sysdeps/sparc/sparc32/sparcv8/addmul_1.S

   1 ! SPARC v8 __mpn_addmul_1 -- Multiply a limb vector with a limb and
   2 ! add the result to a second limb vector.
   3
   4 ! Copyright (C) 1992, 1993, 1994, 1995 Free Software Foundation, Inc.
   5
   6 ! This file is part of the GNU MP Library.
   7
   8 ! The GNU MP Library is free software; you can redistribute it and/or modify
   9 ! it under the terms of the GNU Lesser General Public License as published by
  10 ! the Free Software Foundation; either version 2.1 of the License, or (at your
  11 ! option) any later version.
  12
  13 ! The GNU MP Library is distributed in the hope that it will be useful, but
  14 ! WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
  15 ! or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
  16 ! License for more details.
  17
  18 ! You should have received a copy of the GNU Lesser General Public License
  19 ! along with the GNU MP Library; see the file COPYING.LIB.  If not,
  20 ! see <http://www.gnu.org/licenses/>.
  21
  22
  23 ! INPUT PARAMETERS
  24 ! res_ptr       o0
  25 ! s1_ptr        o1
  26 ! size          o2
  27 ! s2_limb       o3
  28
  29 #include <sysdep.h>
  30
  31 ENTRY(__mpn_addmul_1)
  32         ld      [%o1+0],%o4     ! 1
  33         sll     %o2,4,%g1
  34         orcc    %g0,%g0,%g2
  35         mov     %o7,%g4                 ! Save return address register
  36         and     %g1,(4-1)<<4,%g1
  37 1:      call    2f
  38          add    %o7,3f-1b,%g3
  39 2:      jmp     %g3+%g1
  40          mov    %g4,%o7                 ! Restore return address register
  41
  42         .align  4
  43 3:
  44 LOC(00):
  45         add     %o0,-4,%o0
  46         b       LOC(loop00)             /* 4, 8, 12, ... */
  47          add    %o1,-4,%o1
  48         nop
  49 LOC(01):
  50         b       LOC(loop01)             /* 1, 5, 9, ... */
  51          nop
  52         nop
  53         nop
  54 LOC(10):
  55         add     %o0,-12,%o0     /* 2, 6, 10, ... */
  56         b       LOC(loop10)
  57          add    %o1,4,%o1
  58         nop
  59 LOC(11):
  60         add     %o0,-8,%o0      /* 3, 7, 11, ... */
  61         b       LOC(loop11)
  62          add    %o1,-8,%o1
  63         nop
  64
  65 LOC(loop):
  66         addcc   %g3,%g2,%g3     ! 1
  67         ld      [%o1+4],%o4     ! 2
  68         rd      %y,%g2          ! 1
  69         addx    %g0,%g2,%g2
  70         ld      [%o0+0],%g1     ! 2
  71         addcc   %g1,%g3,%g3
  72         st      %g3,[%o0+0]     ! 1
  73 LOC(loop00):
  74         umul    %o4,%o3,%g3     ! 2
  75         ld      [%o0+4],%g1     ! 2
  76         addxcc  %g3,%g2,%g3     ! 2
  77         ld      [%o1+8],%o4     ! 3
  78         rd      %y,%g2          ! 2
  79         addx    %g0,%g2,%g2
  80         nop
  81         addcc   %g1,%g3,%g3
  82         st      %g3,[%o0+4]     ! 2
  83 LOC(loop11):
  84         umul    %o4,%o3,%g3     ! 3
  85         addxcc  %g3,%g2,%g3     ! 3
  86         ld      [%o1+12],%o4    ! 4
  87         rd      %y,%g2          ! 3
  88         add     %o1,16,%o1
  89         addx    %g0,%g2,%g2
  90         ld      [%o0+8],%g1     ! 2
  91         addcc   %g1,%g3,%g3
  92         st      %g3,[%o0+8]     ! 3
  93 LOC(loop10):
  94         umul    %o4,%o3,%g3     ! 4
  95         addxcc  %g3,%g2,%g3     ! 4
  96         ld      [%o1+0],%o4     ! 1
  97         rd      %y,%g2          ! 4
  98         addx    %g0,%g2,%g2
  99         ld      [%o0+12],%g1    ! 2
 100         addcc   %g1,%g3,%g3
 101         st      %g3,[%o0+12]    ! 4
 102         add     %o0,16,%o0
 103         addx    %g0,%g2,%g2
 104 LOC(loop01):
 105         addcc   %o2,-4,%o2
 106         bg      LOC(loop)
 107          umul   %o4,%o3,%g3     ! 1
 108
 109         addcc   %g3,%g2,%g3     ! 4
 110         rd      %y,%g2          ! 4
 111         addx    %g0,%g2,%g2
 112         ld      [%o0+0],%g1     ! 2
 113         addcc   %g1,%g3,%g3
 114         st      %g3,[%o0+0]     ! 4
 115         retl
 116          addx   %g0,%g2,%o0
 117
 118 END(__mpn_addmul_1)