source/libs/gmp/gmp-src/mpn/pa32/hppa1_1/pa7100/rshift.asm

   1 dnl  HP-PA  mpn_rshift -- Shift a number right.
   2 dnl  Optimized for the PA7100, where is runs at 3.25 cycles/limb.
   3
   4 dnl  Copyright 1992, 1994, 2000-2003 Free Software Foundation, Inc.
   5
   6 dnl  This file is part of the GNU MP Library.
   7 dnl
   8 dnl  The GNU MP Library is free software; you can redistribute it and/or modify
   9 dnl  it under the terms of either:
  10 dnl
  11 dnl    * the GNU Lesser General Public License as published by the Free
  12 dnl      Software Foundation; either version 3 of the License, or (at your
  13 dnl      option) any later version.
  14 dnl
  15 dnl  or
  16 dnl
  17 dnl    * the GNU General Public License as published by the Free Software
  18 dnl      Foundation; either version 2 of the License, or (at your option) any
  19 dnl      later version.
  20 dnl
  21 dnl  or both in parallel, as here.
  22 dnl
  23 dnl  The GNU MP Library is distributed in the hope that it will be useful, but
  24 dnl  WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
  25 dnl  or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  26 dnl  for more details.
  27 dnl
  28 dnl  You should have received copies of the GNU General Public License and the
  29 dnl  GNU Lesser General Public License along with the GNU MP Library.  If not,
  30 dnl  see https://www.gnu.org/licenses/.
  31
  32 include(`../config.m4')
  33
  34 C INPUT PARAMETERS
  35 C res_ptr       r26
  36 C s_ptr         r25
  37 C size          r24
  38 C cnt           r23
  39
  40 ASM_START()
  41 PROLOGUE(mpn_rshift)
  42         ldws,ma         4(0,%r25),%r22
  43         mtsar           %r23
  44         addib,=         -1,%r24,L(0004)
  45         vshd            %r22,%r0,%r28           C compute carry out limb
  46         ldws,ma         4(0,%r25),%r29
  47         addib,<=        -5,%r24,L(rest)
  48         vshd            %r29,%r22,%r20
  49
  50 LDEF(loop)
  51         ldws,ma         4(0,%r25),%r22
  52         stws,ma         %r20,4(0,%r26)
  53         vshd            %r22,%r29,%r20
  54         ldws,ma         4(0,%r25),%r29
  55         stws,ma         %r20,4(0,%r26)
  56         vshd            %r29,%r22,%r20
  57         ldws,ma         4(0,%r25),%r22
  58         stws,ma         %r20,4(0,%r26)
  59         vshd            %r22,%r29,%r20
  60         ldws,ma         4(0,%r25),%r29
  61         stws,ma         %r20,4(0,%r26)
  62         addib,>         -4,%r24,L(loop)
  63         vshd            %r29,%r22,%r20
  64
  65 LDEF(rest)
  66         addib,=         4,%r24,L(end1)
  67         nop
  68
  69 LDEF(eloop)
  70         ldws,ma         4(0,%r25),%r22
  71         stws,ma         %r20,4(0,%r26)
  72         addib,<=        -1,%r24,L(end2)
  73         vshd            %r22,%r29,%r20
  74         ldws,ma         4(0,%r25),%r29
  75         stws,ma         %r20,4(0,%r26)
  76         addib,>         -1,%r24,L(eloop)
  77         vshd            %r29,%r22,%r20
  78
  79 LDEF(end1)
  80         stws,ma         %r20,4(0,%r26)
  81         vshd            %r0,%r29,%r20
  82         bv              0(%r2)
  83         stw             %r20,0(0,%r26)
  84
  85 LDEF(end2)
  86         stws,ma         %r20,4(0,%r26)
  87
  88 LDEF(0004)
  89         vshd            %r0,%r22,%r20
  90         bv              0(%r2)
  91         stw             %r20,0(0,%r26)
  92 EPILOGUE()