dnl Power9 mpn_sqr_basecase.
dnl Copyright 1999-2001, 2003-2006, 2008, 2017-2018 Free Software Foundation,
dnl Inc.
dnl This file is part of the GNU MP Library.
dnl
dnl The GNU MP Library is free software; you can redistribute it and/or modify
dnl it under the terms of either:
dnl
dnl * the GNU Lesser General Public License as published by the Free
dnl Software Foundation; either version 3 of the License, or (at your
dnl option) any later version.
dnl
dnl or
dnl
dnl * the GNU General Public License as published by the Free Software
dnl Foundation; either version 2 of the License, or (at your option) any
dnl later version.
dnl
dnl or both in parallel, as here.
dnl
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 General Public License
dnl for more details.
dnl
dnl You should have received copies of the GNU General Public License and the
dnl GNU Lesser General Public License along with the GNU MP Library. If not,
dnl see https://www.gnu.org/licenses/.
include(`../config.m4')
C cycles/limb
C POWER3/PPC630 -
C POWER4/PPC970 -
C POWER5 -
C POWER6 -
C POWER7 -
C POWER8 -
C POWER9 1.62
C TODO
C * Completely separate evn and odd code into two outer loops. Also consider
C unrolling these two outer loops and thereby eliminate all branches.
C * Avoid the reloading of u1 before every loop start.
C * Reduce register usage.
C * Consider getting rid of cy and instead load 3 u limbs, use addc+adde+adde.
C * Consider skewing conditional adjustments to allow mask creation with subfe
C like in the un=3 code. It might streamline the adjustments (or not).
C INPUT PARAMETERS
define(`rp', `r3')
define(`up', `r4')
define(`un', `r5')
define(`u0', `r0')
define(`u1', `r7')
define(`rp2', `r24')
define(`up2', `r25')
define(`cy', `r6')
define(`LSHU1U0',`
addc u0, u0, u0
adde u1, u1, u1
li cy, 0
addze cy, cy
')
define(`LSHU1U',`
addc u0, u0, u0
add u0, u0, cy
adde u1, u1, u1
li cy, 0
addze cy, cy
')
define(`LSHU1UF',`
addc u0, u0, u0
add u0, u0, cy
adde u1, u1, u1
')
define(`LSHU1UHF',`
add u0, u0, u0
add u0, u0, cy
')
C These are cleverer replacements, but they tend to leave CA set, disturbing
C the main accumulation code! Breaking that false dependency might have a
C positive performance impact. Note that the subfe here results in a mask for
C our adjustments.
define(`xLSHU1U0',`
addc u0, u0, u0
adde u1, u1, u1
subfe cy, cy, cy
')
define(`xLSHU1U',`
subfic cy, cy, 0
adde u0, u0, u0
adde u1, u1, u1
subfe cy, cy, cy
')
define(`xLSHU1U',`
subfic cy, cy, 0
adde u0, u0, u0
')
ASM_START()
PROLOGUE(mpn_sqr_basecase)
ld r0, 0(up) C n = 1
mulld r8, r0, r0 C weight 0
mulhdu r9, r0, r0 C weight 1
std r8, 0(rp)
cmpdi cr0, un, 2
bge cr0, L(ge2)
std r9, 8(rp)
blr
L(ge2): bgt cr0, L(gt2)
ld r6, 8(up)
mulld r10, r6, r6 C u1 * u1
mulhdu r11, r6, r6 C u1 * u1
mulld r4, r6, r0 C u1 * u0
mulhdu r5, r6, r0 C u1 * u0
addc r4, r4, r4
adde r5, r5, r5
addze r11, r11
addc r9, r9, r4
adde r10, r10, r5
addze r11, r11
std r9, 8(rp)
std r10, 16(rp)
std r11, 24(rp)
blr
L(gt2): cmpdi cr0, un, 3
bgt cr0, L(gt3)
std r30, -16(r1)
std r31, -8(r1)
subfo r12, r12, r12 C clear OV (and result register)
ld r8, 8(r4)
mulld r5, r8, r8 C W2
mulhdu r10, r8, r8 C W3
sradi r11, u0, 63 C CAUTION: clobbers CA
and r11, r11, r8 C W3
addc u0, u0, u0
adde u1, r8, r8
subfe r6, r6, r6 C mask
ld r4, 16(r4) C W2
mulld r12, r8, u0 C W1 u1 x u0
mulhdu r8, r8, u0 C W2 u1 x u0
maddld( r31, r4, u0, r11) C W2
maddhdu(r30, r4, u0, r11) C W3
andc r6, r4, r6 C W4
addc r9, r12, r9 C W1
std r9, 8(rp) C W1
mulld r9, r4, u1 C W3
mulhdu r11, r4, u1 C W4
addex( r5, r5, r8, 0) C W2
adde r5, r31, r5 C W2
std r5, 16(rp) C W2
maddld( r5, r4, r4, r6) C W4 u2^2
maddhdu(r6, r4, r4, r6) C W5 u2^2
addex( r9, r9, r30, 0) C W3
adde r9, r9, r10 C W3
std r9, 24(rp) C W3
adde r5, r5, r11 C W4
addze r6, r6 C W5
li r8, 0
addex( r5, r5, r8, 0) C W4
std r5, 32(rp) C W4
addex( r6, r6, r8, 0) C W5
std r6, 40(rp) C W5
ld r30, -16(r1)
ld r31, -8(r1)
blr
L(gt3): std r22, -80(r1)
std r23, -72(r1)
std r24, -64(r1)
std r25, -56(r1)
std r26, -48(r1)
std r27, -40(r1)
std r28, -32(r1)
std r29, -24(r1)
std r30, -16(r1)
std r31, -8(r1)
mr rp2, rp
mr up2, up
addi r22, un, -1 C count for loop FIXME: Adjust
subfo r0, r0, r0 C clear OV (and r0)
rldicl r0, un, 0, 63 C r0 = un & 1
cmpdi cr7, r0, 0
ld u0, 0(up2)
ld u1, 8(up2)
cmpdi cr5, r22, 4
srdi r31, r22, 2
addi r22, r22, -2
mtctr r31
beq cr7, L(m2_evn)
L(m2_odd):
rldicl. r31, r22, 63, 63 C r22 & 2
mulld r23, u0, u0
mulhdu r12, u0, u0
mulld r5, u1, u1
mulhdu r10, u1, u1
sradi r11, u0, 63
and r11, r11, u1
LSHU1U0
ld r8, 8(up2)
ld r9, 16(up2)
mulld r28, r8, u0 C W u1 x u0
mulhdu r31, r8, u0 C W u1 x u0
std r23, 0(rp2)
bne cr0, L(m2_11)
L(m2_01):
addi up, up2, 16
addi rp, rp2, 0
b L(m2_lo2)
L(m2_11):
addi up, up2, 0
addi rp, rp2, -16
b L(m2_lo0)
L(m2_evn):
rldicl. r31, r22, 63, 63 C r22 & 2
mulld r23, u0, u0
mulhdu r5, u0, u0
mulld r12, u1, u1
mulhdu r11, u1, u1
sradi r10, u0, 63
and r10, r10, u1
LSHU1U0
ld r9, 8(up2)
ld r8, 16(up2)
mulld r29, r9, u0 C W u1 x u0
mulhdu r30, r9, u0 C W u1 x u0
std r23, 0(rp2)
beq cr0, L(m2_10)
L(m2_00):
addi up, up2, 8
addi rp, rp2, -8
b L(m2_lo1)
L(m2_10):
addi up, up2, 24
addi rp, rp2, 8
ble cr5, L(m2_end)
L(m2_top):
ld r9, 0(up)
maddld( r28, r8, u0, r10)
maddhdu(r31, r8, u0, r10)
adde r5, r29, r5
std r5, 0(rp)
mulld r5, r8, u1
mulhdu r10, r8, u1
addex( r12, r12, r30, 0)
L(m2_lo2):
ld r8, 8(up)
maddld( r29, r9, u0, r11)
maddhdu(r30, r9, u0, r11)
adde r12, r28, r12
std r12, 8(rp)
mulld r12, r9, u1
mulhdu r11, r9, u1
addex( r5, r5, r31, 0)
L(m2_lo1):
ld r9, 16(up)
maddld( r28, r8, u0, r10)
maddhdu(r31, r8, u0, r10)
adde r5, r29, r5
std r5, 16(rp)
mulld r5, r8, u1
mulhdu r10, r8, u1
addex( r12, r12, r30, 0)
L(m2_lo0):
ld r8, 24(up)
maddld( r29, r9, u0, r11)
maddhdu(r30, r9, u0, r11)
adde r12, r28, r12
std r12, 24(rp)
mulld r12, r9, u1
mulhdu r11, r9, u1
addex( r5, r5, r31, 0)
addi up, up, 32
addi rp, rp, 32
bdnz L(m2_top)
L(m2_end):
ld r9, 0(up)
maddld( r28, r8, u0, r10)
maddhdu(r31, r8, u0, r10)
adde r5, r29, r5
std r5, 0(rp)
mulld r5, r8, u1
mulhdu r10, r8, u1
b L(cj) C jump to addmul_2 tail
L(outer):
addi up2, up2, 16
addi rp2, rp2, 32
ld u0, 0(up2)
ld u1, 8(up2)
cmpdi cr5, r22, 4
srdi r31, r22, 2
addi r22, r22, -2
mtctr r31
ld r26, 0(rp2)
ld r27, 16(rp2)
rldicl. r31, r22, 63, 63 C r22 & 2
beq cr7, L(evn)
L(odd): maddld( r23, u0, u0, r26) C W u2^2
maddhdu(r12, u0, u0, r26) C W u2^2
maddld( r5, u1, u1, r27) C W u3^2
maddhdu(r10, u1, u1, r27) C W u3^2
ld r26, 8(rp2)
ld r8, -8(up2)
sradi r8, r8, 63 C CAUTION: clobbers CA
and r8, r8, u0
sradi r11, u0, 63 C CAUTION: clobbers CA
and r11, r11, u1
LSHU1U
addc r23, r23, r8
ld r8, 8(up2)
ld r9, 16(up2)
maddld( r28, r8, u0, r26) C W u3 x u2
maddhdu(r31, r8, u0, r26) C W u3 x u2
ld r26, 24(rp2)
std r23, 0(rp2) C W0
bne cr0, L(11)
L(01):
addi up, up2, 16
addi rp, rp2, 0
b L(lo2)
L(11):
addi up, up2, 0
addi rp, rp2, -16
b L(lo0)
L(evn): maddld( r23, u0, u0, r26) C W u2^2
maddhdu(r5, u0, u0, r26) C W u2^2
maddld( r12, u1, u1, r27) C W u3^2
maddhdu(r11, u1, u1, r27) C W u3^2
ld r27, 8(rp2)
ld r9, -8(up2)
sradi r9, r9, 63 C CAUTION: clobbers CA
and r9, r9, u0
sradi r10, u0, 63 C CAUTION: clobbers CA
and r10, r10, u1
LSHU1U
addc r23, r23, r9
ld r9, 8(up2)
ld r8, 16(up2)
maddld( r29, r9, u0, r27) C W u3 x u2
maddhdu(r30, r9, u0, r27) C W u3 x u2
ld r27, 24(rp2)
std r23, 0(rp2) C W0
beq cr0, L(10)
L(00):
addi up, up2, 8
addi rp, rp2, -8
b L(lo1)
L(10):
addi up, up2, 24
addi rp, rp2, 8
ble cr5, L(end)
L(top): ld r9, 0(up)
maddld( r28, r8, u0, r10)
maddhdu(r31, r8, u0, r10)
adde r5, r29, r5
ld r26, 24(rp)
std r5, 0(rp)
maddld( r5, r8, u1, r27)
maddhdu(r10, r8, u1, r27)
addex( r12, r12, r30, 0)
L(lo2): ld r8, 8(up)
maddld( r29, r9, u0, r11)
maddhdu(r30, r9, u0, r11)
adde r12, r28, r12
ld r27, 32(rp)
std r12, 8(rp)
maddld( r12, r9, u1, r26)
maddhdu(r11, r9, u1, r26)
addex( r5, r5, r31, 0)
L(lo1): ld r9, 16(up)
maddld( r28, r8, u0, r10)
maddhdu(r31, r8, u0, r10)
adde r5, r29, r5
ld r26, 40(rp)
std r5, 16(rp)
maddld( r5, r8, u1, r27)
maddhdu(r10, r8, u1, r27)
addex( r12, r12, r30, 0)
L(lo0): ld r8, 24(up)
maddld( r29, r9, u0, r11)
maddhdu(r30, r9, u0, r11)
adde r12, r28, r12
ld r27, 48(rp)
std r12, 24(rp)
maddld( r12, r9, u1, r26)
maddhdu(r11, r9, u1, r26)
addex( r5, r5, r31, 0)
addi up, up, 32
addi rp, rp, 32
bdnz L(top)
L(end): ld r9, 0(up)
maddld( r28, r8, u0, r10)
maddhdu(r31, r8, u0, r10)
adde r5, r29, r5
std r5, 0(rp)
maddld( r5, r8, u1, r27)
maddhdu(r10, r8, u1, r27)
L(cj): addex( r12, r12, r30, 0)
maddld( r29, r9, u0, r11)
maddhdu(r30, r9, u0, r11)
adde r12, r28, r12
std r12, 8(rp)
mulld r12, r9, u1
mulhdu r11, r9, u1
addex( r5, r5, r31, 0)
adde r5, r29, r5
std r5, 16(rp)
addex( r12, r12, r30, 0)
adde r12, r12, r10
std r12, 24(rp)
li r4, 0
addze r5, r11
addex( r5, r5, r4, 0)
std r5, 32(rp)
bgt cr5, L(outer)
L(corner):
ld u0, 16(up2)
ld u1, 24(up2)
ld r26, 32(rp2)
bne cr7, L(corner_odd)
L(corner_evn):
ld r27, 40(rp2)
maddld( r23, u0, u0, r26) C W u2^2
maddhdu(r5, u0, u0, r26) C W u2^2
mulld r12, u1, u1 C W u3^2
mulhdu r11, u1, u1 C W u3^2
ld r9, 8(up2)
sradi r9, r9, 63 C CAUTION: clobbers CA
and r9, r9, u0
sradi r10, u0, 63 C CAUTION: clobbers CA
and r10, r10, u1
LSHU1UHF
addc r23, r23, r9
ld r9, 24(up2)
maddld( r29, r9, u0, r27) C W u3 x u2
maddhdu(r30, r9, u0, r27) C W u3 x u2
std r23, 32(rp2)
adde r5, r29, r5
std r5, 40(rp2)
addex( r12, r12, r30, 0)
adde r12, r12, r10 C W FIXME can this co?
std r12, 48(rp2)
li r4, 0
addex( r5, r11, r4, 0)
addze r5, r5
std r5, 56(rp2)
b L(ret)
L(corner_odd):
ld r27, 48(rp2)
maddld( r23, u0, u0, r26) C W u2^2
maddhdu(r12, u0, u0, r26) C W u2^2
maddld( r5, u1, u1, r27) C W u3^2
maddhdu(r10, u1, u1, r27) C W u3^2
ld r26, 40(rp2)
ld r8, 8(up2)
sradi r8, r8, 63 C CAUTION: clobbers CA
and r8, r8, u0
sradi r11, u0, 63 C CAUTION: clobbers CA
and r11, r11, u1
LSHU1UF
addc r23, r23, r8
ld r8, 24(up2)
ld r9, 32(up2)
maddld( r28, r8, u0, r26) C W u3 x u2
maddhdu(r31, r8, u0, r26) C W u3 x u2
std r23, 32(rp2)
maddld( r29, r9, u0, r11)
maddhdu(r30, r9, u0, r11)
adde r12, r28, r12
std r12, 40(rp2)
mulld r12, r9, u1
mulhdu r11, r9, u1
addex( r5, r5, r31, 0)
adde r5, r29, r5
std r5, 48(rp2)
addex( r12, r12, r30, 0)
adde r12, r12, r10
std r12, 56(rp2)
mulld r23, r9, r9 C W u2^2
mulhdu r12, r9, r9 C W u2^2
adde r23, r23, r11
addze r12, r12
sradi r4, r8, 63 C CAUTION: clobbers CA
and r4, r4, r9
addex( r23, r23, r4, 0)
std r23, 64(rp2)
li r4, 0
addex( r12, r12, r4, 0)
std r12, 72(rp2)
L(ret): ld r22, -80(r1)
ld r23, -72(r1)
ld r24, -64(r1)
ld r25, -56(r1)
ld r26, -48(r1)
ld r27, -40(r1)
ld r28, -32(r1)
ld r29, -24(r1)
ld r30, -16(r1)
ld r31, -8(r1)
blr
EPILOGUE()
ASM_END()