xmm1     ; Convert i to double
    vdivsd      %xmm1, %xmm2, %xmm2    ; Compute b/i
    vsubsd      %xmm2, %xmm0, %xmm0    ; Subtract from a*x
    ret

double cel2fahr(double temp) {
    return 1.8 * temp + 32.0;
}

cel2fahr:
.LFB0:
    .cfi_startproc
    vmulsd  .LC0(%rip), %xmm0, %xmm0
    vaddsd  .LC1(%rip), %xmm0, %xmm0
    ret
    .cfi_endproc
.LFE0:
    .size   cel2fahr, .-cel2fahr
    .section        .rodata.cst8,"aM",@progbits,8
    .align 8
.LC0:
    .long   3435973837    ; Low-order 4 bytes of 1.8
    .long   1073532108    ; High-order 4 bytes of 1.8
    .align 8
.LC1:
    .long   0             ; Low-order 4 bytes of 32.0
    .long   1077936128    ; Low-order 4 bytes of 32.0

double zero(void) {
    return 0.0;
}

double negative(double x) {
    return -x;
}

#include <math.h>
double absolute(double x) {
    return fabs(x);
}

zero:
    vxorpd  %xmm0, %xmm0, %xmm0
    ret
negative:
    vxorpd  .LC1(%rip), %xmm0, %xmm0
    ret
absolute:
    vandpd  .LC2(%rip), %xmm0, %xmm0
    ret
.LC1:
    .long   0
    .long   -2147483648
    .long   0
    .long   0
.LC2:
    .long   4294967295
    .long   2147483647
    .long   0
    .long   0

typedef enum { NEG, ZERO, POS, OTHER } range_t;

range_t find_range(float x) {
    int result;
    if (x < 0)
        result = NEG;
    else if (x == 0)
        result = ZERO;
    else if (x > 0)
        result = POS;
    else
        result = OTHER;
    return result;
}

find_range:
    vxorps      %xmm1, %xmm1, %xmm1
    vucomiss    %xmm0, %xmm1    ; Compare 0:x
    ja      .L5
    vucomiss    %xmm1, %xmm0    ; Compare x:0
    jp      .L8
    movl    $1, %eax      ; result = ZERO
    je      .L10
.L8:
    xorl    %eax, %eax
    movl    $3, %edx
    vucomiss    %xmm1, %xmm0    ; Compare x:0
    seta    %al
    subl    %eax, %edx
    movl    %edx, %eax    ; result = (~CF & ~ZF) ? POS : OTHER
    ret
.L5:
    xorl    %eax, %eax    ; result = NEG
.L10:
    rep ret