[PATCH] T3/T4 sparc shifts, plus more timings
David Miller
davem at davemloft.net
Sat Mar 30 02:58:19 CET 2013
From: Torbjorn Granlund <tg at gmplib.org>
Date: Sat, 30 Mar 2013 02:46:09 +0100
> David Miller <davem at davemloft.net> writes:
>
> So here is a working generic 64-bit sparc lshift.asm that seems
> to work well on all chips.
>
> I'm now going to iterate over lshiftc and rshift.
>
> I whacked off some code at the end, and generalised the resulting code
> to become a lorrshift.asm. Some additional things should be
> streamlined, see comments in the code.
Looks good, and here is my lshiftc:
dnl SPARC v9 mpn_lshiftc
dnl Contributed to the GNU project by David Miller.
dnl Copyright 2013 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 UltraSPARC 1&2: 2
C UltraSPARC 3: 2.5
C UltraSPARC T3: 10
C UltraSPARC T4: 3.5
C INPUT PARAMETERS
define(`rp', `%i0')
define(`up', `%i1')
define(`n', `%i2')
define(`cnt', `%i3')
define(`tcnt', `%i4')
define(`retval', `%i5')
define(`u0', `%l0')
define(`u1', `%l1')
define(`r0', `%l6')
define(`r1', `%l7')
define(`u0_off', `%o0')
define(`u1_off', `%o1')
define(`r0_off', `%o2')
define(`r1_off', `%o3')
ASM_START()
REGISTER(%g2,#scratch)
REGISTER(%g3,#scratch)
PROLOGUE(mpn_lshiftc)
save %sp, -176, %sp
sllx n, 3, n
sub %g0, cnt, tcnt
sub up, 8, u1_off
add rp, (5 * 8), r1_off
ldx [n + u1_off], u1 C WAS: up - 8
add u1_off, (3 * 8), u1_off
sub r1_off, 8, r0_off
sub u1_off, 8, u0_off
subcc n, (3 * 8), n
srlx u1, tcnt, retval
bl,pn %xcc, L(end12)
sllx u1, cnt, %l3
ldx [n + u0_off], u0 C WAS: up - 16
subcc n, (2 * 8), n
ldx [n + u1_off], u1 C WAS: up - 24
bl,pn %xcc, L(end34)
srlx u0, tcnt, %l4
b,a L(top)
ALIGN(16)
L(top):
not %l3, %l3
sllx u0, cnt, %l2
andn %l3, %l4, r0
ldx [n + u0_off], u0 C WAS: up - 16
srlx u1, tcnt, %l5
stx r0, [n + r0_off] C WAS: rp - 8
subcc n, (2 * 8), n
not %l2, %l2
sllx u1, cnt, %l3
andn %l2, %l5, r1
ldx [n + u1_off], u1 C WAS: up - 24
srlx u0, tcnt, %l4
bge,pt %xcc, L(top)
stx r1, [n + r1_off] C WAS: rp - 16
L(end34):
not %l3, %l3
sllx u0, cnt, %l2
andn %l3, %l4, r0
srlx u1, tcnt, %l5
stx r0, [n + r0_off] C WAS: rp - 8
not %l2, %l2
andn %l2, %l5, r1
sub n, (2 * 8), %o5
sllx u1, cnt, %l3
stx r1, [%o5 + r1_off] C WAS: rp - 16
L(end12):
add n, r0_off, rp
add n, u0_off, up
addcc n, (2 * 8), n
sub rp, 8, rp
bz,pn %xcc, L(done)
add up, 16, up
L(loop0):
add rp, -8, rp
not %l3, %l3
subcc n, (1 * 8), n
ldx [up - 16], u1
add up, -8, up
srlx u1, tcnt, %l4
andn %l3, %l4, r0
stx r0, [rp + 0]
bnz,pt %xcc, L(loop0)
sllx u1, cnt, %l3
L(done):
not %l3, %l3
stx %l3, [rp - 8]
ret
restore retval, 0, %o0
EPILOGUE()
More information about the gmp-devel
mailing list