From 573b91127359a4c4d7fd5d0155d1e7492e42a20e Mon Sep 17 00:00:00 2001
From: jasonmoxham <jason@njkfrudils.plus.com>
Date: Sun, 23 Aug 2009 22:20:49 +0000
Subject: [PATCH] New asm functions K8/K10 mpn_addlsh_n mpn_sublsh_n and
 carry-in varients

---
 config.in                  |   2 +
 configure.in               |   2 +
 gmp-impl.h                 |   6 ++
 mpn/asm-defs.m4            |   2 +
 mpn/x86_64/k8/addlsh_n.asm | 118 +++++++++++++++++++++++++++++++++++
 mpn/x86_64/k8/sublsh_n.asm | 123 +++++++++++++++++++++++++++++++++++++
 6 files changed, 253 insertions(+)
 create mode 100644 mpn/x86_64/k8/addlsh_n.asm
 create mode 100644 mpn/x86_64/k8/sublsh_n.asm

diff --git a/config.in b/config.in
index d13cc752..cf1b79fc 100644
--- a/config.in
+++ b/config.in
@@ -214,6 +214,7 @@ MA 02110-1301, USA.
 #undef HAVE_NATIVE_mpn_addadd_n
 #undef HAVE_NATIVE_mpn_addlsh1_n
 #undef HAVE_NATIVE_mpn_addlsh_n
+#undef HAVE_NATIVE_mpn_addlsh_nc
 #undef HAVE_NATIVE_mpn_addmul_1c
 #undef HAVE_NATIVE_mpn_addmul_2
 #undef HAVE_NATIVE_mpn_addmul_3
@@ -273,6 +274,7 @@ MA 02110-1301, USA.
 #undef HAVE_NATIVE_mpn_subadd_n
 #undef HAVE_NATIVE_mpn_sublsh1_n
 #undef HAVE_NATIVE_mpn_sublsh_n
+#undef HAVE_NATIVE_mpn_sublsh_nc
 #undef HAVE_NATIVE_mpn_submul_1c
 #undef HAVE_NATIVE_mpn_umul_ppmm
 #undef HAVE_NATIVE_mpn_umul_ppmm_r
diff --git a/configure.in b/configure.in
index e5f7c756..f9de492a 100644
--- a/configure.in
+++ b/configure.in
@@ -2962,6 +2962,7 @@ AH_VERBATIM([HAVE_NATIVE],
 #undef HAVE_NATIVE_mpn_addadd_n
 #undef HAVE_NATIVE_mpn_addlsh1_n
 #undef HAVE_NATIVE_mpn_addlsh_n
+#undef HAVE_NATIVE_mpn_addlsh_nc
 #undef HAVE_NATIVE_mpn_addmul_1c
 #undef HAVE_NATIVE_mpn_addmul_2
 #undef HAVE_NATIVE_mpn_addmul_3
@@ -3021,6 +3022,7 @@ AH_VERBATIM([HAVE_NATIVE],
 #undef HAVE_NATIVE_mpn_subadd_n
 #undef HAVE_NATIVE_mpn_sublsh1_n
 #undef HAVE_NATIVE_mpn_sublsh_n
+#undef HAVE_NATIVE_mpn_sublsh_nc
 #undef HAVE_NATIVE_mpn_submul_1c
 #undef HAVE_NATIVE_mpn_umul_ppmm
 #undef HAVE_NATIVE_mpn_umul_ppmm_r
diff --git a/gmp-impl.h b/gmp-impl.h
index 5502526d..76250caa 100644
--- a/gmp-impl.h
+++ b/gmp-impl.h
@@ -823,6 +823,12 @@ __GMP_DECLSPEC mp_limb_t mpn_addlsh_n __GMP_PROTO ((mp_ptr, mp_srcptr, mp_srcptr
 #define mpn_sublsh_n __MPN(sublsh_n)
 __GMP_DECLSPEC mp_limb_t mpn_sublsh_n __GMP_PROTO ((mp_ptr, mp_srcptr, mp_srcptr, mp_size_t,unsigned int));
 
+#define mpn_addlsh_nc __MPN(addlsh_nc)
+__GMP_DECLSPEC mp_limb_t mpn_addlsh_nc __GMP_PROTO ((mp_ptr, mp_srcptr, mp_srcptr, mp_size_t,unsigned int, mp_limb_t));
+
+#define mpn_sublsh_nc __MPN(sublsh_nc)
+__GMP_DECLSPEC mp_limb_t mpn_sublsh_nc __GMP_PROTO ((mp_ptr, mp_srcptr, mp_srcptr, mp_size_t,unsigned int, mp_limb_t));
+
 /* mpn_rsh1add_n(c,a,b,n), when it exists, sets {c,n} to ({a,n} + {b,n}) >> 1,
    and returns the bit rshifted out (0 or 1).  */
 #define mpn_rsh1add_n __MPN(rsh1add_n)
diff --git a/mpn/asm-defs.m4 b/mpn/asm-defs.m4
index f3fbea60..7701fa96 100644
--- a/mpn/asm-defs.m4
+++ b/mpn/asm-defs.m4
@@ -1316,6 +1316,7 @@ define_mpn(add_n)
 define_mpn(add_nc)
 define_mpn(addadd_n)
 define_mpn(addlsh_n)
+define_mpn(addlsh_nc)
 define_mpn(addlsh1_n)
 define_mpn(addmul_1)
 define_mpn(addmul_1c)
@@ -1404,6 +1405,7 @@ define_mpn(sub_err1_n)
 define_mpn(sub_err2_n)
 define_mpn(sub_n)
 define_mpn(sublsh_n)
+define_mpn(sublsh_nc)
 define_mpn(sublsh1_n)
 define_mpn(sqrtrem)
 define_mpn(store)
diff --git a/mpn/x86_64/k8/addlsh_n.asm b/mpn/x86_64/k8/addlsh_n.asm
new file mode 100644
index 00000000..33a77861
--- /dev/null
+++ b/mpn/x86_64/k8/addlsh_n.asm
@@ -0,0 +1,118 @@
+dnl  AMD64 mpn_addlsh_n
+
+dnl  Copyright 2009 Jason Moxham
+
+dnl  This file is part of the MPIR Library.
+
+dnl  The MPIR 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 2.1 of the License, or (at
+dnl  your option) any later version.
+
+dnl  The MPIR 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 MPIR Library; see the file COPYING.LIB.  If not, write
+dnl  to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
+dnl  Boston, MA 02110-1301, USA.
+
+include(`../config.m4')
+
+C    carry+(xp,n)= (yp,n) + (zp,n)<<c with carry in ci
+C mp_limb_t	mpn_addlsh_nc(mp_ptr xp, mp_srcptr yp,mp_srcptr zp,mp_size_t n,unsigned int c,mp_limb_t ci)
+C xp in rdi	yp in rsi	zp in rdx	n  in rcx	c  in r8	ci in r9
+
+MULFUNC_PROLOGUE(mpn_addlsh_n mpn_addlsh_nc)
+
+ASM_START()
+PROLOGUE(mpn_addlsh_n)
+xor %r9,%r9
+EPILOGUE()
+PROLOGUE(mpn_addlsh_nc)
+mov	%rcx,%r10
+lea	(%rdi,%r10,8),%rdi
+lea	(%rsi,%r10,8),%rsi
+lea	(%rdx,%r10,8),%rdx
+mov	%r8,%rcx
+neg	%rcx
+shr	%cl,%r9
+neg	%r10
+xor	%rax,%rax
+test	$3,%r10
+jz	next
+lp:
+	mov	(%rdx,%r10,8),%r8
+	mov	%r8,%r11
+	neg	%rcx
+	shl	%cl,%r8
+	neg	%rcx
+	shr	%cl,%r11
+	or	%r9,%r8
+	mov	%r11,%r9
+	add	$1,%rax
+	adc	(%rsi,%r10,8),%r8
+	sbb	%rax,%rax
+	mov	%r8,(%rdi,%r10,8)
+	inc	%r10
+	test	$3,%r10
+	jnz	lp
+next:
+cmp	$0,%r10
+jz	end
+push	%rbx
+push	%rbp
+push	%r12
+push	%r13
+push	%r14
+push	%r15
+ALIGN(16)
+loop:
+	mov	(%rdx,%r10,8),%r8
+	mov	8(%rdx,%r10,8),%rbp
+	mov	16(%rdx,%r10,8),%rbx
+	mov	24(%rdx,%r10,8),%r12
+	mov	%r8,%r11
+	mov	%rbp,%r13
+	mov	%rbx,%r14
+	mov	%r12,%r15
+	neg	%rcx
+	shl	%cl,%r8
+	shl	%cl,%rbp
+	shl	%cl,%rbx
+	shl	%cl,%r12
+	neg	%rcx
+	shr	%cl,%r11
+	shr	%cl,%r13
+	shr	%cl,%r14
+	shr	%cl,%r15
+	or	%r9,%r8
+	or	%r11,%rbp
+	or	%r13,%rbx
+	or	%r14,%r12
+	mov	%r15,%r9
+	add	$1,%rax
+	adc	(%rsi,%r10,8),%r8
+	adc	8(%rsi,%r10,8),%rbp
+	adc	16(%rsi,%r10,8),%rbx
+	adc	24(%rsi,%r10,8),%r12
+	sbb	%rax,%rax
+	mov	%r8,(%rdi,%r10,8)
+	mov	%rbp,8(%rdi,%r10,8)
+	mov	%rbx,16(%rdi,%r10,8)
+	mov	%r12,24(%rdi,%r10,8)
+	add	$4,%r10
+	jnz	loop
+pop	%r15
+pop	%r14
+pop	%r13
+pop	%r12
+pop	%rbp
+pop	%rbx
+end:
+neg	%rax
+add	%r9,%rax
+ret
+EPILOGUE()
diff --git a/mpn/x86_64/k8/sublsh_n.asm b/mpn/x86_64/k8/sublsh_n.asm
new file mode 100644
index 00000000..dae96922
--- /dev/null
+++ b/mpn/x86_64/k8/sublsh_n.asm
@@ -0,0 +1,123 @@
+dnl  AMD64 mpn_sublsh 
+
+dnl  Copyright 2009 Jason Moxham
+
+dnl  This file is part of the MPIR Library.
+
+dnl  The MPIR 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 2.1 of the License, or (at
+dnl  your option) any later version.
+
+dnl  The MPIR 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 MPIR Library; see the file COPYING.LIB.  If not, write
+dnl  to the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
+dnl  Boston, MA 02110-1301, USA.
+
+include(`../config.m4')
+
+C	    carry+(xp,n)= (yp,n) - (zp,n)<<c with carry in ci
+C	mp_limb_t	mpn_sublsh_nc(mp_ptr xp, mp_srcptr yp,mp_srcptr zp,mp_size_t n,unsigned int c,mp_limb_t ci)
+C	xp in rdi	yp in rsi	zp in rdx	n  in rcx	c  in r8	ci in r9
+
+MULFUNC_PROLOGUE(mpn_sublsh_n mpn_sublsh_nc)
+
+ASM_START()
+PROLOGUE(mpn_sublsh_n)
+xor %r9,%r9
+EPILOGUE()
+PROLOGUE(mpn_sublsh_nc)
+mov	%rcx,%r10
+lea	(%rdi,%r10,8),%rdi
+lea	(%rsi,%r10,8),%rsi
+lea	(%rdx,%r10,8),%rdx
+mov	%r8,%rcx
+neg	%rcx
+shr	%cl,%r9
+neg	%r10
+xor	%rax,%rax
+test	$3,%r10
+jz	next
+lp:
+	mov	(%rdx,%r10,8),%r8
+	mov	%r8,%r11
+	neg	%rcx
+	shl	%cl,%r8
+	neg	%rcx
+	shr	%cl,%r11
+	or	%r9,%r8
+	mov	%r11,%r9
+	add	$1,%rax
+	mov	(%rsi,%r10,8),%r11
+	sbb	%r8,%r11
+	sbb	%rax,%rax
+	mov	%r11,(%rdi,%r10,8)
+	inc	%r10
+	test	$3,%r10
+	jnz	lp
+next:
+cmp	$0,%r10
+jz	end
+push	%rbx
+push	%rbp
+push	%r12
+push	%r13
+push	%r14
+push	%r15
+ALIGN(16)
+loop:
+	mov	(%rdx,%r10,8),%r8
+	mov	8(%rdx,%r10,8),%rbp
+	mov	16(%rdx,%r10,8),%rbx
+	mov	24(%rdx,%r10,8),%r12
+	mov	%r8,%r11
+	mov	%rbp,%r13
+	mov	%rbx,%r14
+	mov	%r12,%r15
+	neg	%rcx
+	shl	%cl,%r8
+	shl	%cl,%rbp
+	shl	%cl,%rbx
+	shl	%cl,%r12
+	neg	%rcx
+	shr	%cl,%r11
+	shr	%cl,%r13
+	shr	%cl,%r14
+	shr	%cl,%r15
+	or	%r9,%r8
+	or	%r11,%rbp
+	or	%r13,%rbx
+	or	%r14,%r12
+	mov	%r15,%r9
+	add	$1,%rax
+	mov	(%rsi,%r10,8),%r11
+	mov	8(%rsi,%r10,8),%r13
+	mov	16(%rsi,%r10,8),%r14
+	mov	24(%rsi,%r10,8),%r15
+	sbb	%r8,%r11
+	sbb	%rbp,%r13
+	sbb	%rbx,%r14
+	sbb	%r12,%r15
+	sbb	%rax,%rax
+	mov	%r11,(%rdi,%r10,8)
+	mov	%r13,8(%rdi,%r10,8)
+	mov	%r14,16(%rdi,%r10,8)
+	mov	%r15,24(%rdi,%r10,8)
+	add	$4,%r10
+	jnz	loop
+pop	%r15
+pop	%r14
+pop	%r13
+pop	%r12
+pop	%rbp
+pop	%rbx
+end:
+neg	%rax
+add	%r9,%rax
+ret
+EPILOGUE()