github.com/aergoio/aergo@v1.3.1/libtool/src/gmp-6.1.2/mpn/x86_64/pentium4/aors_n.asm (about)

     1  dnl  x86-64 mpn_add_n/mpn_sub_n optimized for Pentium 4.
     2  
     3  dnl  Contributed to the GNU project by Torbjorn Granlund.
     4  
     5  dnl  Copyright 2007, 2008, 2010-2012 Free Software Foundation, Inc.
     6  
     7  dnl  This file is part of the GNU MP Library.
     8  dnl
     9  dnl  The GNU MP Library is free software; you can redistribute it and/or modify
    10  dnl  it under the terms of either:
    11  dnl
    12  dnl    * the GNU Lesser General Public License as published by the Free
    13  dnl      Software Foundation; either version 3 of the License, or (at your
    14  dnl      option) any later version.
    15  dnl
    16  dnl  or
    17  dnl
    18  dnl    * the GNU General Public License as published by the Free Software
    19  dnl      Foundation; either version 2 of the License, or (at your option) any
    20  dnl      later version.
    21  dnl
    22  dnl  or both in parallel, as here.
    23  dnl
    24  dnl  The GNU MP Library is distributed in the hope that it will be useful, but
    25  dnl  WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
    26  dnl  or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
    27  dnl  for more details.
    28  dnl
    29  dnl  You should have received copies of the GNU General Public License and the
    30  dnl  GNU Lesser General Public License along with the GNU MP Library.  If not,
    31  dnl  see https://www.gnu.org/licenses/.
    32  
    33  include(`../config.m4')
    34  
    35  
    36  C	     cycles/limb
    37  C AMD K8,K9	 2.8
    38  C AMD K10	 2.8
    39  C Intel P4	 4
    40  C Intel core2	 3.6-5	(fluctuating)
    41  C Intel corei	 ?
    42  C Intel atom	 ?
    43  C VIA nano	 ?
    44  
    45  
    46  C INPUT PARAMETERS
    47  define(`rp',	`%rdi')
    48  define(`up',	`%rsi')
    49  define(`vp',	`%rdx')
    50  define(`n',	`%rcx')
    51  define(`cy',	`%r8')
    52  
    53  ifdef(`OPERATION_add_n', `
    54  	define(ADDSUB,	      add)
    55  	define(func,	      mpn_add_n)
    56  	define(func_nc,	      mpn_add_nc)')
    57  ifdef(`OPERATION_sub_n', `
    58  	define(ADDSUB,	      sub)
    59  	define(func,	      mpn_sub_n)
    60  	define(func_nc,	      mpn_sub_nc)')
    61  
    62  ABI_SUPPORT(DOS64)
    63  ABI_SUPPORT(STD64)
    64  
    65  MULFUNC_PROLOGUE(mpn_add_n mpn_add_nc mpn_sub_n mpn_sub_nc)
    66  ASM_START()
    67  	TEXT
    68  PROLOGUE(func)
    69  	FUNC_ENTRY(4)
    70  	xor	%r8, %r8
    71  IFDOS(`	jmp	L(ent)		')
    72  EPILOGUE()
    73  PROLOGUE(func_nc)
    74  	FUNC_ENTRY(4)
    75  IFDOS(`	mov	56(%rsp), %r8	')
    76  L(ent):	push	%rbx
    77  	push	%r12
    78  
    79  	mov	(vp), %r9
    80  
    81  	mov	R32(n), R32(%rax)
    82  	and	$3, R32(%rax)
    83  	jne	L(n00)		C n = 0, 4, 8, ...
    84  	mov	R32(%r8), R32(%rbx)
    85  	mov	(up), %r8
    86  	mov	8(up), %r10
    87  	ADDSUB	%r9, %r8
    88  	mov	8(vp), %r9
    89  	setc	R8(%rax)
    90  	lea	-16(rp), rp
    91  	jmp	L(L00)
    92  
    93  L(n00):	cmp	$2, R32(%rax)
    94  	jnc	L(n01)		C n = 1, 5, 9, ...
    95  	mov	(up), %r11
    96  	mov	R32(%r8), R32(%rax)
    97  	xor	R32(%rbx), R32(%rbx)
    98  	dec	n
    99  	jnz	L(gt1)
   100  	ADDSUB	%r9, %r11
   101  	setc	R8(%rbx)
   102  	ADDSUB	%rax, %r11
   103  	adc	$0, R32(%rbx)
   104  	mov	%r11, (rp)
   105  	jmp	L(ret)
   106  L(gt1):	mov	8(up), %r8
   107  	ADDSUB	%r9, %r11
   108  	mov	8(vp), %r9
   109  	setc	R8(%rbx)
   110  	lea	-8(rp), rp
   111  	lea	8(up), up
   112  	lea	8(vp), vp
   113  	jmp	L(L01)
   114  
   115  L(n01):	jne	L(n10)		C n = 2, 6, 10, ...
   116  	mov	(up), %r12
   117  	mov	R32(%r8), R32(%rbx)
   118  	mov	8(up), %r11
   119  	ADDSUB	%r9, %r12
   120  	mov	8(vp), %r9
   121  	setc	R8(%rax)
   122  	lea	-32(rp), rp
   123  	lea	16(up), up
   124  	lea	16(vp), vp
   125  	jmp	L(L10)
   126  
   127  L(n10):	mov	(up), %r10	C n = 3, 7, 11, ...
   128  	mov	R32(%r8), R32(%rax)
   129  	xor	R32(%rbx), R32(%rbx)
   130  	mov	8(up), %r12
   131  	ADDSUB	%r9, %r10
   132  	mov	8(vp), %r9
   133  	setc	R8(%rbx)
   134  	lea	-24(rp), rp
   135  	lea	-8(up), up
   136  	lea	-8(vp), vp
   137  	jmp	L(L11)
   138  
   139  L(c0):	mov	$1, R8(%rbx)
   140  	jmp	L(rc0)
   141  L(c1):	mov	$1, R8(%rax)
   142  	jmp	L(rc1)
   143  L(c2):	mov	$1, R8(%rbx)
   144  	jmp	L(rc2)
   145  L(c3):	mov	$1, R8(%rax)
   146  	jmp	L(rc3)
   147  
   148  	ALIGN(16)
   149  L(top):	mov	(up), %r8	C not on critical path
   150  	ADDSUB	%r9, %r11	C not on critical path
   151  	mov	(vp), %r9	C not on critical path
   152  	setc	R8(%rbx)	C save carry out
   153  	mov	%r12, (rp)
   154  L(L01):	ADDSUB	%rax, %r11	C apply previous carry out
   155  	jc	L(c0)		C jump if ripple
   156  L(rc0):	mov	8(up), %r10
   157  	ADDSUB	%r9, %r8
   158  	mov	8(vp), %r9
   159  	setc	R8(%rax)
   160  	mov	%r11, 8(rp)
   161  L(L00):	ADDSUB	%rbx, %r8
   162  	jc	L(c1)
   163  L(rc1):	mov	16(up), %r12
   164  	ADDSUB	%r9, %r10
   165  	mov	16(vp), %r9
   166  	setc	R8(%rbx)
   167  	mov	%r8, 16(rp)
   168  L(L11):	ADDSUB	%rax, %r10
   169  	jc	L(c2)
   170  L(rc2):	mov	24(up), %r11
   171  	ADDSUB	%r9, %r12
   172  	lea	32(up), up
   173  	mov	24(vp), %r9
   174  	lea	32(vp), vp
   175  	setc	R8(%rax)
   176  	mov	%r10, 24(rp)
   177  L(L10):	ADDSUB	%rbx, %r12
   178  	jc	L(c3)
   179  L(rc3):	lea	32(rp), rp
   180  	sub	$4, n
   181  	ja	L(top)
   182  
   183  L(end):	ADDSUB	%r9, %r11
   184  	setc	R8(%rbx)
   185  	mov	%r12, (rp)
   186  	ADDSUB	%rax, %r11
   187  	jnc	L(1)
   188  	mov	$1, R8(%rbx)
   189  L(1):	mov	%r11, 8(rp)
   190  
   191  L(ret):	mov	R32(%rbx), R32(%rax)
   192  	pop	%r12
   193  	pop	%rbx
   194  	FUNC_EXIT()
   195  	ret
   196  EPILOGUE()