Return to add_n.asm CVS log

Up to [local] / OpenXM_contrib / gmp / mpn / x86 / pentium4 / sse2

Annotation of OpenXM_contrib/gmp/mpn/x86/pentium4/sse2/add_n.asm, Revision 1.1.1.1

1.1       ohara       1: dnl  Intel Pentium-4 mpn_add_n -- mpn addition.
                      2:
                      3: dnl  Copyright 2001, 2002 Free Software Foundation, Inc.
                      4: dnl
                      5: dnl  This file is part of the GNU MP Library.
                      6: dnl
                      7: dnl  The GNU MP Library is free software; you can redistribute it and/or
                      8: dnl  modify it under the terms of the GNU Lesser General Public License as
                      9: dnl  published by the Free Software Foundation; either version 2.1 of the
                     10: dnl  License, or (at your option) any later version.
                     11: dnl
                     12: dnl  The GNU MP Library is distributed in the hope that it will be useful,
                     13: dnl  but WITHOUT ANY WARRANTY; without even the implied warranty of
                     14: dnl  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
                     15: dnl  Lesser General Public License for more details.
                     16: dnl
                     17: dnl  You should have received a copy of the GNU Lesser General Public
                     18: dnl  License along with the GNU MP Library; see the file COPYING.LIB.  If
                     19: dnl  not, write to the Free Software Foundation, Inc., 59 Temple Place -
                     20: dnl  Suite 330, Boston, MA 02111-1307, USA.
                     21:
                     22: include(`../config.m4')
                     23:
                     24:
                     25: C P4: 4.0 cycles/limb if dst!=src1 and dst!=src2
                     26: C     6.0 cycles/limb if dst==src1 or dst==src2
                     27:
                     28:
                     29: C mp_limb_t mpn_add_n (mp_ptr dst, mp_srcptr src1, mp_srcptr src2,
                     30: C                      mp_size_t size);
                     31: C mp_limb_t mpn_add_nc (mp_ptr dst, mp_srcptr src1, mp_srcptr src2,
                     32: C                       mp_size_t size, mp_limb_t carry);
                     33: C
                     34: C The 4 c/l achieved here isn't particularly good, but is better than 9 c/l
                     35: C for a basic adc loop.
                     36:
                     37: defframe(PARAM_CARRY,20)
                     38: defframe(PARAM_SIZE, 16)
                     39: defframe(PARAM_SRC2, 12)
                     40: defframe(PARAM_SRC1, 8)
                     41: defframe(PARAM_DST,  4)
                     42:
                     43: dnl  re-use parameter space
                     44: define(SAVE_EBX,`PARAM_SRC1')
                     45:
                     46:        TEXT
                     47:        ALIGN(8)
                     48:
                     49: PROLOGUE(mpn_add_nc)
                     50: deflit(`FRAME',0)
                     51:
                     52:        movd    PARAM_CARRY, %mm0
                     53:        jmp     L(start_nc)
                     54:
                     55: EPILOGUE()
                     56:
                     57:        ALIGN(8)
                     58: PROLOGUE(mpn_add_n)
                     59: deflit(`FRAME',0)
                     60:
                     61:        pxor    %mm0, %mm0
                     62:
                     63: L(start_nc):
                     64:        movl    PARAM_SRC1, %eax
                     65:        movl    %ebx, SAVE_EBX
                     66:        movl    PARAM_SRC2, %ebx
                     67:        movl    PARAM_DST, %edx
                     68:        movl    PARAM_SIZE, %ecx
                     69:
                     70:        leal    (%eax,%ecx,4), %eax     C src1 end
                     71:        leal    (%ebx,%ecx,4), %ebx     C src2 end
                     72:        leal    (%edx,%ecx,4), %edx     C dst end
                     73:        negl    %ecx                    C -size
                     74:
                     75: L(top):
                     76:        C eax   src1 end
                     77:        C ebx   src2 end
                     78:        C ecx   counter, limbs, negative
                     79:        C edx   dst end
                     80:        C mm0   carry bit
                     81:
                     82:        movd    (%eax,%ecx,4), %mm1
                     83:        movd    (%ebx,%ecx,4), %mm2
                     84:        paddq   %mm2, %mm1
                     85:
                     86:        paddq   %mm1, %mm0
                     87:        movd    %mm0, (%edx,%ecx,4)
                     88:
                     89:        psrlq   $32, %mm0
                     90:
                     91:        addl    $1, %ecx
                     92:        jnz     L(top)
                     93:
                     94:
                     95:        movd    %mm0, %eax
                     96:        movl    SAVE_EBX, %ebx
                     97:        emms
                     98:        ret
                     99:
                    100: EPILOGUE()