nasm sse sin

qsin.asm

qsin:
  movapd xmm1, xmm0    ; xmm1 = 0:x
  mov rax, 1
  cvtsi2sd xmm2, rax
  shufpd xmm1, xmm2, 0
  mulpd xmm1, xmm1     ; xmm1 = 1:x^2
  shufpd xmm0, xmm0, 0 ; xmm0 = x:x
  mulpd xmm1, xmm0     ; xmm1 = x:x^3

  movapd xmm2, xmm0    ; xmm2 = x:x
  mulpd xmm2, xmm2     ; xmm2 = x^2:x^2
  mulpd xmm2, xmm2     ; xmm2 = x^4:x^4
  xorpd xmm0, xmm0     ; xmm0 = 0:0

  mov rax, 1
  ;pinsrq xmm3, rax, 0  ; sse4.1
  cvtsi2sd xmm3, rax
  shufpd xmm3, xmm3, 0 ; xmm3 = 1.0:1.0


  mov rax, 5
  cvtsi2sd xmm4, rax ; xmm4 = 0:5
  movapd xmm5, xmm3  ; xmm5 = 1:1
  addpd xmm5, xmm4   ; xmm5 = 1:6
  divpd xmm1, xmm5   ; xmm1 = x:x^3/6
  movapd xmm0, xmm1  ; xmm0 = x:x^3/6

  mov rax, 2
  cvtsi2sd xmm4, rax ; xmm4 = 0:2
  addpd xmm4, xmm3  ; xmm4 = 1:3

  mov rcx, 10
  qsin_loop:
    addpd xmm4, xmm3
    divpd xmm1, xmm4
    addpd xmm4, xmm3
    divpd xmm1, xmm4
    addpd xmm4, xmm3
    divpd xmm1, xmm4
    addpd xmm4, xmm3
    divpd xmm1, xmm4
    mulpd xmm1, xmm2 ; xmm1 *= x^2(n(n+1):m(m+1))
    addpd xmm0, xmm1
    dec rcx
  jge qsin_loop

  movq xmm1, xmm0
  shufpd xmm0, xmm0, 1
  subsd xmm0, xmm1
  ret