sysdeps/hppa/hppa1.1/addmul_1.s

   1 ;! HP-PA-1.1 __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 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, write to
  20 ;! the Free Software Foundation, Inc., 59 Temple Place - Suite 330, Boston,
  21 ;! MA 02111-1307, USA.
  22
  23
  24 ;! INPUT PARAMETERS
  25 ;! res_ptr      r26
  26 ;! s1_ptr       r25
  27 ;! size         r24
  28 ;! s2_limb      r23
  29
  30 ;! This runs at 11 cycles/limb on a PA7000.  With the used instructions, it
  31 ;! can not become faster due to data cache contention after a store.  On the
  32 ;! PA7100 it runs at 10 cycles/limb, and that can not be improved either,
  33 ;! since only the xmpyu does not need the integer pipeline, so the only
  34 ;! dual-issue we will get are addc+xmpyu.  Unrolling could gain a cycle/limb
  35 ;! on the PA7100.
  36
  37 ;! There are some ideas described in mul_1.s that applies to this code too.
  38
  39         .text
  40         .export         __mpn_addmul_1
  41 __mpn_addmul_1:
  42         .proc
  43         .callinfo       frame=64,no_calls
  44         .entry
  45
  46         ldo             64(%r30),%r30
  47         fldws,ma        4(%r25),%fr5
  48         stw             %r23,-16(%r30)          ;! move s2_limb ...
  49         addib,=         -1,%r24,L$just_one_limb
  50          fldws          -16(%r30),%fr4          ;! ... into fr4
  51         add             %r0,%r0,%r0             ;! clear carry
  52         xmpyu           %fr4,%fr5,%fr6
  53         fldws,ma        4(%r25),%fr7
  54         fstds           %fr6,-16(%r30)
  55         xmpyu           %fr4,%fr7,%fr8
  56         ldw             -12(%r30),%r20          ;! least significant limb in product
  57         ldw             -16(%r30),%r28
  58
  59         fstds           %fr8,-16(%r30)
  60         addib,=         -1,%r24,L$end
  61          ldw            -12(%r30),%r1
  62
  63 ;! Main loop
  64 L$loop:
  65         ldws            0(%r26),%r29
  66         fldws,ma        4(%r25),%fr5
  67         add             %r29,%r20,%r20
  68         stws,ma         %r20,4(%r26)
  69         addc            %r28,%r1,%r20
  70         xmpyu           %fr4,%fr5,%fr6
  71         ldw             -16(%r30),%r28
  72         fstds           %fr6,-16(%r30)
  73         addc            %r0,%r28,%r28
  74         addib,<>        -1,%r24,L$loop
  75          ldw            -12(%r30),%r1
  76
  77 L$end:
  78         ldw             0(%r26),%r29
  79         add             %r29,%r20,%r20
  80         stws,ma         %r20,4(%r26)
  81         addc            %r28,%r1,%r20
  82         ldw             -16(%r30),%r28
  83         ldws            0(%r26),%r29
  84         addc            %r0,%r28,%r28
  85         add             %r29,%r20,%r20
  86         stws,ma         %r20,4(%r26)
  87         addc            %r0,%r28,%r28
  88         bv              0(%r2)
  89         ldo             -64(%r30),%r30
  90
  91 L$just_one_limb:
  92         xmpyu           %fr4,%fr5,%fr6
  93         ldw             0(%r26),%r29
  94         fstds           %fr6,-16(%r30)
  95         ldw             -12(%r30),%r1
  96         ldw             -16(%r30),%r28
  97         add             %r29,%r1,%r20
  98         stw             %r20,0(%r26)
  99         addc            %r0,%r28,%r28
 100         bv              0(%r2)
 101          ldo            -64(%r30),%r30
 102
 103         .exit
 104         .procend