mpir/mpn/ia64/divexact_1.asm

218 lines
6.2 KiB
NASM
Raw Normal View History

2010-03-07 19:26:51 -05:00
dnl IA-64 mpn_divexact_1 -- mpn by limb exact division.
dnl Copyright 2003, 2004, 2005 Free Software Foundation, Inc.
dnl This file is part of the GNU MP Library.
dnl The GNU MP Library is free software; you can redistribute it and/or modify
dnl it under the terms of the GNU Lesser General Public License as published
dnl by the Free Software Foundation; either version 3 of the License, or (at
dnl your option) any later version.
dnl The GNU MP Library is distributed in the hope that it will be useful, but
dnl WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
dnl or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
dnl License for more details.
dnl You should have received a copy of the GNU Lesser General Public License
dnl along with the GNU MP Library. If not, see http://www.gnu.org/licenses/.
include(`../config.m4')
C cycles/limb
C Itanium: 16
C Itanium 2: 8
C INPUT PARAMETERS
define(`rp', `r32')
define(`up', `r33')
define(`n', `r34')
define(`divisor', `r35')
define(`lshift', `r24')
define(`rshift', `r25')
C This code is a bit messy, and not as similar to mode1o.asm as desired.
C The critical path during initialization is for computing the inverse of the
C divisor. Since odd divisors are probably common, we conditionally execute
C the initial count_traling_zeros code and the downshift.
C Possible improvement: Merge more of the feed-in code into the inverse
C computation.
ASM_START()
.text
.align 32
.Ltab:
data1 0,0x01, 0,0xAB, 0,0xCD, 0,0xB7, 0,0x39, 0,0xA3, 0,0xC5, 0,0xEF
data1 0,0xF1, 0,0x1B, 0,0x3D, 0,0xA7, 0,0x29, 0,0x13, 0,0x35, 0,0xDF
data1 0,0xE1, 0,0x8B, 0,0xAD, 0,0x97, 0,0x19, 0,0x83, 0,0xA5, 0,0xCF
data1 0,0xD1, 0,0xFB, 0,0x1D, 0,0x87, 0,0x09, 0,0xF3, 0,0x15, 0,0xBF
data1 0,0xC1, 0,0x6B, 0,0x8D, 0,0x77, 0,0xF9, 0,0x63, 0,0x85, 0,0xAF
data1 0,0xB1, 0,0xDB, 0,0xFD, 0,0x67, 0,0xE9, 0,0xD3, 0,0xF5, 0,0x9F
data1 0,0xA1, 0,0x4B, 0,0x6D, 0,0x57, 0,0xD9, 0,0x43, 0,0x65, 0,0x8F
data1 0,0x91, 0,0xBB, 0,0xDD, 0,0x47, 0,0xC9, 0,0xB3, 0,0xD5, 0,0x7F
data1 0,0x81, 0,0x2B, 0,0x4D, 0,0x37, 0,0xB9, 0,0x23, 0,0x45, 0,0x6F
data1 0,0x71, 0,0x9B, 0,0xBD, 0,0x27, 0,0xA9, 0,0x93, 0,0xB5, 0,0x5F
data1 0,0x61, 0,0x0B, 0,0x2D, 0,0x17, 0,0x99, 0,0x03, 0,0x25, 0,0x4F
data1 0,0x51, 0,0x7B, 0,0x9D, 0,0x07, 0,0x89, 0,0x73, 0,0x95, 0,0x3F
data1 0,0x41, 0,0xEB, 0,0x0D, 0,0xF7, 0,0x79, 0,0xE3, 0,0x05, 0,0x2F
data1 0,0x31, 0,0x5B, 0,0x7D, 0,0xE7, 0,0x69, 0,0x53, 0,0x75, 0,0x1F
data1 0,0x21, 0,0xCB, 0,0xED, 0,0xD7, 0,0x59, 0,0xC3, 0,0xE5, 0,0x0F
data1 0,0x11, 0,0x3B, 0,0x5D, 0,0xC7, 0,0x49, 0,0x33, 0,0x55, 0,0xFF
PROLOGUE(mpn_divexact_1)
.prologue
.save ar.lc, r2
.body
{.mmi; add r8 = -1, divisor C M0
nop 0 C M1
tbit.z p8, p9 = divisor, 0 C I0
}
ifdef(`HAVE_ABI_32',
` addp4 rp = 0, rp C M2 rp extend
addp4 up = 0, up C M3 up extend
sxt4 n = n') C I1 size extend
;;
.Lhere:
{.mmi; ld8 r20 = [up], 8 C M0 up[0]
(p8) andcm r8 = r8, divisor C M1
mov r15 = ip C I0 .Lhere
;;
}{.mii
.pred.rel "mutex", p8, p9
(p9) mov rshift = 0 C M0
(p8) popcnt rshift = r8 C I0 r8 = cnt_lo_zeros(divisor)
cmp.eq p6, p10 = 1, n C I1
;;
}{.mii; add r9 = .Ltab-.Lhere, r15 C M0
(p8) shr.u divisor = divisor, rshift C I0
nop 0 C I1
;;
}{.mmi; add n = -4, n C M0 size-1
(p10) ld8 r21 = [up], 8 C M1 up[1]
mov r14 = 2 C M1 2
}{.mfi; setf.sig f6 = divisor C M2 divisor
mov f9 = f0 C M3 carry FIXME
zxt1 r3 = divisor C I1 divisor low byte
;;
}{.mmi; add r3 = r9, r3 C M0 table offset ip and index
sub r16 = 0, divisor C M1 -divisor
mov r2 = ar.lc C I0
}{.mmi; sub lshift = 64, rshift C M2
setf.sig f13 = r14 C M3 2 in significand
mov r17 = -1 C I1 -1
;;
}{.mmi; ld1 r3 = [r3] C M0 inverse, 8 bits
nop 0 C M1
mov ar.lc = n C I0 size-1 loop count
}{.mmi; setf.sig f12 = r16 C M2 -divisor
setf.sig f8 = r17 C M3 -1
cmp.eq p7, p0 = -2, n C I1
;;
}{.mmi; setf.sig f7 = r3 C M2 inverse, 8 bits
cmp.eq p8, p0 = -1, n C M0
shr.u r23 = r20, rshift C I0
;;
}
C f6 divisor
C f7 inverse, being calculated
C f8 -1, will be -inverse
C f9 carry
C f12 -divisor
C f13 2
C f14 scratch
xmpy.l f14 = f13, f7 C Newton 2*i
xmpy.l f7 = f7, f7 C Newton i*i
;;
xma.l f7 = f7, f12, f14 C Newton i*i*-d + 2*i, 16 bits
;;
setf.sig f10 = r23 C speculative, used iff n = 1
xmpy.l f14 = f13, f7 C Newton 2*i
shl r22 = r21, lshift C speculative, used iff n > 1
xmpy.l f7 = f7, f7 C Newton i*i
;;
or r31 = r22, r23 C speculative, used iff n > 1
xma.l f7 = f7, f12, f14 C Newton i*i*-d + 2*i, 32 bits
shr.u r23 = r21, rshift C speculative, used iff n > 1
;;
setf.sig f11 = r31 C speculative, used iff n > 1
xmpy.l f14 = f13, f7 C Newton 2*i
xmpy.l f7 = f7, f7 C Newton i*i
;;
xma.l f7 = f7, f12, f14 C Newton i*i*-d + 2*i, 64 bits
(p7) br.cond.dptk .Ln2
(p10) br.cond.dptk .grt3
;;
.Ln1: xmpy.l f12 = f10, f7 C q = ulimb * inverse
br .Lx1
.Ln2:
xmpy.l f8 = f7, f8 C -inverse = inverse * -1
xmpy.l f12 = f11, f7 C q = ulimb * inverse
setf.sig f11 = r23
br .Lx2
.grt3:
ld8 r21 = [up], 8 C up[2]
xmpy.l f8 = f7, f8 C -inverse = inverse * -1
;;
shl r22 = r21, lshift
;;
xmpy.l f12 = f11, f7 C q = ulimb * inverse
;;
or r31 = r22, r23
shr.u r23 = r21, rshift
;;
setf.sig f11 = r31
(p8) br.cond.dptk .Lx3 C branch for n = 3
;;
ld8 r21 = [up], 8
br .Lent
.Loop: ld8 r21 = [up], 8
xma.l f12 = f9, f8, f10 C q = c * -inverse + si
;;
.Lent: add r16 = 160, up
shl r22 = r21, lshift
;;
stf8 [rp] = f12, 8
xma.hu f9 = f12, f6, f9 C c = high(q * divisor + c)
xmpy.l f10 = f11, f7 C si = ulimb * inverse
;;
or r31 = r22, r23
shr.u r23 = r21, rshift
;;
lfetch [r16]
setf.sig f11 = r31
br.cloop.sptk.few.clr .Loop
xma.l f12 = f9, f8, f10 C q = c * -inverse + si
;;
.Lx3: stf8 [rp] = f12, 8
xma.hu f9 = f12, f6, f9 C c = high(q * divisor + c)
xmpy.l f10 = f11, f7 C si = ulimb * inverse
;;
setf.sig f11 = r23
;;
xma.l f12 = f9, f8, f10 C q = c * -inverse + si
;;
.Lx2: stf8 [rp] = f12, 8
xma.hu f9 = f12, f6, f9 C c = high(q * divisor + c)
xmpy.l f10 = f11, f7 C si = ulimb * inverse
;;
xma.l f12 = f9, f8, f10 C q = c * -inverse + si
;;
.Lx1: stf8 [rp] = f12, 8
mov ar.lc = r2 C I0
br.ret.sptk.many b0
EPILOGUE()