dnl AMD64 mpn_sqr_basecase optimised for Intel Sandy bridge and Ivy bridge.
dnl Contributed to the GNU project by Torbjörn Granlund.
dnl Copyright 2008, 2009, 2011-2013 Free Software Foundation, 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 mul_2 addmul_2 sqr_diag_addlsh1
C AMD K8,K9 ? ? ?
C AMD K10 ? ? ?
C AMD bull ? ? ?
C AMD pile ? ? ?
C AMD steam ? ? ?
C AMD bobcat ? ? ?
C AMD jaguar ? ? ?
C Intel P4 ? ? ?
C Intel core ? ? ?
C Intel NHM ? ? ?
C Intel SBR 2.57 2.93 3.0
C Intel IBR 2.35 2.66 3.0
C Intel HWL 2.02 2.5 2.5
C Intel BWL ? ? ?
C Intel atom ? ? ?
C VIA nano ? ? ?
C The inner loops of this code are the result of running a code generation and
C optimisation tool suite written by David Harvey and Torbjörn Granlund, except
C that the sqr_diag_addlsh1 loop was manually written.
C TODO
C * Replace current unoptimised sqr_diag_addlsh1 loop, 2.5 c/l should be easy.
C * Streamline pointer updates.
C * Perhaps suppress a few more xor insns in feed-in code.
C * Make sure we write no dead registers in feed-in code.
C * We might use 32-bit size ops, since n >= 2^32 is non-terminating. Watch
C out for negative sizes being zero-extended, though.
C * The straight-line code for n <= 3 comes from the K8 code, and might be
C quite sub-optimal here. Write specific code, and add code for n = 4.
C * The mul_2 loop has a 10 insn common sequence in the loop start and the
C wind-down code. Try re-rolling it.
C * This file has been the subject to just basic micro-optimisation.
C When playing with pointers, set this to $2 to fall back to conservative
C indexing in wind-down code.
define(`I',`$1')
define(`rp', `%rdi')
define(`up', `%rsi')
define(`un_param',`%rdx')
ABI_SUPPORT(DOS64)
ABI_SUPPORT(STD64)
ASM_START()
TEXT
ALIGN(32)
PROLOGUE(mpn_sqr_basecase)
FUNC_ENTRY(3)
cmp $2, un_param
jae L(gt1)
mov (up), %rax
mul %rax
mov %rax, (rp)
mov %rdx, 8(rp)
FUNC_EXIT()
ret
L(gt1): jne L(gt2)
mov (up), %rax
mov %rax, %r8
mul %rax
mov 8(up), %r11
mov %rax, (rp)
mov %r11, %rax
mov %rdx, %r9
mul %rax
mov %rax, %r10
mov %r11, %rax
mov %rdx, %r11
mul %r8
xor %r8, %r8
add %rax, %r9
adc %rdx, %r10
adc %r8, %r11
add %rax, %r9
mov %r9, 8(rp)
adc %rdx, %r10
mov %r10, 16(rp)
adc %r8, %r11
mov %r11, 24(rp)
FUNC_EXIT()
ret
L(gt2): cmp $4, un_param
jae L(gt3)
define(`v0', `%r8')
define(`v1', `%r9')
define(`w0', `%r10')
define(`w2', `%r11')
mov (up), %rax
mov %rax, %r10
mul %rax
mov 8(up), %r11
mov %rax, (rp)
mov %r11, %rax
mov %rdx, 8(rp)
mul %rax
mov 16(up), %rcx
mov %rax, 16(rp)
mov %rcx, %rax
mov %rdx, 24(rp)
mul %rax
mov %rax, 32(rp)
mov %rdx, 40(rp)
mov %r11, %rax
mul %r10
mov %rax, %r8
mov %rcx, %rax
mov %rdx, %r9
mul %r10
xor %r10, %r10
add %rax, %r9
mov %r11, %rax
mov %r10, %r11
adc %rdx, %r10
mul %rcx
add %rax, %r10
adc %r11, %rdx
add %r8, %r8
adc %r9, %r9
adc %r10, %r10
adc %rdx, %rdx
adc %r11, %r11
add %r8, 8(rp)
adc %r9, 16(rp)
adc %r10, 24(rp)
adc %rdx, 32(rp)
adc %r11, 40(rp)
FUNC_EXIT()
ret
L(gt3):
define(`v0', `%r8')
define(`v1', `%r9')
define(`w0', `%r10')
define(`w1', `%r11')
define(`w2', `%rbx')
define(`w3', `%rbp')
define(`un', `%r12')
define(`n', `%rcx')
define(`X0', `%r13')
define(`X1', `%r14')
L(do_mul_2):
mov (up), v0
push %rbx
lea (rp,un_param,8), rp C point rp at R[un]
mov 8(up), %rax
push %rbp
lea (up,un_param,8), up C point up right after U's end
mov %rax, v1
push %r12
mov $1, R32(un) C free up rdx
push %r13
sub un_param, un
push %r14
push un
mul v0
mov %rax, (rp,un,8)
mov 8(up,un,8), %rax
test $1, R8(un)
jnz L(m2b1)
L(m2b0):lea 2(un), n
xor R32(w1), R32(w1) C FIXME
xor R32(w2), R32(w2) C FIXME
mov %rdx, w0
jmp L(m2l0)
L(m2b1):lea 1(un), n
xor R32(w3), R32(w3) C FIXME
xor R32(w0), R32(w0) C FIXME
mov %rdx, w2
jmp L(m2l1)
ALIGN(32)
L(m2tp):
L(m2l0):mul v0
add %rax, w0
mov %rdx, w3
adc $0, w3
mov -8(up,n,8), %rax
mul v1
add w1, w0
adc $0, w3
add %rax, w2
mov w0, -8(rp,n,8)
mov %rdx, w0
adc $0, w0
mov (up,n,8), %rax
L(m2l1):mul v0
add %rax, w2
mov %rdx, w1
adc $0, w1
add w3, w2
mov (up,n,8), %rax
adc $0, w1
mul v1
mov w2, (rp,n,8)
add %rax, w0
mov %rdx, w2
mov 8(up,n,8), %rax
adc $0, w2
add $2, n
jnc L(m2tp)
L(m2ed):mul v0
add %rax, w0
mov %rdx, w3
adc $0, w3
mov I(-8(up),-8(up,n,8)), %rax
mul v1
add w1, w0
adc $0, w3
add %rax, w2
mov w0, I(-8(rp),-8(rp,n,8))
adc $0, %rdx
add w3, w2
mov w2, I((rp),(rp,n,8))
adc $0, %rdx
mov %rdx, I(8(rp),8(rp,n,8))
add $2, un C decrease |un|
L(do_addmul_2):
L(outer):
lea 16(rp), rp
cmp $-2, R32(un) C jump if un C {-1,0} FIXME jump if un C {-2,1}
jge L(corner) C FIXME: move to before the lea above
mov -8(up,un,8), v0
mov (up,un,8), %rax
mov %rax, v1
mul v0
test $1, R8(un)
jnz L(a1x1)
L(a1x0):mov (rp,un,8), X0
xor w0, w0
mov 8(rp,un,8), X1
add %rax, X0
mov %rdx, w1
adc $0, w1
xor w2, w2
mov X0, (rp,un,8)
mov 8(up,un,8), %rax
test $2, R8(un)
jnz L(a110)
L(a100):lea 2(un), n C un = 4, 8, 12, ...
jmp L(lo0)
L(a110):lea (un), n C un = 2, 6, 10, ...
jmp L(lo2)
L(a1x1):mov (rp,un,8), X1
xor w2, w2
mov 8(rp,un,8), X0
add %rax, X1
mov %rdx, w3
adc $0, w3
xor w0, w0
mov 8(up,un,8), %rax
test $2, R8(un)
jz L(a111)
L(a101):lea 3(un), n C un = 1, 5, 9, ...
jmp L(lo1)
L(a111):lea 1(un), n C un = 3, 7, 11, ...
jmp L(lo3)
ALIGN(32)
L(top): mul v1
mov %rdx, w0
add %rax, X0
adc $0, w0
add w1, X1
adc $0, w3
add w2, X0
adc $0, w0
mov -16(up,n,8), %rax
L(lo1): mul v0
add %rax, X0
mov %rdx, w1
adc $0, w1
mov -16(up,n,8), %rax
mul v1
mov X1, -24(rp,n,8)
mov -8(rp,n,8), X1
add w3, X0
adc $0, w1
mov %rdx, w2
mov X0, -16(rp,n,8)
add %rax, X1
adc $0, w2
mov -8(up,n,8), %rax
add w0, X1
adc $0, w2
L(lo0): mul v0
add %rax, X1
mov %rdx, w3
adc $0, w3
mov -8(up,n,8), %rax
mul v1
add w1, X1
mov (rp,n,8), X0
adc $0, w3
mov %rdx, w0
add %rax, X0
adc $0, w0
mov (up,n,8), %rax
L(lo3): mul v0
add w2, X0
mov X1, -8(rp,n,8)
mov %rdx, w1
adc $0, w0
add %rax, X0
adc $0, w1
mov (up,n,8), %rax
add w3, X0
adc $0, w1
mul v1
mov 8(rp,n,8), X1
add %rax, X1
mov %rdx, w2
adc $0, w2
mov 8(up,n,8), %rax
mov X0, (rp,n,8)
L(lo2): mul v0
add w0, X1
mov %rdx, w3
adc $0, w2
add %rax, X1
mov 8(up,n,8), %rax
mov 16(rp,n,8), X0
adc $0, w3
add $4, n
jnc L(top)
L(end): mul v1
add w1, X1
adc $0, w3
add w2, %rax
adc $0, %rdx
mov X1, I(-8(rp),-24(rp,n,8))
add w3, %rax
adc $0, %rdx
mov %rax, I((rp),-16(rp,n,8))
mov %rdx, I(8(rp),-8(rp,n,8))
add $2, un C decrease |un|
jmp L(outer) C loop until a small corner remains
L(corner):
pop n
jg L(small_corner)
lea 8(rp), rp
mov -24(up), v0
mov -16(up), %rax
mov %rax, v1
mul v0
mov -24(rp), X0
mov -16(rp), X1
add %rax, X0
mov %rdx, w1
adc $0, w1
xor w2, w2
mov X0, -24(rp)
mov -8(up), %rax
mul v0
add $0, X1
mov %rdx, w3
adc $0, w2
add %rax, X1
mov -8(up), %rax
adc $0, w3
mul v1
add w1, X1
adc $0, w3
add w2, %rax
adc $0, %rdx
mov X1, -16(rp)
jmp L(com)
L(small_corner):
mov -8(rp), w3
mov -16(up), v0
mov -8(up), %rax
mul v0
L(com): add w3, %rax
adc $0, %rdx
mov %rax, -8(rp)
mov %rdx, (rp)
L(sqr_diag_addlsh1):
mov -8(up,n,8), %rax
shl n
mul %rax
mov %rax, (rp,n,8)
xor R32(%rbx), R32(%rbx)
mov 8(rp,n,8), %r8
mov 16(rp,n,8), %r9
jmp L(dm)
ALIGN(32)
L(dtop):add %r8, %r10
adc %r9, %rax
mov 8(rp,n,8), %r8
mov 16(rp,n,8), %r9
mov %r10, -8(rp,n,8)
mov %rax, (rp,n,8)
L(dm): adc %r8, %r8
adc %r9, %r9
mov (up,n,4), %rax
lea (%rdx,%rbx), %r10
setc R8(%rbx)
mul %rax
add $2, n
js L(dtop)
L(dend):add %r8, %r10
adc %r9, %rax
mov %r10, I(-8(rp),-8(rp,n,8))
mov %rax, I((rp),(rp,n,8))
adc %rbx, %rdx
mov %rdx, I(8(rp),8(rp,n,8))
pop %r14
pop %r13
pop %r12
pop %rbp
pop %rbx
FUNC_EXIT()
ret
EPILOGUE()