public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
* [wide-int] Add fast path for hosts with HWI widening multiplication
@ 2013-12-01 10:29 Richard Sandiford
  2013-12-01 18:48 ` Ian Lance Taylor
  2013-12-04 12:56 ` Richard Sandiford
  0 siblings, 2 replies; 24+ messages in thread
From: Richard Sandiford @ 2013-12-01 10:29 UTC (permalink / raw)
  To: gcc-patches; +Cc: zadeck, mikestump

This patch handles multiplications using a single HWIxHWI->2HWI multiplication
on hosts that have one.  This removes all uses of the slow (half-HWI) path
for insn-recog.ii.  The slow path is still used 58 times for cp/parser.ii
and 168 times for fold-const.ii, but at that kind of level it shouldn't
matter much.

I followed Joseph's suggestion and reused longlong.h.  I copied it from
libgcc rather than glibc since it seemed better for GCC to have a single
version across both gcc/ and libgcc/.  I can put it in include/ if that
seems better.

Tested on x86_64-linux-gnu.  OK to install?

Thanks,
Richard


Index: gcc/longlong.h
===================================================================
--- /dev/null	2013-11-30 13:13:45.045406132 +0000
+++ gcc/longlong.h	2013-11-30 16:15:13.877457660 +0000
@@ -0,0 +1,1724 @@
+/* longlong.h -- definitions for mixed size 32/64 bit arithmetic.
+   Copyright (C) 1991-2013 Free Software Foundation, Inc.
+
+   This file is part of the GNU C Library.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Lesser General Public
+   License as published by the Free Software Foundation; either
+   version 2.1 of the License, or (at your option) any later version.
+
+   In addition to the permissions in the GNU Lesser General Public
+   License, the Free Software Foundation gives you unlimited
+   permission to link the compiled version of this file into
+   combinations with other programs, and to distribute those
+   combinations without any restriction coming from the use of this
+   file.  (The Lesser General Public License restrictions do apply in
+   other respects; for example, they cover modification of the file,
+   and distribution when not linked into a combine executable.)
+
+   The GNU C Library is distributed in the hope that it will be useful,
+   but WITHOUT ANY WARRANTY; without even the implied warranty of
+   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+   Lesser General Public License for more details.
+
+   You should have received a copy of the GNU Lesser General Public
+   License along with the GNU C Library; if not, see
+   <http://www.gnu.org/licenses/>.  */
+
+/* You have to define the following before including this file:
+
+   UWtype -- An unsigned type, default type for operations (typically a "word")
+   UHWtype -- An unsigned type, at least half the size of UWtype.
+   UDWtype -- An unsigned type, at least twice as large a UWtype
+   W_TYPE_SIZE -- size in bits of UWtype
+
+   UQItype -- Unsigned 8 bit type.
+   SItype, USItype -- Signed and unsigned 32 bit types.
+   DItype, UDItype -- Signed and unsigned 64 bit types.
+
+   On a 32 bit machine UWtype should typically be USItype;
+   on a 64 bit machine, UWtype should typically be UDItype.  */
+
+#define __BITS4 (W_TYPE_SIZE / 4)
+#define __ll_B ((UWtype) 1 << (W_TYPE_SIZE / 2))
+#define __ll_lowpart(t) ((UWtype) (t) & (__ll_B - 1))
+#define __ll_highpart(t) ((UWtype) (t) >> (W_TYPE_SIZE / 2))
+
+#ifndef W_TYPE_SIZE
+#define W_TYPE_SIZE	32
+#define UWtype		USItype
+#define UHWtype		USItype
+#define UDWtype		UDItype
+#endif
+
+/* Used in glibc only.  */
+#ifndef attribute_hidden
+#define attribute_hidden
+#endif
+
+extern const UQItype __clz_tab[256] attribute_hidden;
+
+/* Define auxiliary asm macros.
+
+   1) umul_ppmm(high_prod, low_prod, multiplier, multiplicand) multiplies two
+   UWtype integers MULTIPLIER and MULTIPLICAND, and generates a two UWtype
+   word product in HIGH_PROD and LOW_PROD.
+
+   2) __umulsidi3(a,b) multiplies two UWtype integers A and B, and returns a
+   UDWtype product.  This is just a variant of umul_ppmm.
+
+   3) udiv_qrnnd(quotient, remainder, high_numerator, low_numerator,
+   denominator) divides a UDWtype, composed by the UWtype integers
+   HIGH_NUMERATOR and LOW_NUMERATOR, by DENOMINATOR and places the quotient
+   in QUOTIENT and the remainder in REMAINDER.  HIGH_NUMERATOR must be less
+   than DENOMINATOR for correct operation.  If, in addition, the most
+   significant bit of DENOMINATOR must be 1, then the pre-processor symbol
+   UDIV_NEEDS_NORMALIZATION is defined to 1.
+
+   4) sdiv_qrnnd(quotient, remainder, high_numerator, low_numerator,
+   denominator).  Like udiv_qrnnd but the numbers are signed.  The quotient
+   is rounded towards 0.
+
+   5) count_leading_zeros(count, x) counts the number of zero-bits from the
+   msb to the first nonzero bit in the UWtype X.  This is the number of
+   steps X needs to be shifted left to set the msb.  Undefined for X == 0,
+   unless the symbol COUNT_LEADING_ZEROS_0 is defined to some value.
+
+   6) count_trailing_zeros(count, x) like count_leading_zeros, but counts
+   from the least significant end.
+
+   7) add_ssaaaa(high_sum, low_sum, high_addend_1, low_addend_1,
+   high_addend_2, low_addend_2) adds two UWtype integers, composed by
+   HIGH_ADDEND_1 and LOW_ADDEND_1, and HIGH_ADDEND_2 and LOW_ADDEND_2
+   respectively.  The result is placed in HIGH_SUM and LOW_SUM.  Overflow
+   (i.e. carry out) is not stored anywhere, and is lost.
+
+   8) sub_ddmmss(high_difference, low_difference, high_minuend, low_minuend,
+   high_subtrahend, low_subtrahend) subtracts two two-word UWtype integers,
+   composed by HIGH_MINUEND_1 and LOW_MINUEND_1, and HIGH_SUBTRAHEND_2 and
+   LOW_SUBTRAHEND_2 respectively.  The result is placed in HIGH_DIFFERENCE
+   and LOW_DIFFERENCE.  Overflow (i.e. carry out) is not stored anywhere,
+   and is lost.
+
+   If any of these macros are left undefined for a particular CPU,
+   C macros are used.  */
+
+/* The CPUs come in alphabetical order below.
+
+   Please add support for more CPUs here, or improve the current support
+   for the CPUs below!
+   (E.g. WE32100, IBM360.)  */
+
+#if defined (__GNUC__) && !defined (NO_ASM)
+
+/* We sometimes need to clobber "cc" with gcc2, but that would not be
+   understood by gcc1.  Use cpp to avoid major code duplication.  */
+#if __GNUC__ < 2
+#define __CLOBBER_CC
+#define __AND_CLOBBER_CC
+#else /* __GNUC__ >= 2 */
+#define __CLOBBER_CC : "cc"
+#define __AND_CLOBBER_CC , "cc"
+#endif /* __GNUC__ < 2 */
+
+#if defined (__alpha) && W_TYPE_SIZE == 64
+#define umul_ppmm(ph, pl, m0, m1) \
+  do {									\
+    UDItype __m0 = (m0), __m1 = (m1);					\
+    (ph) = __builtin_alpha_umulh (__m0, __m1);				\
+    (pl) = __m0 * __m1;							\
+  } while (0)
+#define UMUL_TIME 46
+#ifndef LONGLONG_STANDALONE
+#define udiv_qrnnd(q, r, n1, n0, d) \
+  do { UDItype __r;							\
+    (q) = __udiv_qrnnd (&__r, (n1), (n0), (d));				\
+    (r) = __r;								\
+  } while (0)
+extern UDItype __udiv_qrnnd (UDItype *, UDItype, UDItype, UDItype);
+#define UDIV_TIME 220
+#endif /* LONGLONG_STANDALONE */
+#ifdef __alpha_cix__
+#define count_leading_zeros(COUNT,X)	((COUNT) = __builtin_clzl (X))
+#define count_trailing_zeros(COUNT,X)	((COUNT) = __builtin_ctzl (X))
+#define COUNT_LEADING_ZEROS_0 64
+#else
+#define count_leading_zeros(COUNT,X) \
+  do {									\
+    UDItype __xr = (X), __t, __a;					\
+    __t = __builtin_alpha_cmpbge (0, __xr);				\
+    __a = __clz_tab[__t ^ 0xff] - 1;					\
+    __t = __builtin_alpha_extbl (__xr, __a);				\
+    (COUNT) = 64 - (__clz_tab[__t] + __a*8);				\
+  } while (0)
+#define count_trailing_zeros(COUNT,X) \
+  do {									\
+    UDItype __xr = (X), __t, __a;					\
+    __t = __builtin_alpha_cmpbge (0, __xr);				\
+    __t = ~__t & -~__t;							\
+    __a = ((__t & 0xCC) != 0) * 2;					\
+    __a += ((__t & 0xF0) != 0) * 4;					\
+    __a += ((__t & 0xAA) != 0);						\
+    __t = __builtin_alpha_extbl (__xr, __a);				\
+    __a <<= 3;								\
+    __t &= -__t;							\
+    __a += ((__t & 0xCC) != 0) * 2;					\
+    __a += ((__t & 0xF0) != 0) * 4;					\
+    __a += ((__t & 0xAA) != 0);						\
+    (COUNT) = __a;							\
+  } while (0)
+#endif /* __alpha_cix__ */
+#endif /* __alpha */
+
+#if defined (__arc__) && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("add.f	%1, %4, %5\n\tadc	%0, %2, %3"		\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "%r" ((USItype) (ah)),					\
+	     "rIJ" ((USItype) (bh)),					\
+	     "%r" ((USItype) (al)),					\
+	     "rIJ" ((USItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("sub.f	%1, %4, %5\n\tsbc	%0, %2, %3"		\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "r" ((USItype) (ah)),					\
+	     "rIJ" ((USItype) (bh)),					\
+	     "r" ((USItype) (al)),					\
+	     "rIJ" ((USItype) (bl)))
+
+#define __umulsidi3(u,v) ((UDItype)(USItype)u*(USItype)v)
+#ifdef __ARC_NORM__
+#define count_leading_zeros(count, x) \
+  do									\
+    {									\
+      SItype c_;							\
+									\
+      __asm__ ("norm.f\t%0,%1\n\tmov.mi\t%0,-1" : "=r" (c_) : "r" (x) : "cc");\
+      (count) = c_ + 1;							\
+    }									\
+  while (0)
+#define COUNT_LEADING_ZEROS_0 32
+#endif
+#endif
+
+#if defined (__arm__) && (defined (__thumb2__) || !defined (__thumb__)) \
+ && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("adds	%1, %4, %5\n\tadc	%0, %2, %3"		\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "%r" ((USItype) (ah)),					\
+	     "rI" ((USItype) (bh)),					\
+	     "%r" ((USItype) (al)),					\
+	     "rI" ((USItype) (bl)) __CLOBBER_CC)
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("subs	%1, %4, %5\n\tsbc	%0, %2, %3"		\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "r" ((USItype) (ah)),					\
+	     "rI" ((USItype) (bh)),					\
+	     "r" ((USItype) (al)),					\
+	     "rI" ((USItype) (bl)) __CLOBBER_CC)
+# if defined(__ARM_ARCH_2__) || defined(__ARM_ARCH_2A__) \
+     || defined(__ARM_ARCH_3__)
+#  define umul_ppmm(xh, xl, a, b)					\
+  do {									\
+    register USItype __t0, __t1, __t2;					\
+    __asm__ ("%@ Inlined umul_ppmm\n"					\
+	   "	mov	%2, %5, lsr #16\n"				\
+	   "	mov	%0, %6, lsr #16\n"				\
+	   "	bic	%3, %5, %2, lsl #16\n"				\
+	   "	bic	%4, %6, %0, lsl #16\n"				\
+	   "	mul	%1, %3, %4\n"					\
+	   "	mul	%4, %2, %4\n"					\
+	   "	mul	%3, %0, %3\n"					\
+	   "	mul	%0, %2, %0\n"					\
+	   "	adds	%3, %4, %3\n"					\
+	   "	addcs	%0, %0, #65536\n"				\
+	   "	adds	%1, %1, %3, lsl #16\n"				\
+	   "	adc	%0, %0, %3, lsr #16"				\
+	   : "=&r" ((USItype) (xh)),					\
+	     "=r" ((USItype) (xl)),					\
+	     "=&r" (__t0), "=&r" (__t1), "=r" (__t2)			\
+	   : "r" ((USItype) (a)),					\
+	     "r" ((USItype) (b)) __CLOBBER_CC );			\
+  } while (0)
+#  define UMUL_TIME 20
+# else
+#  define umul_ppmm(xh, xl, a, b)					\
+  do {									\
+    /* Generate umull, under compiler control.  */			\
+    register UDItype __t0 = (UDItype)(USItype)(a) * (USItype)(b);	\
+    (xl) = (USItype)__t0;						\
+    (xh) = (USItype)(__t0 >> 32);					\
+  } while (0)
+#  define UMUL_TIME 3
+# endif
+# define UDIV_TIME 100
+#endif /* __arm__ */
+
+#if defined(__arm__)
+/* Let gcc decide how best to implement count_leading_zeros.  */
+#define count_leading_zeros(COUNT,X)	((COUNT) = __builtin_clz (X))
+#define count_trailing_zeros(COUNT,X)   ((COUNT) = __builtin_ctz (X))
+#define COUNT_LEADING_ZEROS_0 32
+#endif
+
+#if defined (__AVR__)
+
+#if W_TYPE_SIZE == 16
+#define count_leading_zeros(COUNT,X)  ((COUNT) = __builtin_clz (X))
+#define count_trailing_zeros(COUNT,X) ((COUNT) = __builtin_ctz (X))
+#define COUNT_LEADING_ZEROS_0 16
+#endif /* W_TYPE_SIZE == 16 */
+
+#if W_TYPE_SIZE == 32
+#define count_leading_zeros(COUNT,X)  ((COUNT) = __builtin_clzl (X))
+#define count_trailing_zeros(COUNT,X) ((COUNT) = __builtin_ctzl (X))
+#define COUNT_LEADING_ZEROS_0 32
+#endif /* W_TYPE_SIZE == 32 */
+
+#if W_TYPE_SIZE == 64
+#define count_leading_zeros(COUNT,X)  ((COUNT) = __builtin_clzll (X))
+#define count_trailing_zeros(COUNT,X) ((COUNT) = __builtin_ctzll (X))
+#define COUNT_LEADING_ZEROS_0 64
+#endif /* W_TYPE_SIZE == 64 */
+
+#endif /* defined (__AVR__) */
+
+#if defined (__CRIS__)
+
+#if __CRIS_arch_version >= 3
+#define count_leading_zeros(COUNT, X) ((COUNT) = __builtin_clz (X))
+#define COUNT_LEADING_ZEROS_0 32
+#endif /* __CRIS_arch_version >= 3 */
+
+#if __CRIS_arch_version >= 8
+#define count_trailing_zeros(COUNT, X) ((COUNT) = __builtin_ctz (X))
+#endif /* __CRIS_arch_version >= 8 */
+
+#if __CRIS_arch_version >= 10
+#define __umulsidi3(u,v) ((UDItype)(USItype) (u) * (UDItype)(USItype) (v))
+#else
+#define __umulsidi3 __umulsidi3
+extern UDItype __umulsidi3 (USItype, USItype);
+#endif /* __CRIS_arch_version >= 10 */
+
+#define umul_ppmm(w1, w0, u, v)		\
+  do {					\
+    UDItype __x = __umulsidi3 (u, v);	\
+    (w0) = (USItype) (__x);		\
+    (w1) = (USItype) (__x >> 32);	\
+  } while (0)
+
+/* FIXME: defining add_ssaaaa and sub_ddmmss should be advantageous for
+   DFmode ("double" intrinsics, avoiding two of the three insns handling
+   carry), but defining them as open-code C composing and doing the
+   operation in DImode (UDImode) shows that the DImode needs work:
+   register pressure from requiring neighboring registers and the
+   traffic to and from them come to dominate, in the 4.7 series.  */
+
+#endif /* defined (__CRIS__) */
+
+#if defined (__hppa) && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("add %4,%5,%1\n\taddc %2,%3,%0"				\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "%rM" ((USItype) (ah)),					\
+	     "rM" ((USItype) (bh)),					\
+	     "%rM" ((USItype) (al)),					\
+	     "rM" ((USItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("sub %4,%5,%1\n\tsubb %2,%3,%0"				\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "rM" ((USItype) (ah)),					\
+	     "rM" ((USItype) (bh)),					\
+	     "rM" ((USItype) (al)),					\
+	     "rM" ((USItype) (bl)))
+#if defined (_PA_RISC1_1)
+#define umul_ppmm(w1, w0, u, v) \
+  do {									\
+    union								\
+      {									\
+	UDItype __f;							\
+	struct {USItype __w1, __w0;} __w1w0;				\
+      } __t;								\
+    __asm__ ("xmpyu %1,%2,%0"						\
+	     : "=x" (__t.__f)						\
+	     : "x" ((USItype) (u)),					\
+	       "x" ((USItype) (v)));					\
+    (w1) = __t.__w1w0.__w1;						\
+    (w0) = __t.__w1w0.__w0;						\
+     } while (0)
+#define UMUL_TIME 8
+#else
+#define UMUL_TIME 30
+#endif
+#define UDIV_TIME 40
+#define count_leading_zeros(count, x) \
+  do {									\
+    USItype __tmp;							\
+    __asm__ (								\
+       "ldi		1,%0\n"						\
+"	extru,=		%1,15,16,%%r0		; Bits 31..16 zero?\n"	\
+"	extru,tr	%1,15,16,%1		; No.  Shift down, skip add.\n"\
+"	ldo		16(%0),%0		; Yes.  Perform add.\n"	\
+"	extru,=		%1,23,8,%%r0		; Bits 15..8 zero?\n"	\
+"	extru,tr	%1,23,8,%1		; No.  Shift down, skip add.\n"\
+"	ldo		8(%0),%0		; Yes.  Perform add.\n"	\
+"	extru,=		%1,27,4,%%r0		; Bits 7..4 zero?\n"	\
+"	extru,tr	%1,27,4,%1		; No.  Shift down, skip add.\n"\
+"	ldo		4(%0),%0		; Yes.  Perform add.\n"	\
+"	extru,=		%1,29,2,%%r0		; Bits 3..2 zero?\n"	\
+"	extru,tr	%1,29,2,%1		; No.  Shift down, skip add.\n"\
+"	ldo		2(%0),%0		; Yes.  Perform add.\n"	\
+"	extru		%1,30,1,%1		; Extract bit 1.\n"	\
+"	sub		%0,%1,%0		; Subtract it.\n"	\
+	: "=r" (count), "=r" (__tmp) : "1" (x));			\
+  } while (0)
+#endif
+
+#if (defined (__i370__) || defined (__s390__) || defined (__mvs__)) && W_TYPE_SIZE == 32
+#if !defined (__zarch__)
+#define smul_ppmm(xh, xl, m0, m1) \
+  do {									\
+    union {DItype __ll;							\
+	   struct {USItype __h, __l;} __i;				\
+	  } __x;							\
+    __asm__ ("lr %N0,%1\n\tmr %0,%2"					\
+	     : "=&r" (__x.__ll)						\
+	     : "r" (m0), "r" (m1));					\
+    (xh) = __x.__i.__h; (xl) = __x.__i.__l;				\
+  } while (0)
+#define sdiv_qrnnd(q, r, n1, n0, d) \
+  do {									\
+    union {DItype __ll;							\
+	   struct {USItype __h, __l;} __i;				\
+	  } __x;							\
+    __x.__i.__h = n1; __x.__i.__l = n0;					\
+    __asm__ ("dr %0,%2"							\
+	     : "=r" (__x.__ll)						\
+	     : "0" (__x.__ll), "r" (d));				\
+    (q) = __x.__i.__l; (r) = __x.__i.__h;				\
+  } while (0)
+#else
+#define smul_ppmm(xh, xl, m0, m1) \
+  do {                                                                  \
+    register SItype __r0 __asm__ ("0");					\
+    register SItype __r1 __asm__ ("1") = (m0);				\
+									\
+    __asm__ ("mr\t%%r0,%3"                                              \
+	     : "=r" (__r0), "=r" (__r1)					\
+	     : "r"  (__r1),  "r" (m1));					\
+    (xh) = __r0; (xl) = __r1;						\
+  } while (0)
+
+#define sdiv_qrnnd(q, r, n1, n0, d) \
+  do {									\
+    register SItype __r0 __asm__ ("0") = (n1);				\
+    register SItype __r1 __asm__ ("1") = (n0);				\
+									\
+    __asm__ ("dr\t%%r0,%4"                                              \
+	     : "=r" (__r0), "=r" (__r1)					\
+	     : "r" (__r0), "r" (__r1), "r" (d));			\
+    (q) = __r1; (r) = __r0;						\
+  } while (0)
+#endif /* __zarch__ */
+#endif
+
+#if (defined (__i386__) || defined (__i486__)) && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("add{l} {%5,%1|%1,%5}\n\tadc{l} {%3,%0|%0,%3}"		\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "%0" ((USItype) (ah)),					\
+	     "g" ((USItype) (bh)),					\
+	     "%1" ((USItype) (al)),					\
+	     "g" ((USItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("sub{l} {%5,%1|%1,%5}\n\tsbb{l} {%3,%0|%0,%3}"		\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "0" ((USItype) (ah)),					\
+	     "g" ((USItype) (bh)),					\
+	     "1" ((USItype) (al)),					\
+	     "g" ((USItype) (bl)))
+#define umul_ppmm(w1, w0, u, v) \
+  __asm__ ("mul{l} %3"							\
+	   : "=a" ((USItype) (w0)),					\
+	     "=d" ((USItype) (w1))					\
+	   : "%0" ((USItype) (u)),					\
+	     "rm" ((USItype) (v)))
+#define udiv_qrnnd(q, r, n1, n0, dv) \
+  __asm__ ("div{l} %4"							\
+	   : "=a" ((USItype) (q)),					\
+	     "=d" ((USItype) (r))					\
+	   : "0" ((USItype) (n0)),					\
+	     "1" ((USItype) (n1)),					\
+	     "rm" ((USItype) (dv)))
+#define count_leading_zeros(count, x)	((count) = __builtin_clz (x))
+#define count_trailing_zeros(count, x)	((count) = __builtin_ctz (x))
+#define UMUL_TIME 40
+#define UDIV_TIME 40
+#endif /* 80x86 */
+
+#if (defined (__x86_64__) || defined (__i386__)) && W_TYPE_SIZE == 64
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("add{q} {%5,%1|%1,%5}\n\tadc{q} {%3,%0|%0,%3}"		\
+	   : "=r" ((UDItype) (sh)),					\
+	     "=&r" ((UDItype) (sl))					\
+	   : "%0" ((UDItype) (ah)),					\
+	     "rme" ((UDItype) (bh)),					\
+	     "%1" ((UDItype) (al)),					\
+	     "rme" ((UDItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("sub{q} {%5,%1|%1,%5}\n\tsbb{q} {%3,%0|%0,%3}"		\
+	   : "=r" ((UDItype) (sh)),					\
+	     "=&r" ((UDItype) (sl))					\
+	   : "0" ((UDItype) (ah)),					\
+	     "rme" ((UDItype) (bh)),					\
+	     "1" ((UDItype) (al)),					\
+	     "rme" ((UDItype) (bl)))
+#define umul_ppmm(w1, w0, u, v) \
+  __asm__ ("mul{q} %3"							\
+	   : "=a" ((UDItype) (w0)),					\
+	     "=d" ((UDItype) (w1))					\
+	   : "%0" ((UDItype) (u)),					\
+	     "rm" ((UDItype) (v)))
+#define udiv_qrnnd(q, r, n1, n0, dv) \
+  __asm__ ("div{q} %4"							\
+	   : "=a" ((UDItype) (q)),					\
+	     "=d" ((UDItype) (r))					\
+	   : "0" ((UDItype) (n0)),					\
+	     "1" ((UDItype) (n1)),					\
+	     "rm" ((UDItype) (dv)))
+#define count_leading_zeros(count, x)	((count) = __builtin_clzll (x))
+#define count_trailing_zeros(count, x)	((count) = __builtin_ctzll (x))
+#define UMUL_TIME 40
+#define UDIV_TIME 40
+#endif /* x86_64 */
+
+#if defined (__i960__) && W_TYPE_SIZE == 32
+#define umul_ppmm(w1, w0, u, v) \
+  ({union {UDItype __ll;						\
+	   struct {USItype __l, __h;} __i;				\
+	  } __xx;							\
+  __asm__ ("emul	%2,%1,%0"					\
+	   : "=d" (__xx.__ll)						\
+	   : "%dI" ((USItype) (u)),					\
+	     "dI" ((USItype) (v)));					\
+  (w1) = __xx.__i.__h; (w0) = __xx.__i.__l;})
+#define __umulsidi3(u, v) \
+  ({UDItype __w;							\
+    __asm__ ("emul	%2,%1,%0"					\
+	     : "=d" (__w)						\
+	     : "%dI" ((USItype) (u)),					\
+	       "dI" ((USItype) (v)));					\
+    __w; })
+#endif /* __i960__ */
+
+#if defined (__ia64) && W_TYPE_SIZE == 64
+/* This form encourages gcc (pre-release 3.4 at least) to emit predicated
+   "sub r=r,r" and "sub r=r,r,1", giving a 2 cycle latency.  The generic
+   code using "al<bl" arithmetically comes out making an actual 0 or 1 in a
+   register, which takes an extra cycle.  */
+#define sub_ddmmss(sh, sl, ah, al, bh, bl)				\
+  do {									\
+    UWtype __x;								\
+    __x = (al) - (bl);							\
+    if ((al) < (bl))							\
+      (sh) = (ah) - (bh) - 1;						\
+    else								\
+      (sh) = (ah) - (bh);						\
+    (sl) = __x;								\
+  } while (0)
+
+/* Do both product parts in assembly, since that gives better code with
+   all gcc versions.  Some callers will just use the upper part, and in
+   that situation we waste an instruction, but not any cycles.  */
+#define umul_ppmm(ph, pl, m0, m1)					\
+  __asm__ ("xma.hu %0 = %2, %3, f0\n\txma.l %1 = %2, %3, f0"		\
+	   : "=&f" (ph), "=f" (pl)					\
+	   : "f" (m0), "f" (m1))
+#define count_leading_zeros(count, x)					\
+  do {									\
+    UWtype _x = (x), _y, _a, _c;					\
+    __asm__ ("mux1 %0 = %1, @rev" : "=r" (_y) : "r" (_x));		\
+    __asm__ ("czx1.l %0 = %1" : "=r" (_a) : "r" (-_y | _y));		\
+    _c = (_a - 1) << 3;							\
+    _x >>= _c;								\
+    if (_x >= 1 << 4)							\
+      _x >>= 4, _c += 4;						\
+    if (_x >= 1 << 2)							\
+      _x >>= 2, _c += 2;						\
+    _c += _x >> 1;							\
+    (count) =  W_TYPE_SIZE - 1 - _c;					\
+  } while (0)
+/* similar to what gcc does for __builtin_ffs, but 0 based rather than 1
+   based, and we don't need a special case for x==0 here */
+#define count_trailing_zeros(count, x)					\
+  do {									\
+    UWtype __ctz_x = (x);						\
+    __asm__ ("popcnt %0 = %1"						\
+	     : "=r" (count)						\
+	     : "r" ((__ctz_x-1) & ~__ctz_x));				\
+  } while (0)
+#define UMUL_TIME 14
+#endif
+
+#if defined (__M32R__) && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  /* The cmp clears the condition bit.  */ \
+  __asm__ ("cmp %0,%0\n\taddx %1,%5\n\taddx %0,%3"			\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "0" ((USItype) (ah)),					\
+	     "r" ((USItype) (bh)),					\
+	     "1" ((USItype) (al)),					\
+	     "r" ((USItype) (bl))					\
+	   : "cbit")
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  /* The cmp clears the condition bit.  */ \
+  __asm__ ("cmp %0,%0\n\tsubx %1,%5\n\tsubx %0,%3"			\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "0" ((USItype) (ah)),					\
+	     "r" ((USItype) (bh)),					\
+	     "1" ((USItype) (al)),					\
+	     "r" ((USItype) (bl))					\
+	   : "cbit")
+#endif /* __M32R__ */
+
+#if defined (__mc68000__) && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("add%.l %5,%1\n\taddx%.l %3,%0"				\
+	   : "=d" ((USItype) (sh)),					\
+	     "=&d" ((USItype) (sl))					\
+	   : "%0" ((USItype) (ah)),					\
+	     "d" ((USItype) (bh)),					\
+	     "%1" ((USItype) (al)),					\
+	     "g" ((USItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("sub%.l %5,%1\n\tsubx%.l %3,%0"				\
+	   : "=d" ((USItype) (sh)),					\
+	     "=&d" ((USItype) (sl))					\
+	   : "0" ((USItype) (ah)),					\
+	     "d" ((USItype) (bh)),					\
+	     "1" ((USItype) (al)),					\
+	     "g" ((USItype) (bl)))
+
+/* The '020, '030, '040, '060 and CPU32 have 32x32->64 and 64/32->32q-32r.  */
+#if (defined (__mc68020__) && !defined (__mc68060__))
+#define umul_ppmm(w1, w0, u, v) \
+  __asm__ ("mulu%.l %3,%1:%0"						\
+	   : "=d" ((USItype) (w0)),					\
+	     "=d" ((USItype) (w1))					\
+	   : "%0" ((USItype) (u)),					\
+	     "dmi" ((USItype) (v)))
+#define UMUL_TIME 45
+#define udiv_qrnnd(q, r, n1, n0, d) \
+  __asm__ ("divu%.l %4,%1:%0"						\
+	   : "=d" ((USItype) (q)),					\
+	     "=d" ((USItype) (r))					\
+	   : "0" ((USItype) (n0)),					\
+	     "1" ((USItype) (n1)),					\
+	     "dmi" ((USItype) (d)))
+#define UDIV_TIME 90
+#define sdiv_qrnnd(q, r, n1, n0, d) \
+  __asm__ ("divs%.l %4,%1:%0"						\
+	   : "=d" ((USItype) (q)),					\
+	     "=d" ((USItype) (r))					\
+	   : "0" ((USItype) (n0)),					\
+	     "1" ((USItype) (n1)),					\
+	     "dmi" ((USItype) (d)))
+
+#elif defined (__mcoldfire__) /* not mc68020 */
+
+#define umul_ppmm(xh, xl, a, b) \
+  __asm__ ("| Inlined umul_ppmm\n"					\
+	   "	move%.l	%2,%/d0\n"					\
+	   "	move%.l	%3,%/d1\n"					\
+	   "	move%.l	%/d0,%/d2\n"					\
+	   "	swap	%/d0\n"						\
+	   "	move%.l	%/d1,%/d3\n"					\
+	   "	swap	%/d1\n"						\
+	   "	move%.w	%/d2,%/d4\n"					\
+	   "	mulu	%/d3,%/d4\n"					\
+	   "	mulu	%/d1,%/d2\n"					\
+	   "	mulu	%/d0,%/d3\n"					\
+	   "	mulu	%/d0,%/d1\n"					\
+	   "	move%.l	%/d4,%/d0\n"					\
+	   "	clr%.w	%/d0\n"						\
+	   "	swap	%/d0\n"						\
+	   "	add%.l	%/d0,%/d2\n"					\
+	   "	add%.l	%/d3,%/d2\n"					\
+	   "	jcc	1f\n"						\
+	   "	add%.l	%#65536,%/d1\n"					\
+	   "1:	swap	%/d2\n"						\
+	   "	moveq	%#0,%/d0\n"					\
+	   "	move%.w	%/d2,%/d0\n"					\
+	   "	move%.w	%/d4,%/d2\n"					\
+	   "	move%.l	%/d2,%1\n"					\
+	   "	add%.l	%/d1,%/d0\n"					\
+	   "	move%.l	%/d0,%0"					\
+	   : "=g" ((USItype) (xh)),					\
+	     "=g" ((USItype) (xl))					\
+	   : "g" ((USItype) (a)),					\
+	     "g" ((USItype) (b))					\
+	   : "d0", "d1", "d2", "d3", "d4")
+#define UMUL_TIME 100
+#define UDIV_TIME 400
+#else /* not ColdFire */
+/* %/ inserts REGISTER_PREFIX, %# inserts IMMEDIATE_PREFIX.  */
+#define umul_ppmm(xh, xl, a, b) \
+  __asm__ ("| Inlined umul_ppmm\n"					\
+	   "	move%.l	%2,%/d0\n"					\
+	   "	move%.l	%3,%/d1\n"					\
+	   "	move%.l	%/d0,%/d2\n"					\
+	   "	swap	%/d0\n"						\
+	   "	move%.l	%/d1,%/d3\n"					\
+	   "	swap	%/d1\n"						\
+	   "	move%.w	%/d2,%/d4\n"					\
+	   "	mulu	%/d3,%/d4\n"					\
+	   "	mulu	%/d1,%/d2\n"					\
+	   "	mulu	%/d0,%/d3\n"					\
+	   "	mulu	%/d0,%/d1\n"					\
+	   "	move%.l	%/d4,%/d0\n"					\
+	   "	eor%.w	%/d0,%/d0\n"					\
+	   "	swap	%/d0\n"						\
+	   "	add%.l	%/d0,%/d2\n"					\
+	   "	add%.l	%/d3,%/d2\n"					\
+	   "	jcc	1f\n"						\
+	   "	add%.l	%#65536,%/d1\n"					\
+	   "1:	swap	%/d2\n"						\
+	   "	moveq	%#0,%/d0\n"					\
+	   "	move%.w	%/d2,%/d0\n"					\
+	   "	move%.w	%/d4,%/d2\n"					\
+	   "	move%.l	%/d2,%1\n"					\
+	   "	add%.l	%/d1,%/d0\n"					\
+	   "	move%.l	%/d0,%0"					\
+	   : "=g" ((USItype) (xh)),					\
+	     "=g" ((USItype) (xl))					\
+	   : "g" ((USItype) (a)),					\
+	     "g" ((USItype) (b))					\
+	   : "d0", "d1", "d2", "d3", "d4")
+#define UMUL_TIME 100
+#define UDIV_TIME 400
+
+#endif /* not mc68020 */
+
+/* The '020, '030, '040 and '060 have bitfield insns.
+   cpu32 disguises as a 68020, but lacks them.  */
+#if defined (__mc68020__) && !defined (__mcpu32__)
+#define count_leading_zeros(count, x) \
+  __asm__ ("bfffo %1{%b2:%b2},%0"					\
+	   : "=d" ((USItype) (count))					\
+	   : "od" ((USItype) (x)), "n" (0))
+/* Some ColdFire architectures have a ff1 instruction supported via
+   __builtin_clz. */
+#elif defined (__mcfisaaplus__) || defined (__mcfisac__)
+#define count_leading_zeros(count,x) ((count) = __builtin_clz (x))
+#define COUNT_LEADING_ZEROS_0 32
+#endif
+#endif /* mc68000 */
+
+#if defined (__m88000__) && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("addu.co %1,%r4,%r5\n\taddu.ci %0,%r2,%r3"			\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "%rJ" ((USItype) (ah)),					\
+	     "rJ" ((USItype) (bh)),					\
+	     "%rJ" ((USItype) (al)),					\
+	     "rJ" ((USItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("subu.co %1,%r4,%r5\n\tsubu.ci %0,%r2,%r3"			\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "rJ" ((USItype) (ah)),					\
+	     "rJ" ((USItype) (bh)),					\
+	     "rJ" ((USItype) (al)),					\
+	     "rJ" ((USItype) (bl)))
+#define count_leading_zeros(count, x) \
+  do {									\
+    USItype __cbtmp;							\
+    __asm__ ("ff1 %0,%1"						\
+	     : "=r" (__cbtmp)						\
+	     : "r" ((USItype) (x)));					\
+    (count) = __cbtmp ^ 31;						\
+  } while (0)
+#define COUNT_LEADING_ZEROS_0 63 /* sic */
+#if defined (__mc88110__)
+#define umul_ppmm(wh, wl, u, v) \
+  do {									\
+    union {UDItype __ll;						\
+	   struct {USItype __h, __l;} __i;				\
+	  } __xx;							\
+    __asm__ ("mulu.d	%0,%1,%2"					\
+	     : "=r" (__xx.__ll)						\
+	     : "r" ((USItype) (u)),					\
+	       "r" ((USItype) (v)));					\
+    (wh) = __xx.__i.__h;						\
+    (wl) = __xx.__i.__l;						\
+  } while (0)
+#define udiv_qrnnd(q, r, n1, n0, d) \
+  ({union {UDItype __ll;						\
+	   struct {USItype __h, __l;} __i;				\
+	  } __xx;							\
+  USItype __q;								\
+  __xx.__i.__h = (n1); __xx.__i.__l = (n0);				\
+  __asm__ ("divu.d %0,%1,%2"						\
+	   : "=r" (__q)							\
+	   : "r" (__xx.__ll),						\
+	     "r" ((USItype) (d)));					\
+  (r) = (n0) - __q * (d); (q) = __q; })
+#define UMUL_TIME 5
+#define UDIV_TIME 25
+#else
+#define UMUL_TIME 17
+#define UDIV_TIME 150
+#endif /* __mc88110__ */
+#endif /* __m88000__ */
+
+#if defined (__mn10300__)
+# if defined (__AM33__)
+#  define count_leading_zeros(COUNT,X)	((COUNT) = __builtin_clz (X))
+#  define umul_ppmm(w1, w0, u, v)		\
+    asm("mulu %3,%2,%1,%0" : "=r"(w0), "=r"(w1) : "r"(u), "r"(v))
+#  define smul_ppmm(w1, w0, u, v)		\
+    asm("mul %3,%2,%1,%0" : "=r"(w0), "=r"(w1) : "r"(u), "r"(v))
+# else
+#  define umul_ppmm(w1, w0, u, v)		\
+    asm("nop; nop; mulu %3,%0" : "=d"(w0), "=z"(w1) : "%0"(u), "d"(v))
+#  define smul_ppmm(w1, w0, u, v)		\
+    asm("nop; nop; mul %3,%0" : "=d"(w0), "=z"(w1) : "%0"(u), "d"(v))
+# endif
+# define add_ssaaaa(sh, sl, ah, al, bh, bl)	\
+  do {						\
+    DWunion __s, __a, __b;			\
+    __a.s.low = (al); __a.s.high = (ah);	\
+    __b.s.low = (bl); __b.s.high = (bh);	\
+    __s.ll = __a.ll + __b.ll;			\
+    (sl) = __s.s.low; (sh) = __s.s.high;	\
+  } while (0)
+# define sub_ddmmss(sh, sl, ah, al, bh, bl)	\
+  do {						\
+    DWunion __s, __a, __b;			\
+    __a.s.low = (al); __a.s.high = (ah);	\
+    __b.s.low = (bl); __b.s.high = (bh);	\
+    __s.ll = __a.ll - __b.ll;			\
+    (sl) = __s.s.low; (sh) = __s.s.high;	\
+  } while (0)
+# define udiv_qrnnd(q, r, nh, nl, d)		\
+  asm("divu %2,%0" : "=D"(q), "=z"(r) : "D"(d), "0"(nl), "1"(nh))
+# define sdiv_qrnnd(q, r, nh, nl, d)		\
+  asm("div %2,%0" : "=D"(q), "=z"(r) : "D"(d), "0"(nl), "1"(nh))
+# define UMUL_TIME 3
+# define UDIV_TIME 38
+#endif
+
+#if defined (__mips__) && W_TYPE_SIZE == 32
+#define umul_ppmm(w1, w0, u, v)						\
+  do {									\
+    UDItype __x = (UDItype) (USItype) (u) * (USItype) (v);		\
+    (w1) = (USItype) (__x >> 32);					\
+    (w0) = (USItype) (__x);						\
+  } while (0)
+#define UMUL_TIME 10
+#define UDIV_TIME 100
+
+#if (__mips == 32 || __mips == 64) && ! __mips16
+#define count_leading_zeros(COUNT,X)	((COUNT) = __builtin_clz (X))
+#define COUNT_LEADING_ZEROS_0 32
+#endif
+#endif /* __mips__ */
+
+#if defined (__ns32000__) && W_TYPE_SIZE == 32
+#define umul_ppmm(w1, w0, u, v) \
+  ({union {UDItype __ll;						\
+	   struct {USItype __l, __h;} __i;				\
+	  } __xx;							\
+  __asm__ ("meid %2,%0"							\
+	   : "=g" (__xx.__ll)						\
+	   : "%0" ((USItype) (u)),					\
+	     "g" ((USItype) (v)));					\
+  (w1) = __xx.__i.__h; (w0) = __xx.__i.__l;})
+#define __umulsidi3(u, v) \
+  ({UDItype __w;							\
+    __asm__ ("meid %2,%0"						\
+	     : "=g" (__w)						\
+	     : "%0" ((USItype) (u)),					\
+	       "g" ((USItype) (v)));					\
+    __w; })
+#define udiv_qrnnd(q, r, n1, n0, d) \
+  ({union {UDItype __ll;						\
+	   struct {USItype __l, __h;} __i;				\
+	  } __xx;							\
+  __xx.__i.__h = (n1); __xx.__i.__l = (n0);				\
+  __asm__ ("deid %2,%0"							\
+	   : "=g" (__xx.__ll)						\
+	   : "0" (__xx.__ll),						\
+	     "g" ((USItype) (d)));					\
+  (r) = __xx.__i.__l; (q) = __xx.__i.__h; })
+#define count_trailing_zeros(count,x) \
+  do {									\
+    __asm__ ("ffsd     %2,%0"						\
+	    : "=r" ((USItype) (count))					\
+	    : "0" ((USItype) 0),					\
+	      "r" ((USItype) (x)));					\
+  } while (0)
+#endif /* __ns32000__ */
+
+/* FIXME: We should test _IBMR2 here when we add assembly support for the
+   system vendor compilers.
+   FIXME: What's needed for gcc PowerPC VxWorks?  __vxworks__ is not good
+   enough, since that hits ARM and m68k too.  */
+#if (defined (_ARCH_PPC)	/* AIX */				\
+     || defined (__powerpc__)	/* gcc */				\
+     || defined (__POWERPC__)	/* BEOS */				\
+     || defined (__ppc__)	/* Darwin */				\
+     || (defined (PPC) && ! defined (CPU_FAMILY)) /* gcc 2.7.x GNU&SysV */    \
+     || (defined (PPC) && defined (CPU_FAMILY)    /* VxWorks */               \
+	 && CPU_FAMILY == PPC)                                                \
+     ) && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  do {									\
+    if (__builtin_constant_p (bh) && (bh) == 0)				\
+      __asm__ ("add%I4c %1,%3,%4\n\taddze %0,%2"		\
+	     : "=r" (sh), "=&r" (sl) : "r" (ah), "%r" (al), "rI" (bl));\
+    else if (__builtin_constant_p (bh) && (bh) == ~(USItype) 0)		\
+      __asm__ ("add%I4c %1,%3,%4\n\taddme %0,%2"		\
+	     : "=r" (sh), "=&r" (sl) : "r" (ah), "%r" (al), "rI" (bl));\
+    else								\
+      __asm__ ("add%I5c %1,%4,%5\n\tadde %0,%2,%3"		\
+	     : "=r" (sh), "=&r" (sl)					\
+	     : "%r" (ah), "r" (bh), "%r" (al), "rI" (bl));		\
+  } while (0)
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  do {									\
+    if (__builtin_constant_p (ah) && (ah) == 0)				\
+      __asm__ ("subf%I3c %1,%4,%3\n\tsubfze %0,%2"	\
+	       : "=r" (sh), "=&r" (sl) : "r" (bh), "rI" (al), "r" (bl));\
+    else if (__builtin_constant_p (ah) && (ah) == ~(USItype) 0)		\
+      __asm__ ("subf%I3c %1,%4,%3\n\tsubfme %0,%2"	\
+	       : "=r" (sh), "=&r" (sl) : "r" (bh), "rI" (al), "r" (bl));\
+    else if (__builtin_constant_p (bh) && (bh) == 0)			\
+      __asm__ ("subf%I3c %1,%4,%3\n\taddme %0,%2"		\
+	       : "=r" (sh), "=&r" (sl) : "r" (ah), "rI" (al), "r" (bl));\
+    else if (__builtin_constant_p (bh) && (bh) == ~(USItype) 0)		\
+      __asm__ ("subf%I3c %1,%4,%3\n\taddze %0,%2"		\
+	       : "=r" (sh), "=&r" (sl) : "r" (ah), "rI" (al), "r" (bl));\
+    else								\
+      __asm__ ("subf%I4c %1,%5,%4\n\tsubfe %0,%3,%2"	\
+	       : "=r" (sh), "=&r" (sl)					\
+	       : "r" (ah), "r" (bh), "rI" (al), "r" (bl));		\
+  } while (0)
+#define count_leading_zeros(count, x) \
+  __asm__ ("cntlzw %0,%1" : "=r" (count) : "r" (x))
+#define COUNT_LEADING_ZEROS_0 32
+#if defined (_ARCH_PPC) || defined (__powerpc__) || defined (__POWERPC__) \
+  || defined (__ppc__)                                                    \
+  || (defined (PPC) && ! defined (CPU_FAMILY)) /* gcc 2.7.x GNU&SysV */       \
+  || (defined (PPC) && defined (CPU_FAMILY)    /* VxWorks */                  \
+	 && CPU_FAMILY == PPC)
+#define umul_ppmm(ph, pl, m0, m1) \
+  do {									\
+    USItype __m0 = (m0), __m1 = (m1);					\
+    __asm__ ("mulhwu %0,%1,%2" : "=r" (ph) : "%r" (m0), "r" (m1));	\
+    (pl) = __m0 * __m1;							\
+  } while (0)
+#define UMUL_TIME 15
+#define smul_ppmm(ph, pl, m0, m1) \
+  do {									\
+    SItype __m0 = (m0), __m1 = (m1);					\
+    __asm__ ("mulhw %0,%1,%2" : "=r" (ph) : "%r" (m0), "r" (m1));	\
+    (pl) = __m0 * __m1;							\
+  } while (0)
+#define SMUL_TIME 14
+#define UDIV_TIME 120
+#endif
+#endif /* 32-bit POWER architecture variants.  */
+
+/* We should test _IBMR2 here when we add assembly support for the system
+   vendor compilers.  */
+#if (defined (_ARCH_PPC64) || defined (__powerpc64__)) && W_TYPE_SIZE == 64
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  do {									\
+    if (__builtin_constant_p (bh) && (bh) == 0)				\
+      __asm__ ("add%I4c %1,%3,%4\n\taddze %0,%2"		\
+	     : "=r" (sh), "=&r" (sl) : "r" (ah), "%r" (al), "rI" (bl));\
+    else if (__builtin_constant_p (bh) && (bh) == ~(UDItype) 0)		\
+      __asm__ ("add%I4c %1,%3,%4\n\taddme %0,%2"		\
+	     : "=r" (sh), "=&r" (sl) : "r" (ah), "%r" (al), "rI" (bl));\
+    else								\
+      __asm__ ("add%I5c %1,%4,%5\n\tadde %0,%2,%3"		\
+	     : "=r" (sh), "=&r" (sl)					\
+	     : "%r" (ah), "r" (bh), "%r" (al), "rI" (bl));		\
+  } while (0)
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  do {									\
+    if (__builtin_constant_p (ah) && (ah) == 0)				\
+      __asm__ ("subf%I3c %1,%4,%3\n\tsubfze %0,%2"	\
+	       : "=r" (sh), "=&r" (sl) : "r" (bh), "rI" (al), "r" (bl));\
+    else if (__builtin_constant_p (ah) && (ah) == ~(UDItype) 0)		\
+      __asm__ ("subf%I3c %1,%4,%3\n\tsubfme %0,%2"	\
+	       : "=r" (sh), "=&r" (sl) : "r" (bh), "rI" (al), "r" (bl));\
+    else if (__builtin_constant_p (bh) && (bh) == 0)			\
+      __asm__ ("subf%I3c %1,%4,%3\n\taddme %0,%2"		\
+	       : "=r" (sh), "=&r" (sl) : "r" (ah), "rI" (al), "r" (bl));\
+    else if (__builtin_constant_p (bh) && (bh) == ~(UDItype) 0)		\
+      __asm__ ("subf%I3c %1,%4,%3\n\taddze %0,%2"		\
+	       : "=r" (sh), "=&r" (sl) : "r" (ah), "rI" (al), "r" (bl));\
+    else								\
+      __asm__ ("subf%I4c %1,%5,%4\n\tsubfe %0,%3,%2"	\
+	       : "=r" (sh), "=&r" (sl)					\
+	       : "r" (ah), "r" (bh), "rI" (al), "r" (bl));		\
+  } while (0)
+#define count_leading_zeros(count, x) \
+  __asm__ ("cntlzd %0,%1" : "=r" (count) : "r" (x))
+#define COUNT_LEADING_ZEROS_0 64
+#define umul_ppmm(ph, pl, m0, m1) \
+  do {									\
+    UDItype __m0 = (m0), __m1 = (m1);					\
+    __asm__ ("mulhdu %0,%1,%2" : "=r" (ph) : "%r" (m0), "r" (m1));	\
+    (pl) = __m0 * __m1;							\
+  } while (0)
+#define UMUL_TIME 15
+#define smul_ppmm(ph, pl, m0, m1) \
+  do {									\
+    DItype __m0 = (m0), __m1 = (m1);					\
+    __asm__ ("mulhd %0,%1,%2" : "=r" (ph) : "%r" (m0), "r" (m1));	\
+    (pl) = __m0 * __m1;							\
+  } while (0)
+#define SMUL_TIME 14  /* ??? */
+#define UDIV_TIME 120 /* ??? */
+#endif /* 64-bit PowerPC.  */
+
+#if defined (__ibm032__) /* RT/ROMP */ && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("a %1,%5\n\tae %0,%3"					\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "%0" ((USItype) (ah)),					\
+	     "r" ((USItype) (bh)),					\
+	     "%1" ((USItype) (al)),					\
+	     "r" ((USItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("s %1,%5\n\tse %0,%3"					\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "0" ((USItype) (ah)),					\
+	     "r" ((USItype) (bh)),					\
+	     "1" ((USItype) (al)),					\
+	     "r" ((USItype) (bl)))
+#define umul_ppmm(ph, pl, m0, m1) \
+  do {									\
+    USItype __m0 = (m0), __m1 = (m1);					\
+    __asm__ (								\
+       "s	r2,r2\n"						\
+"	mts	r10,%2\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	m	r2,%3\n"						\
+"	cas	%0,r2,r0\n"						\
+"	mfs	r10,%1"							\
+	     : "=r" ((USItype) (ph)),					\
+	       "=r" ((USItype) (pl))					\
+	     : "%r" (__m0),						\
+		"r" (__m1)						\
+	     : "r2");							\
+    (ph) += ((((SItype) __m0 >> 31) & __m1)				\
+	     + (((SItype) __m1 >> 31) & __m0));				\
+  } while (0)
+#define UMUL_TIME 20
+#define UDIV_TIME 200
+#define count_leading_zeros(count, x) \
+  do {									\
+    if ((x) >= 0x10000)							\
+      __asm__ ("clz	%0,%1"						\
+	       : "=r" ((USItype) (count))				\
+	       : "r" ((USItype) (x) >> 16));				\
+    else								\
+      {									\
+	__asm__ ("clz	%0,%1"						\
+		 : "=r" ((USItype) (count))				\
+		 : "r" ((USItype) (x)));					\
+	(count) += 16;							\
+      }									\
+  } while (0)
+#endif
+
+#if defined(__sh__) && !__SHMEDIA__ && W_TYPE_SIZE == 32
+#ifndef __sh1__
+#define umul_ppmm(w1, w0, u, v) \
+  __asm__ (								\
+       "dmulu.l	%2,%3\n\tsts%M1	macl,%1\n\tsts%M0	mach,%0"	\
+	   : "=r<" ((USItype)(w1)),					\
+	     "=r<" ((USItype)(w0))					\
+	   : "r" ((USItype)(u)),					\
+	     "r" ((USItype)(v))						\
+	   : "macl", "mach")
+#define UMUL_TIME 5
+#endif
+
+/* This is the same algorithm as __udiv_qrnnd_c.  */
+#define UDIV_NEEDS_NORMALIZATION 1
+
+#define udiv_qrnnd(q, r, n1, n0, d) \
+  do {									\
+    extern UWtype __udiv_qrnnd_16 (UWtype, UWtype)			\
+			__attribute__ ((visibility ("hidden")));	\
+    /* r0: rn r1: qn */ /* r0: n1 r4: n0 r5: d r6: d1 */ /* r2: __m */	\
+    __asm__ (								\
+	"mov%M4 %4,r5\n"						\
+"	swap.w %3,r4\n"							\
+"	swap.w r5,r6\n"							\
+"	jsr @%5\n"							\
+"	shll16 r6\n"							\
+"	swap.w r4,r4\n"							\
+"	jsr @%5\n"							\
+"	swap.w r1,%0\n"							\
+"	or r1,%0"							\
+	: "=r" (q), "=&z" (r)						\
+	: "1" (n1), "r" (n0), "rm" (d), "r" (&__udiv_qrnnd_16)		\
+	: "r1", "r2", "r4", "r5", "r6", "pr", "t");			\
+  } while (0)
+
+#define UDIV_TIME 80
+
+#define sub_ddmmss(sh, sl, ah, al, bh, bl)				\
+  __asm__ ("clrt;subc %5,%1; subc %4,%0"				\
+	   : "=r" (sh), "=r" (sl)					\
+	   : "0" (ah), "1" (al), "r" (bh), "r" (bl) : "t")
+
+#endif /* __sh__ */
+
+#if defined (__SH5__) && __SHMEDIA__ && W_TYPE_SIZE == 32
+#define __umulsidi3(u,v) ((UDItype)(USItype)u*(USItype)v)
+#define count_leading_zeros(count, x) \
+  do									\
+    {									\
+      UDItype x_ = (USItype)(x);					\
+      SItype c_;							\
+									\
+      __asm__ ("nsb %1, %0" : "=r" (c_) : "r" (x_));			\
+      (count) = c_ - 31;						\
+    }									\
+  while (0)
+#define COUNT_LEADING_ZEROS_0 32
+#endif
+
+#if defined (__sparc__) && !defined (__arch64__) && !defined (__sparcv9) \
+    && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("addcc %r4,%5,%1\n\taddx %r2,%3,%0"				\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "%rJ" ((USItype) (ah)),					\
+	     "rI" ((USItype) (bh)),					\
+	     "%rJ" ((USItype) (al)),					\
+	     "rI" ((USItype) (bl))					\
+	   __CLOBBER_CC)
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("subcc %r4,%5,%1\n\tsubx %r2,%3,%0"				\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "rJ" ((USItype) (ah)),					\
+	     "rI" ((USItype) (bh)),					\
+	     "rJ" ((USItype) (al)),					\
+	     "rI" ((USItype) (bl))					\
+	   __CLOBBER_CC)
+#if defined (__sparc_v9__)
+#define umul_ppmm(w1, w0, u, v) \
+  do {									\
+    register USItype __g1 asm ("g1");					\
+    __asm__ ("umul\t%2,%3,%1\n\t"					\
+	     "srlx\t%1, 32, %0"						\
+	     : "=r" ((USItype) (w1)),					\
+	       "=r" (__g1)						\
+	     : "r" ((USItype) (u)),					\
+	       "r" ((USItype) (v)));					\
+    (w0) = __g1;							\
+  } while (0)
+#define udiv_qrnnd(__q, __r, __n1, __n0, __d) \
+  __asm__ ("mov\t%2,%%y\n\t"						\
+	   "udiv\t%3,%4,%0\n\t"						\
+	   "umul\t%0,%4,%1\n\t"						\
+	   "sub\t%3,%1,%1"						\
+	   : "=&r" ((USItype) (__q)),					\
+	     "=&r" ((USItype) (__r))					\
+	   : "r" ((USItype) (__n1)),					\
+	     "r" ((USItype) (__n0)),					\
+	     "r" ((USItype) (__d)))
+#else
+#if defined (__sparc_v8__)
+#define umul_ppmm(w1, w0, u, v) \
+  __asm__ ("umul %2,%3,%1;rd %%y,%0"					\
+	   : "=r" ((USItype) (w1)),					\
+	     "=r" ((USItype) (w0))					\
+	   : "r" ((USItype) (u)),					\
+	     "r" ((USItype) (v)))
+#define udiv_qrnnd(__q, __r, __n1, __n0, __d) \
+  __asm__ ("mov %2,%%y;nop;nop;nop;udiv %3,%4,%0;umul %0,%4,%1;sub %3,%1,%1"\
+	   : "=&r" ((USItype) (__q)),					\
+	     "=&r" ((USItype) (__r))					\
+	   : "r" ((USItype) (__n1)),					\
+	     "r" ((USItype) (__n0)),					\
+	     "r" ((USItype) (__d)))
+#else
+#if defined (__sparclite__)
+/* This has hardware multiply but not divide.  It also has two additional
+   instructions scan (ffs from high bit) and divscc.  */
+#define umul_ppmm(w1, w0, u, v) \
+  __asm__ ("umul %2,%3,%1;rd %%y,%0"					\
+	   : "=r" ((USItype) (w1)),					\
+	     "=r" ((USItype) (w0))					\
+	   : "r" ((USItype) (u)),					\
+	     "r" ((USItype) (v)))
+#define udiv_qrnnd(q, r, n1, n0, d) \
+  __asm__ ("! Inlined udiv_qrnnd\n"					\
+"	wr	%%g0,%2,%%y	! Not a delayed write for sparclite\n"	\
+"	tst	%%g0\n"							\
+"	divscc	%3,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%%g1\n"						\
+"	divscc	%%g1,%4,%0\n"						\
+"	rd	%%y,%1\n"						\
+"	bl,a 1f\n"							\
+"	add	%1,%4,%1\n"						\
+"1:	! End of inline udiv_qrnnd"					\
+	   : "=r" ((USItype) (q)),					\
+	     "=r" ((USItype) (r))					\
+	   : "r" ((USItype) (n1)),					\
+	     "r" ((USItype) (n0)),					\
+	     "rI" ((USItype) (d))					\
+	   : "g1" __AND_CLOBBER_CC)
+#define UDIV_TIME 37
+#define count_leading_zeros(count, x) \
+  do {                                                                  \
+  __asm__ ("scan %1,1,%0"                                               \
+	   : "=r" ((USItype) (count))                                   \
+	   : "r" ((USItype) (x)));					\
+  } while (0)
+/* Early sparclites return 63 for an argument of 0, but they warn that future
+   implementations might change this.  Therefore, leave COUNT_LEADING_ZEROS_0
+   undefined.  */
+#else
+/* SPARC without integer multiplication and divide instructions.
+   (i.e. at least Sun4/20,40,60,65,75,110,260,280,330,360,380,470,490) */
+#define umul_ppmm(w1, w0, u, v) \
+  __asm__ ("! Inlined umul_ppmm\n"					\
+"	wr	%%g0,%2,%%y	! SPARC has 0-3 delay insn after a wr\n"\
+"	sra	%3,31,%%o5	! Don't move this insn\n"		\
+"	and	%2,%%o5,%%o5	! Don't move this insn\n"		\
+"	andcc	%%g0,0,%%g1	! Don't move this insn\n"		\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,%3,%%g1\n"						\
+"	mulscc	%%g1,0,%%g1\n"						\
+"	add	%%g1,%%o5,%0\n"						\
+"	rd	%%y,%1"							\
+	   : "=r" ((USItype) (w1)),					\
+	     "=r" ((USItype) (w0))					\
+	   : "%rI" ((USItype) (u)),					\
+	     "r" ((USItype) (v))						\
+	   : "g1", "o5" __AND_CLOBBER_CC)
+#define UMUL_TIME 39		/* 39 instructions */
+/* It's quite necessary to add this much assembler for the sparc.
+   The default udiv_qrnnd (in C) is more than 10 times slower!  */
+#define udiv_qrnnd(__q, __r, __n1, __n0, __d) \
+  __asm__ ("! Inlined udiv_qrnnd\n"					\
+"	mov	32,%%g1\n"						\
+"	subcc	%1,%2,%%g0\n"						\
+"1:	bcs	5f\n"							\
+"	 addxcc %0,%0,%0	! shift n1n0 and a q-bit in lsb\n"	\
+"	sub	%1,%2,%1	! this kills msb of n\n"		\
+"	addx	%1,%1,%1	! so this can't give carry\n"		\
+"	subcc	%%g1,1,%%g1\n"						\
+"2:	bne	1b\n"							\
+"	 subcc	%1,%2,%%g0\n"						\
+"	bcs	3f\n"							\
+"	 addxcc %0,%0,%0	! shift n1n0 and a q-bit in lsb\n"	\
+"	b	3f\n"							\
+"	 sub	%1,%2,%1	! this kills msb of n\n"		\
+"4:	sub	%1,%2,%1\n"						\
+"5:	addxcc	%1,%1,%1\n"						\
+"	bcc	2b\n"							\
+"	 subcc	%%g1,1,%%g1\n"						\
+"! Got carry from n.  Subtract next step to cancel this carry.\n"	\
+"	bne	4b\n"							\
+"	 addcc	%0,%0,%0	! shift n1n0 and a 0-bit in lsb\n"	\
+"	sub	%1,%2,%1\n"						\
+"3:	xnor	%0,0,%0\n"						\
+"	! End of inline udiv_qrnnd"					\
+	   : "=&r" ((USItype) (__q)),					\
+	     "=&r" ((USItype) (__r))					\
+	   : "r" ((USItype) (__d)),					\
+	     "1" ((USItype) (__n1)),					\
+	     "0" ((USItype) (__n0)) : "g1" __AND_CLOBBER_CC)
+#define UDIV_TIME (3+7*32)	/* 7 instructions/iteration. 32 iterations.  */
+#endif /* __sparclite__ */
+#endif /* __sparc_v8__ */
+#endif /* __sparc_v9__ */
+#endif /* sparc32 */
+
+#if ((defined (__sparc__) && defined (__arch64__)) || defined (__sparcv9)) \
+    && W_TYPE_SIZE == 64
+#define add_ssaaaa(sh, sl, ah, al, bh, bl)				\
+  do {									\
+    UDItype __carry = 0;						\
+    __asm__ ("addcc\t%r5,%6,%1\n\t"					\
+	     "add\t%r3,%4,%0\n\t"					\
+	     "movcs\t%%xcc, 1, %2\n\t"					\
+	     "add\t%0, %2, %0"						\
+	     : "=r" ((UDItype)(sh)),				      	\
+	       "=&r" ((UDItype)(sl)),				      	\
+	       "+r" (__carry)				      		\
+	     : "%rJ" ((UDItype)(ah)),				     	\
+	       "rI" ((UDItype)(bh)),				      	\
+	       "%rJ" ((UDItype)(al)),				     	\
+	       "rI" ((UDItype)(bl))				       	\
+	     __CLOBBER_CC);						\
+  } while (0)
+
+#define sub_ddmmss(sh, sl, ah, al, bh, bl)				\
+  do {									\
+    UDItype __carry = 0;						\
+    __asm__ ("subcc\t%r5,%6,%1\n\t"					\
+	     "sub\t%r3,%4,%0\n\t"					\
+	     "movcs\t%%xcc, 1, %2\n\t"					\
+	     "sub\t%0, %2, %0"						\
+	     : "=r" ((UDItype)(sh)),				      	\
+	       "=&r" ((UDItype)(sl)),				      	\
+	       "+r" (__carry)				      		\
+	     : "%rJ" ((UDItype)(ah)),				     	\
+	       "rI" ((UDItype)(bh)),				      	\
+	       "%rJ" ((UDItype)(al)),				     	\
+	       "rI" ((UDItype)(bl))				       	\
+	     __CLOBBER_CC);						\
+  } while (0)
+
+#define umul_ppmm(wh, wl, u, v)						\
+  do {									\
+	  UDItype tmp1, tmp2, tmp3, tmp4;				\
+	  __asm__ __volatile__ (					\
+		   "srl %7,0,%3\n\t"					\
+		   "mulx %3,%6,%1\n\t"					\
+		   "srlx %6,32,%2\n\t"					\
+		   "mulx %2,%3,%4\n\t"					\
+		   "sllx %4,32,%5\n\t"					\
+		   "srl %6,0,%3\n\t"					\
+		   "sub %1,%5,%5\n\t"					\
+		   "srlx %5,32,%5\n\t"					\
+		   "addcc %4,%5,%4\n\t"					\
+		   "srlx %7,32,%5\n\t"					\
+		   "mulx %3,%5,%3\n\t"					\
+		   "mulx %2,%5,%5\n\t"					\
+		   "sethi %%hi(0x80000000),%2\n\t"			\
+		   "addcc %4,%3,%4\n\t"					\
+		   "srlx %4,32,%4\n\t"					\
+		   "add %2,%2,%2\n\t"					\
+		   "movcc %%xcc,%%g0,%2\n\t"				\
+		   "addcc %5,%4,%5\n\t"					\
+		   "sllx %3,32,%3\n\t"					\
+		   "add %1,%3,%1\n\t"					\
+		   "add %5,%2,%0"					\
+	   : "=r" ((UDItype)(wh)),					\
+	     "=&r" ((UDItype)(wl)),					\
+	     "=&r" (tmp1), "=&r" (tmp2), "=&r" (tmp3), "=&r" (tmp4)	\
+	   : "r" ((UDItype)(u)),					\
+	     "r" ((UDItype)(v))						\
+	   __CLOBBER_CC);						\
+  } while (0)
+#define UMUL_TIME 96
+#define UDIV_TIME 230
+#endif /* sparc64 */
+
+#if defined (__vax__) && W_TYPE_SIZE == 32
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("addl2 %5,%1\n\tadwc %3,%0"					\
+	   : "=g" ((USItype) (sh)),					\
+	     "=&g" ((USItype) (sl))					\
+	   : "%0" ((USItype) (ah)),					\
+	     "g" ((USItype) (bh)),					\
+	     "%1" ((USItype) (al)),					\
+	     "g" ((USItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("subl2 %5,%1\n\tsbwc %3,%0"					\
+	   : "=g" ((USItype) (sh)),					\
+	     "=&g" ((USItype) (sl))					\
+	   : "0" ((USItype) (ah)),					\
+	     "g" ((USItype) (bh)),					\
+	     "1" ((USItype) (al)),					\
+	     "g" ((USItype) (bl)))
+#define umul_ppmm(xh, xl, m0, m1) \
+  do {									\
+    union {								\
+	UDItype __ll;							\
+	struct {USItype __l, __h;} __i;					\
+      } __xx;								\
+    USItype __m0 = (m0), __m1 = (m1);					\
+    __asm__ ("emul %1,%2,$0,%0"						\
+	     : "=r" (__xx.__ll)						\
+	     : "g" (__m0),						\
+	       "g" (__m1));						\
+    (xh) = __xx.__i.__h;						\
+    (xl) = __xx.__i.__l;						\
+    (xh) += ((((SItype) __m0 >> 31) & __m1)				\
+	     + (((SItype) __m1 >> 31) & __m0));				\
+  } while (0)
+#define sdiv_qrnnd(q, r, n1, n0, d) \
+  do {									\
+    union {DItype __ll;							\
+	   struct {SItype __l, __h;} __i;				\
+	  } __xx;							\
+    __xx.__i.__h = n1; __xx.__i.__l = n0;				\
+    __asm__ ("ediv %3,%2,%0,%1"						\
+	     : "=g" (q), "=g" (r)					\
+	     : "g" (__xx.__ll), "g" (d));				\
+  } while (0)
+#endif /* __vax__ */
+
+#ifdef _TMS320C6X
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  do									\
+    {									\
+      UDItype __ll;							\
+      __asm__ ("addu .l1 %1, %2, %0"					\
+	       : "=a" (__ll) : "a" (al), "a" (bl));			\
+      (sl) = (USItype)__ll;						\
+      (sh) = ((USItype)(__ll >> 32)) + (ah) + (bh);			\
+    }									\
+  while (0)
+
+#ifdef _TMS320C6400_PLUS
+#define __umulsidi3(u,v) ((UDItype)(USItype)u*(USItype)v)
+#define umul_ppmm(w1, w0, u, v)						\
+  do {									\
+    UDItype __x = (UDItype) (USItype) (u) * (USItype) (v);		\
+    (w1) = (USItype) (__x >> 32);					\
+    (w0) = (USItype) (__x);						\
+  } while (0)
+#endif  /* _TMS320C6400_PLUS */
+
+#define count_leading_zeros(count, x)	((count) = __builtin_clz (x))
+#ifdef _TMS320C6400
+#define count_trailing_zeros(count, x)	((count) = __builtin_ctz (x))
+#endif
+#define UMUL_TIME 4
+#define UDIV_TIME 40
+#endif /* _TMS320C6X */
+
+#if defined (__xtensa__) && W_TYPE_SIZE == 32
+/* This code is not Xtensa-configuration-specific, so rely on the compiler
+   to expand builtin functions depending on what configuration features
+   are available.  This avoids library calls when the operation can be
+   performed in-line.  */
+#define umul_ppmm(w1, w0, u, v)						\
+  do {									\
+    DWunion __w;							\
+    __w.ll = __builtin_umulsidi3 (u, v);				\
+    w1 = __w.s.high;							\
+    w0 = __w.s.low;							\
+  } while (0)
+#define __umulsidi3(u, v)		__builtin_umulsidi3 (u, v)
+#define count_leading_zeros(COUNT, X)	((COUNT) = __builtin_clz (X))
+#define count_trailing_zeros(COUNT, X)	((COUNT) = __builtin_ctz (X))
+#endif /* __xtensa__ */
+
+#if defined xstormy16
+extern UHItype __stormy16_count_leading_zeros (UHItype);
+#define count_leading_zeros(count, x)					\
+  do									\
+    {									\
+      UHItype size;							\
+									\
+      /* We assume that W_TYPE_SIZE is a multiple of 16...  */		\
+      for ((count) = 0, size = W_TYPE_SIZE; size; size -= 16)		\
+	{								\
+	  UHItype c;							\
+									\
+	  c = __clzhi2 ((x) >> (size - 16));				\
+	  (count) += c;							\
+	  if (c != 16)							\
+	    break;							\
+	}								\
+    }									\
+  while (0)
+#define COUNT_LEADING_ZEROS_0 W_TYPE_SIZE
+#endif
+
+#if defined (__z8000__) && W_TYPE_SIZE == 16
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("add	%H1,%H5\n\tadc	%H0,%H3"				\
+	   : "=r" ((unsigned int)(sh)),					\
+	     "=&r" ((unsigned int)(sl))					\
+	   : "%0" ((unsigned int)(ah)),					\
+	     "r" ((unsigned int)(bh)),					\
+	     "%1" ((unsigned int)(al)),					\
+	     "rQR" ((unsigned int)(bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("sub	%H1,%H5\n\tsbc	%H0,%H3"				\
+	   : "=r" ((unsigned int)(sh)),					\
+	     "=&r" ((unsigned int)(sl))					\
+	   : "0" ((unsigned int)(ah)),					\
+	     "r" ((unsigned int)(bh)),					\
+	     "1" ((unsigned int)(al)),					\
+	     "rQR" ((unsigned int)(bl)))
+#define umul_ppmm(xh, xl, m0, m1) \
+  do {									\
+    union {long int __ll;						\
+	   struct {unsigned int __h, __l;} __i;				\
+	  } __xx;							\
+    unsigned int __m0 = (m0), __m1 = (m1);				\
+    __asm__ ("mult	%S0,%H3"					\
+	     : "=r" (__xx.__i.__h),					\
+	       "=r" (__xx.__i.__l)					\
+	     : "%1" (__m0),						\
+	       "rQR" (__m1));						\
+    (xh) = __xx.__i.__h; (xl) = __xx.__i.__l;				\
+    (xh) += ((((signed int) __m0 >> 15) & __m1)				\
+	     + (((signed int) __m1 >> 15) & __m0));			\
+  } while (0)
+#endif /* __z8000__ */
+
+#endif /* __GNUC__ */
+
+/* If this machine has no inline assembler, use C macros.  */
+
+#if !defined (add_ssaaaa)
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  do {									\
+    UWtype __x;								\
+    __x = (al) + (bl);							\
+    (sh) = (ah) + (bh) + (__x < (al));					\
+    (sl) = __x;								\
+  } while (0)
+#endif
+
+#if !defined (sub_ddmmss)
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  do {									\
+    UWtype __x;								\
+    __x = (al) - (bl);							\
+    (sh) = (ah) - (bh) - (__x > (al));					\
+    (sl) = __x;								\
+  } while (0)
+#endif
+
+/* If we lack umul_ppmm but have smul_ppmm, define umul_ppmm in terms of
+   smul_ppmm.  */
+#if !defined (umul_ppmm) && defined (smul_ppmm)
+#define umul_ppmm(w1, w0, u, v)						\
+  do {									\
+    UWtype __w1;							\
+    UWtype __xm0 = (u), __xm1 = (v);					\
+    smul_ppmm (__w1, w0, __xm0, __xm1);					\
+    (w1) = __w1 + (-(__xm0 >> (W_TYPE_SIZE - 1)) & __xm1)		\
+		+ (-(__xm1 >> (W_TYPE_SIZE - 1)) & __xm0);		\
+  } while (0)
+#endif
+
+/* If we still don't have umul_ppmm, define it using plain C.  */
+#if !defined (umul_ppmm)
+#define umul_ppmm(w1, w0, u, v)						\
+  do {									\
+    UWtype __x0, __x1, __x2, __x3;					\
+    UHWtype __ul, __vl, __uh, __vh;					\
+									\
+    __ul = __ll_lowpart (u);						\
+    __uh = __ll_highpart (u);						\
+    __vl = __ll_lowpart (v);						\
+    __vh = __ll_highpart (v);						\
+									\
+    __x0 = (UWtype) __ul * __vl;					\
+    __x1 = (UWtype) __ul * __vh;					\
+    __x2 = (UWtype) __uh * __vl;					\
+    __x3 = (UWtype) __uh * __vh;					\
+									\
+    __x1 += __ll_highpart (__x0);/* this can't give carry */		\
+    __x1 += __x2;		/* but this indeed can */		\
+    if (__x1 < __x2)		/* did we get it? */			\
+      __x3 += __ll_B;		/* yes, add it in the proper pos.  */	\
+									\
+    (w1) = __x3 + __ll_highpart (__x1);					\
+    (w0) = __ll_lowpart (__x1) * __ll_B + __ll_lowpart (__x0);		\
+  } while (0)
+#endif
+
+#if !defined (__umulsidi3)
+#define __umulsidi3(u, v) \
+  ({DWunion __w;							\
+    umul_ppmm (__w.s.high, __w.s.low, u, v);				\
+    __w.ll; })
+#endif
+
+/* Define this unconditionally, so it can be used for debugging.  */
+#define __udiv_qrnnd_c(q, r, n1, n0, d) \
+  do {									\
+    UWtype __d1, __d0, __q1, __q0;					\
+    UWtype __r1, __r0, __m;						\
+    __d1 = __ll_highpart (d);						\
+    __d0 = __ll_lowpart (d);						\
+									\
+    __r1 = (n1) % __d1;							\
+    __q1 = (n1) / __d1;							\
+    __m = (UWtype) __q1 * __d0;						\
+    __r1 = __r1 * __ll_B | __ll_highpart (n0);				\
+    if (__r1 < __m)							\
+      {									\
+	__q1--, __r1 += (d);						\
+	if (__r1 >= (d)) /* i.e. we didn't get carry when adding to __r1 */\
+	  if (__r1 < __m)						\
+	    __q1--, __r1 += (d);					\
+      }									\
+    __r1 -= __m;							\
+									\
+    __r0 = __r1 % __d1;							\
+    __q0 = __r1 / __d1;							\
+    __m = (UWtype) __q0 * __d0;						\
+    __r0 = __r0 * __ll_B | __ll_lowpart (n0);				\
+    if (__r0 < __m)							\
+      {									\
+	__q0--, __r0 += (d);						\
+	if (__r0 >= (d))						\
+	  if (__r0 < __m)						\
+	    __q0--, __r0 += (d);					\
+      }									\
+    __r0 -= __m;							\
+									\
+    (q) = (UWtype) __q1 * __ll_B | __q0;				\
+    (r) = __r0;								\
+  } while (0)
+
+/* If the processor has no udiv_qrnnd but sdiv_qrnnd, go through
+   __udiv_w_sdiv (defined in libgcc or elsewhere).  */
+#if !defined (udiv_qrnnd) && defined (sdiv_qrnnd)
+#define udiv_qrnnd(q, r, nh, nl, d) \
+  do {									\
+    USItype __r;							\
+    (q) = __udiv_w_sdiv (&__r, nh, nl, d);				\
+    (r) = __r;								\
+  } while (0)
+#endif
+
+/* If udiv_qrnnd was not defined for this processor, use __udiv_qrnnd_c.  */
+#if !defined (udiv_qrnnd)
+#define UDIV_NEEDS_NORMALIZATION 1
+#define udiv_qrnnd __udiv_qrnnd_c
+#endif
+
+#if !defined (count_leading_zeros)
+#define count_leading_zeros(count, x) \
+  do {									\
+    UWtype __xr = (x);							\
+    UWtype __a;								\
+									\
+    if (W_TYPE_SIZE <= 32)						\
+      {									\
+	__a = __xr < ((UWtype)1<<2*__BITS4)				\
+	  ? (__xr < ((UWtype)1<<__BITS4) ? 0 : __BITS4)			\
+	  : (__xr < ((UWtype)1<<3*__BITS4) ?  2*__BITS4 : 3*__BITS4);	\
+      }									\
+    else								\
+      {									\
+	for (__a = W_TYPE_SIZE - 8; __a > 0; __a -= 8)			\
+	  if (((__xr >> __a) & 0xff) != 0)				\
+	    break;							\
+      }									\
+									\
+    (count) = W_TYPE_SIZE - (__clz_tab[__xr >> __a] + __a);		\
+  } while (0)
+#define COUNT_LEADING_ZEROS_0 W_TYPE_SIZE
+#endif
+
+#if !defined (count_trailing_zeros)
+/* Define count_trailing_zeros using count_leading_zeros.  The latter might be
+   defined in asm, but if it is not, the C version above is good enough.  */
+#define count_trailing_zeros(count, x) \
+  do {									\
+    UWtype __ctz_x = (x);						\
+    UWtype __ctz_c;							\
+    count_leading_zeros (__ctz_c, __ctz_x & -__ctz_x);			\
+    (count) = W_TYPE_SIZE - 1 - __ctz_c;				\
+  } while (0)
+#endif
+
+#ifndef UDIV_NEEDS_NORMALIZATION
+#define UDIV_NEEDS_NORMALIZATION 0
+#endif
Index: gcc/wide-int.cc
===================================================================
--- gcc/wide-int.cc	2013-11-30 16:12:28.936850266 +0000
+++ gcc/wide-int.cc	2013-12-01 10:26:10.387979913 +0000
@@ -27,6 +27,16 @@ along with GCC; see the file COPYING3.
 #include "tree.h"
 #include "dumpfile.h"
 
+#if GCC_VERSION >= 3000
+#define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
+typedef unsigned HOST_HALF_WIDE_INT UHWtype;
+typedef unsigned HOST_WIDE_INT UWtype;
+typedef unsigned int UQItype __attribute__ ((mode (QI)));
+typedef unsigned int USItype __attribute__ ((mode (SI)));
+typedef unsigned int UDItype __attribute__ ((mode (DI)));
+#include "longlong.h"
+#endif
+
 /* This is the maximal size of the buffer needed for dump.  */
 const unsigned int MAX_SIZE = (4 * (MAX_BITSIZE_MODE_ANY_INT / 4
 				    + (MAX_BITSIZE_MODE_ANY_INT
@@ -1259,8 +1269,8 @@ wi_pack (unsigned HOST_WIDE_INT *result,
    controls the signedness and is used to check overflow or if HIGH or
    FULL is set.  */
 unsigned int
-wi::mul_internal (HOST_WIDE_INT *val, const HOST_WIDE_INT *op1,
-		  unsigned int op1len, const HOST_WIDE_INT *op2,
+wi::mul_internal (HOST_WIDE_INT *val, const HOST_WIDE_INT *op1val,
+		  unsigned int op1len, const HOST_WIDE_INT *op2val,
 		  unsigned int op2len, unsigned int prec, signop sgn,
 		  bool *overflow, bool high, bool full)
 {
@@ -1289,24 +1299,53 @@ wi::mul_internal (HOST_WIDE_INT *val, co
   if (needs_overflow)
     *overflow = false;
 
+  wide_int_ref op1 = wi::storage_ref (op1val, op1len, prec);
+  wide_int_ref op2 = wi::storage_ref (op2val, op2len, prec);
+
   /* This is a surprisingly common case, so do it first.  */
-  if ((op1len == 1 && op1[0] == 0) || (op2len == 1 && op2[0] == 0))
+  if (op1 == 0 || op2 == 0)
     {
       val[0] = 0;
       return 1;
     }
 
+#ifdef umul_ppmm
+  if (sgn == UNSIGNED)
+    {
+      /* If the inputs are single HWIs and the output has room for at
+	 least two HWIs, we can use umul_ppmm directly.  */
+      if (prec >= HOST_BITS_PER_WIDE_INT * 2
+	  && wi::fits_uhwi_p (op1)
+	  && wi::fits_uhwi_p (op2))
+	{
+	  umul_ppmm (val[1], val[0], op1.ulow (), op2.ulow ());
+	  return 1 + (val[1] != 0 || val[0] < 0);
+	}
+      /* Likewise if the output is a full single HWI, except that the
+	 upper HWI of the result is only used for determining overflow.
+	 (We handle this case inline when overflow isn't needed.)  */
+      else if (prec == HOST_BITS_PER_WIDE_INT)
+	{
+	  unsigned HOST_WIDE_INT upper;
+	  umul_ppmm (upper, val[0], op1.ulow (), op2.ulow ());
+	  if (needs_overflow)
+	    *overflow = (upper != 0);
+	  return 1;
+	}
+    }
+#endif
+
   /* Handle multiplications by 1.  */
-  if (op1len == 1 && op1[0] == 1)
+  if (op1 == 1)
     {
       for (i = 0; i < op2len; i++)
-	val[i] = op2[i];
+	val[i] = op2val[i];
       return op2len;
     }
-  if (op2len == 1 && op2[0] == 1)
+  if (op2 == 1)
     {
       for (i = 0; i < op1len; i++)
-	val[i] = op1[i];
+	val[i] = op1val[i];
       return op1len;
     }
 
@@ -1320,13 +1359,13 @@ wi::mul_internal (HOST_WIDE_INT *val, co
 
       if (sgn == SIGNED)
 	{
-	  o0 = sext_hwi (op1[0], prec);
-	  o1 = sext_hwi (op2[0], prec);
+	  o0 = op1.to_shwi ();
+	  o1 = op2.to_shwi ();
 	}
       else
 	{
-	  o0 = zext_hwi (op1[0], prec);
-	  o1 = zext_hwi (op2[0], prec);
+	  o0 = op1.to_uhwi ();
+	  o1 = op2.to_uhwi ();
 	}
 
       r = o0 * o1;
@@ -1348,9 +1387,9 @@ wi::mul_internal (HOST_WIDE_INT *val, co
     }
 
   /* We do unsigned mul and then correct it.  */
-  wi_unpack (u, (const unsigned HOST_WIDE_INT*)op1, op1len,
+  wi_unpack (u, (const unsigned HOST_WIDE_INT *) op1val, op1len,
 	     half_blocks_needed, prec, SIGNED);
-  wi_unpack (v, (const unsigned HOST_WIDE_INT*)op2, op2len,
+  wi_unpack (v, (const unsigned HOST_WIDE_INT *) op2val, op2len,
 	     half_blocks_needed, prec, SIGNED);
 
   /* The 2 is for a full mult.  */
@@ -1375,7 +1414,7 @@ wi::mul_internal (HOST_WIDE_INT *val, co
   if (sgn == SIGNED && (full || high || needs_overflow))
     {
       unsigned HOST_WIDE_INT b;
-      if (op1[op1len-1] < 0)
+      if (wi::neg_p (op1))
 	{
 	  b = 0;
 	  for (i = 0; i < half_blocks_needed; i++)
@@ -1386,7 +1425,7 @@ wi::mul_internal (HOST_WIDE_INT *val, co
 	      b = t >> (HOST_BITS_PER_WIDE_INT - 1);
 	    }
 	}
-      if (op2[op2len-1] < 0)
+      if (wi::neg_p (op2))
 	{
 	  b = 0;
 	  for (i = 0; i < half_blocks_needed; i++)

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2013-12-01 10:29 [wide-int] Add fast path for hosts with HWI widening multiplication Richard Sandiford
@ 2013-12-01 18:48 ` Ian Lance Taylor
  2013-12-02 19:35   ` Richard Sandiford
  2013-12-04 12:56 ` Richard Sandiford
  1 sibling, 1 reply; 24+ messages in thread
From: Ian Lance Taylor @ 2013-12-01 18:48 UTC (permalink / raw)
  To: gcc-patches, Kenneth Zadeck, Mike Stump, Richard Sandiford

On Sun, Dec 1, 2013 at 2:28 AM, Richard Sandiford
<rdsandiford@googlemail.com> wrote:
>
> I followed Joseph's suggestion and reused longlong.h.  I copied it from
> libgcc rather than glibc since it seemed better for GCC to have a single
> version across both gcc/ and libgcc/.  I can put it in include/ if that
> seems better.

Actually copying complex code like this does not seem maintainable.  I
think there needs to be only one copy in the GCC sources.  If that
requires moving it back from libgcc to gcc, or moving it to include,
do that.

Ian

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2013-12-01 18:48 ` Ian Lance Taylor
@ 2013-12-02 19:35   ` Richard Sandiford
  2013-12-02 19:48     ` Paolo Bonzini
  0 siblings, 1 reply; 24+ messages in thread
From: Richard Sandiford @ 2013-12-02 19:35 UTC (permalink / raw)
  To: Ian Lance Taylor; +Cc: gcc-patches, Kenneth Zadeck, Mike Stump

Ian Lance Taylor <iant@google.com> writes:
> On Sun, Dec 1, 2013 at 2:28 AM, Richard Sandiford
> <rdsandiford@googlemail.com> wrote:
>> I followed Joseph's suggestion and reused longlong.h.  I copied it from
>> libgcc rather than glibc since it seemed better for GCC to have a single
>> version across both gcc/ and libgcc/.  I can put it in include/ if that
>> seems better.
>
> Actually copying complex code like this does not seem maintainable.  I
> think there needs to be only one copy in the GCC sources.  If that
> requires moving it back from libgcc to gcc, or moving it to include,
> do that.

OK, will do, but which do you prefer?

Thanks,
Richard

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2013-12-02 19:35   ` Richard Sandiford
@ 2013-12-02 19:48     ` Paolo Bonzini
  2013-12-03 23:55       ` Richard Sandiford
  0 siblings, 1 reply; 24+ messages in thread
From: Paolo Bonzini @ 2013-12-02 19:48 UTC (permalink / raw)
  To: Ian Lance Taylor, gcc-patches, Kenneth Zadeck, Mike Stump, rdsandiford

Il 02/12/2013 20:34, Richard Sandiford ha scritto:
>>> >> I followed Joseph's suggestion and reused longlong.h.  I copied it from
>>> >> libgcc rather than glibc since it seemed better for GCC to have a single
>>> >> version across both gcc/ and libgcc/.  I can put it in include/ if that
>>> >> seems better.
>> >
>> > Actually copying complex code like this does not seem maintainable.  I
>> > think there needs to be only one copy in the GCC sources.  If that
>> > requires moving it back from libgcc to gcc, or moving it to include,
>> > do that.
> OK, will do, but which do you prefer?

libgcc/ should not use gcc/ sources too much.  Please put it in include/.

Paolo

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2013-12-02 19:48     ` Paolo Bonzini
@ 2013-12-03 23:55       ` Richard Sandiford
  2013-12-04  0:18         ` Ian Lance Taylor
  0 siblings, 1 reply; 24+ messages in thread
From: Richard Sandiford @ 2013-12-03 23:55 UTC (permalink / raw)
  To: Paolo Bonzini; +Cc: Ian Lance Taylor, gcc-patches, Kenneth Zadeck, Mike Stump

Paolo Bonzini <bonzini@gnu.org> writes:
> Il 02/12/2013 20:34, Richard Sandiford ha scritto:
>>>> >> I followed Joseph's suggestion and reused longlong.h.  I copied it from
>>>> >> libgcc rather than glibc since it seemed better for GCC to have a single
>>>> >> version across both gcc/ and libgcc/.  I can put it in include/ if that
>>>> >> seems better.
>>> >
>>> > Actually copying complex code like this does not seem maintainable.  I
>>> > think there needs to be only one copy in the GCC sources.  If that
>>> > requires moving it back from libgcc to gcc, or moving it to include,
>>> > do that.
>> OK, will do, but which do you prefer?
>
> libgcc/ should not use gcc/ sources too much.  Please put it in include/.

OK, how about this, against trunk?  Tested on x86_64-linux-gnu.

Thanks,
Richard


include/
	* longlong.h: New file.

libgcc/
	* longlong.h: Delete (moved to include/).

libquadmath/
	* Makefile.am (AM_CPPFLAGS): Define.
	* Makefile.in: Regenerate.
	* printf/gmp-impl.h: Remove path from longlong.h include.

Index: libquadmath/Makefile.am
===================================================================
--- libquadmath/Makefile.am	2012-11-03 09:21:21.000000000 +0000
+++ libquadmath/Makefile.am	2013-12-03 19:30:02.551636610 +0000
@@ -6,6 +6,7 @@ AUTOMAKE_OPTIONS = 1.8 foreign
 ## Skip over everything if the quadlib is not available:
 if BUILD_LIBQUADMATH
 ACLOCAL_AMFLAGS = -I .. -I ../config
+AM_CPPFLAGS = -I $(top_srcdir)/../include
 
 ## May be used by toolexeclibdir.
 gcc_version := $(shell cat $(top_srcdir)/../gcc/BASE-VER)
Index: libquadmath/Makefile.in
===================================================================
--- libquadmath/Makefile.in	2013-11-18 14:46:49.994728318 +0000
+++ libquadmath/Makefile.in	2013-12-03 19:31:06.828134039 +0000
@@ -310,6 +310,7 @@ top_builddir = @top_builddir@
 top_srcdir = @top_srcdir@
 AUTOMAKE_OPTIONS = 1.8 foreign
 @BUILD_LIBQUADMATH_TRUE@ACLOCAL_AMFLAGS = -I .. -I ../config
+@BUILD_LIBQUADMATH_TRUE@AM_CPPFLAGS = -I $(top_srcdir)/../include
 @BUILD_LIBQUADMATH_TRUE@gcc_version := $(shell cat $(top_srcdir)/../gcc/BASE-VER)
 @BUILD_LIBQUADMATH_TRUE@@LIBQUAD_USE_SYMVER_FALSE@version_arg = 
 @BUILD_LIBQUADMATH_TRUE@@LIBQUAD_USE_SYMVER_GNU_TRUE@@LIBQUAD_USE_SYMVER_TRUE@version_arg = -Wl,--version-script=$(srcdir)/quadmath.map
Index: libquadmath/printf/gmp-impl.h
===================================================================
--- libquadmath/printf/gmp-impl.h	2013-12-03 19:20:58.167441116 +0000
+++ libquadmath/printf/gmp-impl.h	2013-12-03 19:21:07.093509627 +0000
@@ -60,7 +60,7 @@ #define attribute_hidden __attribute__((
 #define attribute_hidden
 #endif
 
-#include "../../libgcc/longlong.h"
+#include "longlong.h"
 
 /* Copy NLIMBS *limbs* from SRC to DST.  */
 #define MPN_COPY_INCR(DST, SRC, NLIMBS) \

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2013-12-03 23:55       ` Richard Sandiford
@ 2013-12-04  0:18         ` Ian Lance Taylor
  0 siblings, 0 replies; 24+ messages in thread
From: Ian Lance Taylor @ 2013-12-04  0:18 UTC (permalink / raw)
  To: Paolo Bonzini, Ian Lance Taylor, gcc-patches, Kenneth Zadeck,
	Mike Stump, Richard Sandiford

On Tue, Dec 3, 2013 at 3:55 PM, Richard Sandiford
<rdsandiford@googlemail.com> wrote:
>
> OK, how about this, against trunk?  Tested on x86_64-linux-gnu.
>
> Thanks,
> Richard
>
>
> include/
>         * longlong.h: New file.
>
> libgcc/
>         * longlong.h: Delete (moved to include/).
>
> libquadmath/
>         * Makefile.am (AM_CPPFLAGS): Define.
>         * Makefile.in: Regenerate.
>         * printf/gmp-impl.h: Remove path from longlong.h include.

This is OK.

Thanks.

Ian

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2013-12-01 10:29 [wide-int] Add fast path for hosts with HWI widening multiplication Richard Sandiford
  2013-12-01 18:48 ` Ian Lance Taylor
@ 2013-12-04 12:56 ` Richard Sandiford
  2013-12-04 14:04   ` Kenneth Zadeck
  2014-05-08 12:49   ` Ramana Radhakrishnan
  1 sibling, 2 replies; 24+ messages in thread
From: Richard Sandiford @ 2013-12-04 12:56 UTC (permalink / raw)
  To: gcc-patches; +Cc: zadeck, mikestump

Richard Sandiford <rdsandiford@googlemail.com> writes:
> This patch handles multiplications using a single HWIxHWI->2HWI multiplication
> on hosts that have one.  This removes all uses of the slow (half-HWI) path
> for insn-recog.ii.  The slow path is still used 58 times for cp/parser.ii
> and 168 times for fold-const.ii, but at that kind of level it shouldn't
> matter much.
>
> I followed Joseph's suggestion and reused longlong.h.  I copied it from
> libgcc rather than glibc since it seemed better for GCC to have a single
> version across both gcc/ and libgcc/.  I can put it in include/ if that
> seems better.

I've committed the patch to move longlong.h to trunk and merged back
to the branch, so all that's left is the wide-int.cc patch.  OK to install?

Thanks,
Richard


Index: gcc/wide-int.cc
===================================================================
--- gcc/wide-int.cc	2013-12-03 23:59:08.133658567 +0000
+++ gcc/wide-int.cc	2013-12-04 12:55:28.466895358 +0000
@@ -27,6 +27,16 @@ along with GCC; see the file COPYING3.
 #include "tree.h"
 #include "dumpfile.h"
 
+#if GCC_VERSION >= 3000
+#define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
+typedef unsigned HOST_HALF_WIDE_INT UHWtype;
+typedef unsigned HOST_WIDE_INT UWtype;
+typedef unsigned int UQItype __attribute__ ((mode (QI)));
+typedef unsigned int USItype __attribute__ ((mode (SI)));
+typedef unsigned int UDItype __attribute__ ((mode (DI)));
+#include "longlong.h"
+#endif
+
 /* This is the maximal size of the buffer needed for dump.  */
 const unsigned int MAX_SIZE = (4 * (MAX_BITSIZE_MODE_ANY_INT / 4
 				    + (MAX_BITSIZE_MODE_ANY_INT
@@ -1255,8 +1265,8 @@ wi_pack (unsigned HOST_WIDE_INT *result,
    record in *OVERFLOW whether the result overflowed.  SGN controls
    the signedness and is used to check overflow or if HIGH is set.  */
 unsigned int
-wi::mul_internal (HOST_WIDE_INT *val, const HOST_WIDE_INT *op1,
-		  unsigned int op1len, const HOST_WIDE_INT *op2,
+wi::mul_internal (HOST_WIDE_INT *val, const HOST_WIDE_INT *op1val,
+		  unsigned int op1len, const HOST_WIDE_INT *op2val,
 		  unsigned int op2len, unsigned int prec, signop sgn,
 		  bool *overflow, bool high)
 {
@@ -1285,24 +1295,53 @@ wi::mul_internal (HOST_WIDE_INT *val, co
   if (needs_overflow)
     *overflow = false;
 
+  wide_int_ref op1 = wi::storage_ref (op1val, op1len, prec);
+  wide_int_ref op2 = wi::storage_ref (op2val, op2len, prec);
+
   /* This is a surprisingly common case, so do it first.  */
-  if ((op1len == 1 && op1[0] == 0) || (op2len == 1 && op2[0] == 0))
+  if (op1 == 0 || op2 == 0)
     {
       val[0] = 0;
       return 1;
     }
 
+#ifdef umul_ppmm
+  if (sgn == UNSIGNED)
+    {
+      /* If the inputs are single HWIs and the output has room for at
+	 least two HWIs, we can use umul_ppmm directly.  */
+      if (prec >= HOST_BITS_PER_WIDE_INT * 2
+	  && wi::fits_uhwi_p (op1)
+	  && wi::fits_uhwi_p (op2))
+	{
+	  umul_ppmm (val[1], val[0], op1.ulow (), op2.ulow ());
+	  return 1 + (val[1] != 0 || val[0] < 0);
+	}
+      /* Likewise if the output is a full single HWI, except that the
+	 upper HWI of the result is only used for determining overflow.
+	 (We handle this case inline when overflow isn't needed.)  */
+      else if (prec == HOST_BITS_PER_WIDE_INT)
+	{
+	  unsigned HOST_WIDE_INT upper;
+	  umul_ppmm (upper, val[0], op1.ulow (), op2.ulow ());
+	  if (needs_overflow)
+	    *overflow = (upper != 0);
+	  return 1;
+	}
+    }
+#endif
+
   /* Handle multiplications by 1.  */
-  if (op1len == 1 && op1[0] == 1)
+  if (op1 == 1)
     {
       for (i = 0; i < op2len; i++)
-	val[i] = op2[i];
+	val[i] = op2val[i];
       return op2len;
     }
-  if (op2len == 1 && op2[0] == 1)
+  if (op2 == 1)
     {
       for (i = 0; i < op1len; i++)
-	val[i] = op1[i];
+	val[i] = op1val[i];
       return op1len;
     }
 
@@ -1316,13 +1355,13 @@ wi::mul_internal (HOST_WIDE_INT *val, co
 
       if (sgn == SIGNED)
 	{
-	  o0 = sext_hwi (op1[0], prec);
-	  o1 = sext_hwi (op2[0], prec);
+	  o0 = op1.to_shwi ();
+	  o1 = op2.to_shwi ();
 	}
       else
 	{
-	  o0 = zext_hwi (op1[0], prec);
-	  o1 = zext_hwi (op2[0], prec);
+	  o0 = op1.to_uhwi ();
+	  o1 = op2.to_uhwi ();
 	}
 
       r = o0 * o1;
@@ -1344,9 +1383,9 @@ wi::mul_internal (HOST_WIDE_INT *val, co
     }
 
   /* We do unsigned mul and then correct it.  */
-  wi_unpack (u, (const unsigned HOST_WIDE_INT*)op1, op1len,
+  wi_unpack (u, (const unsigned HOST_WIDE_INT *) op1val, op1len,
 	     half_blocks_needed, prec, SIGNED);
-  wi_unpack (v, (const unsigned HOST_WIDE_INT*)op2, op2len,
+  wi_unpack (v, (const unsigned HOST_WIDE_INT *) op2val, op2len,
 	     half_blocks_needed, prec, SIGNED);
 
   /* The 2 is for a full mult.  */
@@ -1371,7 +1410,7 @@ wi::mul_internal (HOST_WIDE_INT *val, co
   if (sgn == SIGNED && (high || needs_overflow))
     {
       unsigned HOST_WIDE_INT b;
-      if (op1[op1len-1] < 0)
+      if (wi::neg_p (op1))
 	{
 	  b = 0;
 	  for (i = 0; i < half_blocks_needed; i++)
@@ -1382,7 +1421,7 @@ wi::mul_internal (HOST_WIDE_INT *val, co
 	      b = t >> (HOST_BITS_PER_WIDE_INT - 1);
 	    }
 	}
-      if (op2[op2len-1] < 0)
+      if (wi::neg_p (op2))
 	{
 	  b = 0;
 	  for (i = 0; i < half_blocks_needed; i++)

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2013-12-04 12:56 ` Richard Sandiford
@ 2013-12-04 14:04   ` Kenneth Zadeck
  2014-05-08 12:49   ` Ramana Radhakrishnan
  1 sibling, 0 replies; 24+ messages in thread
From: Kenneth Zadeck @ 2013-12-04 14:04 UTC (permalink / raw)
  To: gcc-patches, mikestump, rdsandiford

On 12/04/2013 07:56 AM, Richard Sandiford wrote:
> Richard Sandiford <rdsandiford@googlemail.com> writes:
>> This patch handles multiplications using a single HWIxHWI->2HWI multiplication
>> on hosts that have one.  This removes all uses of the slow (half-HWI) path
>> for insn-recog.ii.  The slow path is still used 58 times for cp/parser.ii
>> and 168 times for fold-const.ii, but at that kind of level it shouldn't
>> matter much.
>>
>> I followed Joseph's suggestion and reused longlong.h.  I copied it from
>> libgcc rather than glibc since it seemed better for GCC to have a single
>> version across both gcc/ and libgcc/.  I can put it in include/ if that
>> seems better.
> I've committed the patch to move longlong.h to trunk and merged back
> to the branch, so all that's left is the wide-int.cc patch.  OK to install?
>
> Thanks,
> Richard
>
>
> Index: gcc/wide-int.cc
> ===================================================================
> --- gcc/wide-int.cc	2013-12-03 23:59:08.133658567 +0000
> +++ gcc/wide-int.cc	2013-12-04 12:55:28.466895358 +0000
> @@ -27,6 +27,16 @@ along with GCC; see the file COPYING3.
>   #include "tree.h"
>   #include "dumpfile.h"
>   
> +#if GCC_VERSION >= 3000
> +#define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
> +typedef unsigned HOST_HALF_WIDE_INT UHWtype;
> +typedef unsigned HOST_WIDE_INT UWtype;
> +typedef unsigned int UQItype __attribute__ ((mode (QI)));
> +typedef unsigned int USItype __attribute__ ((mode (SI)));
> +typedef unsigned int UDItype __attribute__ ((mode (DI)));
> +#include "longlong.h"
> +#endif
> +
>   /* This is the maximal size of the buffer needed for dump.  */
>   const unsigned int MAX_SIZE = (4 * (MAX_BITSIZE_MODE_ANY_INT / 4
>   				    + (MAX_BITSIZE_MODE_ANY_INT
> @@ -1255,8 +1265,8 @@ wi_pack (unsigned HOST_WIDE_INT *result,
>      record in *OVERFLOW whether the result overflowed.  SGN controls
>      the signedness and is used to check overflow or if HIGH is set.  */
>   unsigned int
> -wi::mul_internal (HOST_WIDE_INT *val, const HOST_WIDE_INT *op1,
> -		  unsigned int op1len, const HOST_WIDE_INT *op2,
> +wi::mul_internal (HOST_WIDE_INT *val, const HOST_WIDE_INT *op1val,
> +		  unsigned int op1len, const HOST_WIDE_INT *op2val,
>   		  unsigned int op2len, unsigned int prec, signop sgn,
>   		  bool *overflow, bool high)
>   {
> @@ -1285,24 +1295,53 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>     if (needs_overflow)
>       *overflow = false;
>   
> +  wide_int_ref op1 = wi::storage_ref (op1val, op1len, prec);
> +  wide_int_ref op2 = wi::storage_ref (op2val, op2len, prec);
> +
>     /* This is a surprisingly common case, so do it first.  */
> -  if ((op1len == 1 && op1[0] == 0) || (op2len == 1 && op2[0] == 0))
> +  if (op1 == 0 || op2 == 0)
>       {
>         val[0] = 0;
>         return 1;
>       }
>   
> +#ifdef umul_ppmm
> +  if (sgn == UNSIGNED)
> +    {
> +      /* If the inputs are single HWIs and the output has room for at
> +	 least two HWIs, we can use umul_ppmm directly.  */
> +      if (prec >= HOST_BITS_PER_WIDE_INT * 2
> +	  && wi::fits_uhwi_p (op1)
> +	  && wi::fits_uhwi_p (op2))
> +	{
> +	  umul_ppmm (val[1], val[0], op1.ulow (), op2.ulow ());
> +	  return 1 + (val[1] != 0 || val[0] < 0);
> +	}
> +      /* Likewise if the output is a full single HWI, except that the
> +	 upper HWI of the result is only used for determining overflow.
> +	 (We handle this case inline when overflow isn't needed.)  */
> +      else if (prec == HOST_BITS_PER_WIDE_INT)
> +	{
> +	  unsigned HOST_WIDE_INT upper;
> +	  umul_ppmm (upper, val[0], op1.ulow (), op2.ulow ());
> +	  if (needs_overflow)
> +	    *overflow = (upper != 0);
> +	  return 1;
> +	}
> +    }
> +#endif
> +
>     /* Handle multiplications by 1.  */
> -  if (op1len == 1 && op1[0] == 1)
> +  if (op1 == 1)
>       {
>         for (i = 0; i < op2len; i++)
> -	val[i] = op2[i];
> +	val[i] = op2val[i];
>         return op2len;
>       }
> -  if (op2len == 1 && op2[0] == 1)
> +  if (op2 == 1)
>       {
>         for (i = 0; i < op1len; i++)
> -	val[i] = op1[i];
> +	val[i] = op1val[i];
>         return op1len;
>       }
>   
> @@ -1316,13 +1355,13 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>   
>         if (sgn == SIGNED)
>   	{
> -	  o0 = sext_hwi (op1[0], prec);
> -	  o1 = sext_hwi (op2[0], prec);
> +	  o0 = op1.to_shwi ();
> +	  o1 = op2.to_shwi ();
>   	}
>         else
>   	{
> -	  o0 = zext_hwi (op1[0], prec);
> -	  o1 = zext_hwi (op2[0], prec);
> +	  o0 = op1.to_uhwi ();
> +	  o1 = op2.to_uhwi ();
>   	}
>   
>         r = o0 * o1;
> @@ -1344,9 +1383,9 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>       }
>   
>     /* We do unsigned mul and then correct it.  */
> -  wi_unpack (u, (const unsigned HOST_WIDE_INT*)op1, op1len,
> +  wi_unpack (u, (const unsigned HOST_WIDE_INT *) op1val, op1len,
>   	     half_blocks_needed, prec, SIGNED);
> -  wi_unpack (v, (const unsigned HOST_WIDE_INT*)op2, op2len,
> +  wi_unpack (v, (const unsigned HOST_WIDE_INT *) op2val, op2len,
>   	     half_blocks_needed, prec, SIGNED);
>   
>     /* The 2 is for a full mult.  */
> @@ -1371,7 +1410,7 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>     if (sgn == SIGNED && (high || needs_overflow))
>       {
>         unsigned HOST_WIDE_INT b;
> -      if (op1[op1len-1] < 0)
> +      if (wi::neg_p (op1))
>   	{
>   	  b = 0;
>   	  for (i = 0; i < half_blocks_needed; i++)
> @@ -1382,7 +1421,7 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>   	      b = t >> (HOST_BITS_PER_WIDE_INT - 1);
>   	    }
>   	}
> -      if (op2[op2len-1] < 0)
> +      if (wi::neg_p (op2))
>   	{
>   	  b = 0;
>   	  for (i = 0; i < half_blocks_needed; i++)
this is fine with me.

kenny

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2013-12-04 12:56 ` Richard Sandiford
  2013-12-04 14:04   ` Kenneth Zadeck
@ 2014-05-08 12:49   ` Ramana Radhakrishnan
  2014-05-08 13:13     ` Richard Sandiford
  2014-05-08 16:38     ` Joseph S. Myers
  1 sibling, 2 replies; 24+ messages in thread
From: Ramana Radhakrishnan @ 2014-05-08 12:49 UTC (permalink / raw)
  To: gcc-patches, zadeck, Mike Stump, Richard Sandiford, Richard Henderson

[-- Attachment #1: Type: text/plain, Size: 6524 bytes --]

On Wed, Dec 4, 2013 at 12:56 PM, Richard Sandiford
<rdsandiford@googlemail.com> wrote:
> Richard Sandiford <rdsandiford@googlemail.com> writes:
>> This patch handles multiplications using a single HWIxHWI->2HWI multiplication
>> on hosts that have one.  This removes all uses of the slow (half-HWI) path
>> for insn-recog.ii.  The slow path is still used 58 times for cp/parser.ii
>> and 168 times for fold-const.ii, but at that kind of level it shouldn't
>> matter much.
>>
>> I followed Joseph's suggestion and reused longlong.h.  I copied it from
>> libgcc rather than glibc since it seemed better for GCC to have a single
>> version across both gcc/ and libgcc/.  I can put it in include/ if that
>> seems better.
>
> I've committed the patch to move longlong.h to trunk and merged back
> to the branch, so all that's left is the wide-int.cc patch.  OK to install?
>
> Thanks,
> Richard
>
>
> Index: gcc/wide-int.cc
> ===================================================================
> --- gcc/wide-int.cc     2013-12-03 23:59:08.133658567 +0000
> +++ gcc/wide-int.cc     2013-12-04 12:55:28.466895358 +0000
> @@ -27,6 +27,16 @@ along with GCC; see the file COPYING3.
>  #include "tree.h"
>  #include "dumpfile.h"
>
> +#if GCC_VERSION >= 3000
> +#define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
> +typedef unsigned HOST_HALF_WIDE_INT UHWtype;
> +typedef unsigned HOST_WIDE_INT UWtype;
> +typedef unsigned int UQItype __attribute__ ((mode (QI)));
> +typedef unsigned int USItype __attribute__ ((mode (SI)));
> +typedef unsigned int UDItype __attribute__ ((mode (DI)));
> +#include "longlong.h"

We also need something like the attached patch to handle architectures
which use UDWType in longlong.h. I noticed this when trying rth's
patch stack http://gcc.gnu.org/ml/gcc-patches/2014-01/msg00391.html to
improve longlong.h for AArch64.

It's needed when rth's patches go in finally for aarch64 but could
probably go in now - after all the comment in longlong.h says you need
to define UDWtype ... Don't mind either way.

regards
Ramana

<DATE>  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>

       * wide-int.cc (UTItype): Define.
       (UDWtype): Define for appropriate W_TYPE_SIZE.


> +#endif
> +
>  /* This is the maximal size of the buffer needed for dump.  */
>  const unsigned int MAX_SIZE = (4 * (MAX_BITSIZE_MODE_ANY_INT / 4
>                                     + (MAX_BITSIZE_MODE_ANY_INT
> @@ -1255,8 +1265,8 @@ wi_pack (unsigned HOST_WIDE_INT *result,
>     record in *OVERFLOW whether the result overflowed.  SGN controls
>     the signedness and is used to check overflow or if HIGH is set.  */
>  unsigned int
> -wi::mul_internal (HOST_WIDE_INT *val, const HOST_WIDE_INT *op1,
> -                 unsigned int op1len, const HOST_WIDE_INT *op2,
> +wi::mul_internal (HOST_WIDE_INT *val, const HOST_WIDE_INT *op1val,
> +                 unsigned int op1len, const HOST_WIDE_INT *op2val,
>                   unsigned int op2len, unsigned int prec, signop sgn,
>                   bool *overflow, bool high)
>  {
> @@ -1285,24 +1295,53 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>    if (needs_overflow)
>      *overflow = false;
>
> +  wide_int_ref op1 = wi::storage_ref (op1val, op1len, prec);
> +  wide_int_ref op2 = wi::storage_ref (op2val, op2len, prec);
> +
>    /* This is a surprisingly common case, so do it first.  */
> -  if ((op1len == 1 && op1[0] == 0) || (op2len == 1 && op2[0] == 0))
> +  if (op1 == 0 || op2 == 0)
>      {
>        val[0] = 0;
>        return 1;
>      }
>
> +#ifdef umul_ppmm
> +  if (sgn == UNSIGNED)
> +    {
> +      /* If the inputs are single HWIs and the output has room for at
> +        least two HWIs, we can use umul_ppmm directly.  */
> +      if (prec >= HOST_BITS_PER_WIDE_INT * 2
> +         && wi::fits_uhwi_p (op1)
> +         && wi::fits_uhwi_p (op2))
> +       {
> +         umul_ppmm (val[1], val[0], op1.ulow (), op2.ulow ());
> +         return 1 + (val[1] != 0 || val[0] < 0);
> +       }
> +      /* Likewise if the output is a full single HWI, except that the
> +        upper HWI of the result is only used for determining overflow.
> +        (We handle this case inline when overflow isn't needed.)  */
> +      else if (prec == HOST_BITS_PER_WIDE_INT)
> +       {
> +         unsigned HOST_WIDE_INT upper;
> +         umul_ppmm (upper, val[0], op1.ulow (), op2.ulow ());
> +         if (needs_overflow)
> +           *overflow = (upper != 0);
> +         return 1;
> +       }
> +    }
> +#endif
> +
>    /* Handle multiplications by 1.  */
> -  if (op1len == 1 && op1[0] == 1)
> +  if (op1 == 1)
>      {
>        for (i = 0; i < op2len; i++)
> -       val[i] = op2[i];
> +       val[i] = op2val[i];
>        return op2len;
>      }
> -  if (op2len == 1 && op2[0] == 1)
> +  if (op2 == 1)
>      {
>        for (i = 0; i < op1len; i++)
> -       val[i] = op1[i];
> +       val[i] = op1val[i];
>        return op1len;
>      }
>
> @@ -1316,13 +1355,13 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>
>        if (sgn == SIGNED)
>         {
> -         o0 = sext_hwi (op1[0], prec);
> -         o1 = sext_hwi (op2[0], prec);
> +         o0 = op1.to_shwi ();
> +         o1 = op2.to_shwi ();
>         }
>        else
>         {
> -         o0 = zext_hwi (op1[0], prec);
> -         o1 = zext_hwi (op2[0], prec);
> +         o0 = op1.to_uhwi ();
> +         o1 = op2.to_uhwi ();
>         }
>
>        r = o0 * o1;
> @@ -1344,9 +1383,9 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>      }
>
>    /* We do unsigned mul and then correct it.  */
> -  wi_unpack (u, (const unsigned HOST_WIDE_INT*)op1, op1len,
> +  wi_unpack (u, (const unsigned HOST_WIDE_INT *) op1val, op1len,
>              half_blocks_needed, prec, SIGNED);
> -  wi_unpack (v, (const unsigned HOST_WIDE_INT*)op2, op2len,
> +  wi_unpack (v, (const unsigned HOST_WIDE_INT *) op2val, op2len,
>              half_blocks_needed, prec, SIGNED);
>
>    /* The 2 is for a full mult.  */
> @@ -1371,7 +1410,7 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>    if (sgn == SIGNED && (high || needs_overflow))
>      {
>        unsigned HOST_WIDE_INT b;
> -      if (op1[op1len-1] < 0)
> +      if (wi::neg_p (op1))
>         {
>           b = 0;
>           for (i = 0; i < half_blocks_needed; i++)
> @@ -1382,7 +1421,7 @@ wi::mul_internal (HOST_WIDE_INT *val, co
>               b = t >> (HOST_BITS_PER_WIDE_INT - 1);
>             }
>         }
> -      if (op2[op2len-1] < 0)
> +      if (wi::neg_p (op2))
>         {
>           b = 0;
>           for (i = 0; i < half_blocks_needed; i++)

[-- Attachment #2: p1.txt --]
[-- Type: text/plain, Size: 1102 bytes --]

commit 62672d152281515c1d97ec1370698544cb2c81b6
Author: Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
Date:   Thu May 8 09:56:59 2014 +0000

    Fix up UDWtype for wide-int merge.

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 354cdb9..8ef9a0f 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,8 @@
+2014-05-08  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
+
+	* wide-int.cc (UTItype): Define.
+	(UDWtype): Define for appropriate W_TYPE_SIZE.
+
 2014-05-08  Alan Modra  <amodra@gmail.com>
 
 	PR target/60737
diff --git a/gcc/wide-int.cc b/gcc/wide-int.cc
index 69a15bc..3552e03 100644
--- a/gcc/wide-int.cc
+++ b/gcc/wide-int.cc
@@ -34,6 +34,12 @@ typedef unsigned HOST_WIDE_INT UWtype;
 typedef unsigned int UQItype __attribute__ ((mode (QI)));
 typedef unsigned int USItype __attribute__ ((mode (SI)));
 typedef unsigned int UDItype __attribute__ ((mode (DI)));
+typedef unsigned int UTItype __attribute__ ((mode (TI)));
+#if W_TYPE_SIZE == 32
+# define UDWtype       UDItype
+#elif W_TYPE_SIZE == 64
+# define UDWtype       UTItype
+#endif
 #include "longlong.h"
 #endif
 

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 12:49   ` Ramana Radhakrishnan
@ 2014-05-08 13:13     ` Richard Sandiford
  2014-05-08 16:38     ` Joseph S. Myers
  1 sibling, 0 replies; 24+ messages in thread
From: Richard Sandiford @ 2014-05-08 13:13 UTC (permalink / raw)
  To: ramrad01; +Cc: gcc-patches, zadeck, Mike Stump, Richard Henderson

Ramana Radhakrishnan <ramana.gcc@googlemail.com> writes:
> diff --git a/gcc/ChangeLog b/gcc/ChangeLog
> index 354cdb9..8ef9a0f 100644
> --- a/gcc/ChangeLog
> +++ b/gcc/ChangeLog
> @@ -1,3 +1,8 @@
> +2014-05-08  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
> +
> +	* wide-int.cc (UTItype): Define.
> +	(UDWtype): Define for appropriate W_TYPE_SIZE.
> +
>  2014-05-08  Alan Modra  <amodra@gmail.com>
>  
>  	PR target/60737
> diff --git a/gcc/wide-int.cc b/gcc/wide-int.cc
> index 69a15bc..3552e03 100644
> --- a/gcc/wide-int.cc
> +++ b/gcc/wide-int.cc
> @@ -34,6 +34,12 @@ typedef unsigned HOST_WIDE_INT UWtype;
>  typedef unsigned int UQItype __attribute__ ((mode (QI)));
>  typedef unsigned int USItype __attribute__ ((mode (SI)));
>  typedef unsigned int UDItype __attribute__ ((mode (DI)));
> +typedef unsigned int UTItype __attribute__ ((mode (TI)));
> +#if W_TYPE_SIZE == 32
> +# define UDWtype       UDItype
> +#elif W_TYPE_SIZE == 64
> +# define UDWtype       UTItype
> +#endif
>  #include "longlong.h"
>  #endif

OK, thanks.  Technically the W_TYPE_SIZE == 32 should be dead now that
HWI is always 64 bits, but there's still a debate about whether wide-int
should continue using HOST_WIDE_INT/int64_t or something else, so I think
it would be best to keep it.

Richard

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 12:49   ` Ramana Radhakrishnan
  2014-05-08 13:13     ` Richard Sandiford
@ 2014-05-08 16:38     ` Joseph S. Myers
  2014-05-08 18:31       ` Richard Sandiford
  1 sibling, 1 reply; 24+ messages in thread
From: Joseph S. Myers @ 2014-05-08 16:38 UTC (permalink / raw)
  To: ramrad01
  Cc: gcc-patches, zadeck, Mike Stump, Richard Sandiford, Richard Henderson

On Thu, 8 May 2014, Ramana Radhakrishnan wrote:

> <DATE>  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
> 
>        * wide-int.cc (UTItype): Define.
>        (UDWtype): Define for appropriate W_TYPE_SIZE.

This breaks builds for 32-bit hosts, where TImode isn't supported.  You 
can only use TImode on the host if it's 64-bit.

wide-int.cc:37:56: error: unable to emulate 'TI'

-- 
Joseph S. Myers
joseph@codesourcery.com

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 16:38     ` Joseph S. Myers
@ 2014-05-08 18:31       ` Richard Sandiford
  2014-05-08 18:37         ` Kenneth Zadeck
  2014-05-09  9:01         ` Richard Earnshaw
  0 siblings, 2 replies; 24+ messages in thread
From: Richard Sandiford @ 2014-05-08 18:31 UTC (permalink / raw)
  To: Joseph S. Myers
  Cc: ramrad01, gcc-patches, zadeck, Mike Stump, Richard Henderson

"Joseph S. Myers" <joseph@codesourcery.com> writes:
> On Thu, 8 May 2014, Ramana Radhakrishnan wrote:
>
>> <DATE>  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
>> 
>>        * wide-int.cc (UTItype): Define.
>>        (UDWtype): Define for appropriate W_TYPE_SIZE.
>
> This breaks builds for 32-bit hosts, where TImode isn't supported.  You 
> can only use TImode on the host if it's 64-bit.
>
> wide-int.cc:37:56: error: unable to emulate 'TI'

The longlong.h interface seems to be designed to be as difficult to use
as possible :-(  So maybe we really do need to limit it to hosts that are
known to work and benefit from it.

How about the following?  I tested that it produces identical
wide-int.o .text for x86_64.

I think additions to or removals from the list should be treated as
pre-approved.

Thanks,
Richard


gcc/
	* wide-int.cc: Only include longlong.h for certain targets.

Index: gcc/wide-int.cc
===================================================================
--- gcc/wide-int.cc	2014-05-08 19:13:15.782158808 +0100
+++ gcc/wide-int.cc	2014-05-08 19:28:52.880742385 +0100
@@ -27,19 +27,20 @@ along with GCC; see the file COPYING3.
 #include "tree.h"
 #include "dumpfile.h"
 
-#if GCC_VERSION >= 3000
+#if (GCC_VERSION >= 3000 \
+     && (defined __aarch64 \
+         || defined __alpha \
+         || defined __ia64 \
+         || defined __powerpc64__ \
+         || defined __sparcv9 \
+         || defined __x86_64__))
 #define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
-typedef unsigned HOST_HALF_WIDE_INT UHWtype;
-typedef unsigned HOST_WIDE_INT UWtype;
 typedef unsigned int UQItype __attribute__ ((mode (QI)));
 typedef unsigned int USItype __attribute__ ((mode (SI)));
 typedef unsigned int UDItype __attribute__ ((mode (DI)));
-typedef unsigned int UTItype __attribute__ ((mode (TI)));
-#if W_TYPE_SIZE == 32
-# define UDWtype       UDItype
-#elif W_TYPE_SIZE == 64
-# define UDWtype       UTItype
-#endif
+typedef unsigned HOST_HALF_WIDE_INT UHWtype;
+typedef unsigned HOST_WIDE_INT UWtype;
+typedef unsigned int UDWtype __attribute__ ((mode (TI)));
 #include "longlong.h"
 #endif
 

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 18:31       ` Richard Sandiford
@ 2014-05-08 18:37         ` Kenneth Zadeck
  2014-05-08 19:18           ` Richard Sandiford
  2014-05-09  9:01         ` Richard Earnshaw
  1 sibling, 1 reply; 24+ messages in thread
From: Kenneth Zadeck @ 2014-05-08 18:37 UTC (permalink / raw)
  To: Joseph S. Myers, ramrad01, gcc-patches, Mike Stump,
	Richard Henderson, rdsandiford

everyone who has a private port will hate you forever.   note that i 
have 2 of them.
On 05/08/2014 02:31 PM, Richard Sandiford wrote:
> "Joseph S. Myers" <joseph@codesourcery.com> writes:
>> On Thu, 8 May 2014, Ramana Radhakrishnan wrote:
>>
>>> <DATE>  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
>>>
>>>         * wide-int.cc (UTItype): Define.
>>>         (UDWtype): Define for appropriate W_TYPE_SIZE.
>> This breaks builds for 32-bit hosts, where TImode isn't supported.  You
>> can only use TImode on the host if it's 64-bit.
>>
>> wide-int.cc:37:56: error: unable to emulate 'TI'
> The longlong.h interface seems to be designed to be as difficult to use
> as possible :-(  So maybe we really do need to limit it to hosts that are
> known to work and benefit from it.
>
> How about the following?  I tested that it produces identical
> wide-int.o .text for x86_64.
>
> I think additions to or removals from the list should be treated as
> pre-approved.
>
> Thanks,
> Richard
>
>
> gcc/
> 	* wide-int.cc: Only include longlong.h for certain targets.
>
> Index: gcc/wide-int.cc
> ===================================================================
> --- gcc/wide-int.cc	2014-05-08 19:13:15.782158808 +0100
> +++ gcc/wide-int.cc	2014-05-08 19:28:52.880742385 +0100
> @@ -27,19 +27,20 @@ along with GCC; see the file COPYING3.
>   #include "tree.h"
>   #include "dumpfile.h"
>   
> -#if GCC_VERSION >= 3000
> +#if (GCC_VERSION >= 3000 \
> +     && (defined __aarch64 \
> +         || defined __alpha \
> +         || defined __ia64 \
> +         || defined __powerpc64__ \
> +         || defined __sparcv9 \
> +         || defined __x86_64__))
>   #define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
> -typedef unsigned HOST_HALF_WIDE_INT UHWtype;
> -typedef unsigned HOST_WIDE_INT UWtype;
>   typedef unsigned int UQItype __attribute__ ((mode (QI)));
>   typedef unsigned int USItype __attribute__ ((mode (SI)));
>   typedef unsigned int UDItype __attribute__ ((mode (DI)));
> -typedef unsigned int UTItype __attribute__ ((mode (TI)));
> -#if W_TYPE_SIZE == 32
> -# define UDWtype       UDItype
> -#elif W_TYPE_SIZE == 64
> -# define UDWtype       UTItype
> -#endif
> +typedef unsigned HOST_HALF_WIDE_INT UHWtype;
> +typedef unsigned HOST_WIDE_INT UWtype;
> +typedef unsigned int UDWtype __attribute__ ((mode (TI)));
>   #include "longlong.h"
>   #endif
>   

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 18:37         ` Kenneth Zadeck
@ 2014-05-08 19:18           ` Richard Sandiford
  2014-05-08 19:34             ` H.J. Lu
  0 siblings, 1 reply; 24+ messages in thread
From: Richard Sandiford @ 2014-05-08 19:18 UTC (permalink / raw)
  To: Kenneth Zadeck
  Cc: Joseph S. Myers, ramrad01, gcc-patches, Mike Stump, Richard Henderson

Kenneth Zadeck <zadeck@naturalbridge.com> writes:
> everyone who has a private port will hate you forever.   note that i 
> have 2 of them.

Got any other ideas though?  I suppose if we're prepared to break
compatibility with whatever the upstream of longlong.h is, we could
make more use of intN_t and uintN_t.

Having a whitelist of hosts seems like the best fix though.
I'm not sure the default umul_ppmm is going to be any better
than not defining it.

Thanks,
Richard

> On 05/08/2014 02:31 PM, Richard Sandiford wrote:
>> "Joseph S. Myers" <joseph@codesourcery.com> writes:
>>> On Thu, 8 May 2014, Ramana Radhakrishnan wrote:
>>>
>>>> <DATE>  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
>>>>
>>>>         * wide-int.cc (UTItype): Define.
>>>>         (UDWtype): Define for appropriate W_TYPE_SIZE.
>>> This breaks builds for 32-bit hosts, where TImode isn't supported.  You
>>> can only use TImode on the host if it's 64-bit.
>>>
>>> wide-int.cc:37:56: error: unable to emulate 'TI'
>> The longlong.h interface seems to be designed to be as difficult to use
>> as possible :-(  So maybe we really do need to limit it to hosts that are
>> known to work and benefit from it.
>>
>> How about the following?  I tested that it produces identical
>> wide-int.o .text for x86_64.
>>
>> I think additions to or removals from the list should be treated as
>> pre-approved.
>>
>> Thanks,
>> Richard
>>
>>
>> gcc/
>> 	* wide-int.cc: Only include longlong.h for certain targets.
>>
>> Index: gcc/wide-int.cc
>> ===================================================================
>> --- gcc/wide-int.cc	2014-05-08 19:13:15.782158808 +0100
>> +++ gcc/wide-int.cc	2014-05-08 19:28:52.880742385 +0100
>> @@ -27,19 +27,20 @@ along with GCC; see the file COPYING3.
>>   #include "tree.h"
>>   #include "dumpfile.h"
>>   
>> -#if GCC_VERSION >= 3000
>> +#if (GCC_VERSION >= 3000 \
>> +     && (defined __aarch64 \
>> +         || defined __alpha \
>> +         || defined __ia64 \
>> +         || defined __powerpc64__ \
>> +         || defined __sparcv9 \
>> +         || defined __x86_64__))
>>   #define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
>> -typedef unsigned HOST_HALF_WIDE_INT UHWtype;
>> -typedef unsigned HOST_WIDE_INT UWtype;
>>   typedef unsigned int UQItype __attribute__ ((mode (QI)));
>>   typedef unsigned int USItype __attribute__ ((mode (SI)));
>>   typedef unsigned int UDItype __attribute__ ((mode (DI)));
>> -typedef unsigned int UTItype __attribute__ ((mode (TI)));
>> -#if W_TYPE_SIZE == 32
>> -# define UDWtype       UDItype
>> -#elif W_TYPE_SIZE == 64
>> -# define UDWtype       UTItype
>> -#endif
>> +typedef unsigned HOST_HALF_WIDE_INT UHWtype;
>> +typedef unsigned HOST_WIDE_INT UWtype;
>> +typedef unsigned int UDWtype __attribute__ ((mode (TI)));
>>   #include "longlong.h"
>>   #endif
>>   

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 19:18           ` Richard Sandiford
@ 2014-05-08 19:34             ` H.J. Lu
  2014-05-08 19:42               ` Jakub Jelinek
  0 siblings, 1 reply; 24+ messages in thread
From: H.J. Lu @ 2014-05-08 19:34 UTC (permalink / raw)
  To: Kenneth Zadeck, Joseph S. Myers, ramrad01, gcc-patches,
	Mike Stump, Richard Henderson, Richard Sandiford

On Thu, May 8, 2014 at 12:18 PM, Richard Sandiford
<rdsandiford@googlemail.com> wrote:
> Kenneth Zadeck <zadeck@naturalbridge.com> writes:
>> everyone who has a private port will hate you forever.   note that i
>> have 2 of them.
>
> Got any other ideas though?  I suppose if we're prepared to break
> compatibility with whatever the upstream of longlong.h is, we could
> make more use of intN_t and uintN_t.
>
> Having a whitelist of hosts seems like the best fix though.
> I'm not sure the default umul_ppmm is going to be any better
> than not defining it.
>

Can you add a configure time check if

typedef unsigned int UTItype __attribute__ ((mode (TI)));

is supported?

-- 
H.J.

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 19:34             ` H.J. Lu
@ 2014-05-08 19:42               ` Jakub Jelinek
  2014-05-08 19:48                 ` H.J. Lu
  0 siblings, 1 reply; 24+ messages in thread
From: Jakub Jelinek @ 2014-05-08 19:42 UTC (permalink / raw)
  To: H.J. Lu
  Cc: Kenneth Zadeck, Joseph S. Myers, ramrad01, gcc-patches,
	Mike Stump, Richard Henderson, Richard Sandiford

On Thu, May 08, 2014 at 12:34:28PM -0700, H.J. Lu wrote:
> On Thu, May 8, 2014 at 12:18 PM, Richard Sandiford
> <rdsandiford@googlemail.com> wrote:
> > Kenneth Zadeck <zadeck@naturalbridge.com> writes:
> >> everyone who has a private port will hate you forever.   note that i
> >> have 2 of them.
> >
> > Got any other ideas though?  I suppose if we're prepared to break
> > compatibility with whatever the upstream of longlong.h is, we could
> > make more use of intN_t and uintN_t.
> >
> > Having a whitelist of hosts seems like the best fix though.
> > I'm not sure the default umul_ppmm is going to be any better
> > than not defining it.
> >
> 
> Can you add a configure time check if
> 
> typedef unsigned int UTItype __attribute__ ((mode (TI)));
> 
> is supported?

Why?  Isn't that #ifdef __SIZEOF_INT128__ ?

	Jakub

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 19:42               ` Jakub Jelinek
@ 2014-05-08 19:48                 ` H.J. Lu
  2014-05-08 20:12                   ` Richard Sandiford
  0 siblings, 1 reply; 24+ messages in thread
From: H.J. Lu @ 2014-05-08 19:48 UTC (permalink / raw)
  To: Jakub Jelinek
  Cc: Kenneth Zadeck, Joseph S. Myers, ramrad01, gcc-patches,
	Mike Stump, Richard Henderson, Richard Sandiford

On Thu, May 8, 2014 at 12:42 PM, Jakub Jelinek <jakub@redhat.com> wrote:
> On Thu, May 08, 2014 at 12:34:28PM -0700, H.J. Lu wrote:
>> On Thu, May 8, 2014 at 12:18 PM, Richard Sandiford
>> <rdsandiford@googlemail.com> wrote:
>> > Kenneth Zadeck <zadeck@naturalbridge.com> writes:
>> >> everyone who has a private port will hate you forever.   note that i
>> >> have 2 of them.
>> >
>> > Got any other ideas though?  I suppose if we're prepared to break
>> > compatibility with whatever the upstream of longlong.h is, we could
>> > make more use of intN_t and uintN_t.
>> >
>> > Having a whitelist of hosts seems like the best fix though.
>> > I'm not sure the default umul_ppmm is going to be any better
>> > than not defining it.
>> >
>>
>> Can you add a configure time check if
>>
>> typedef unsigned int UTItype __attribute__ ((mode (TI)));
>>
>> is supported?
>
> Why?  Isn't that #ifdef __SIZEOF_INT128__ ?
>

Yes, we can use that.  Will it work?


-- 
H.J.

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 19:48                 ` H.J. Lu
@ 2014-05-08 20:12                   ` Richard Sandiford
  2014-05-08 20:48                     ` H.J. Lu
  2014-05-12 18:23                     ` Richard Henderson
  0 siblings, 2 replies; 24+ messages in thread
From: Richard Sandiford @ 2014-05-08 20:12 UTC (permalink / raw)
  To: H.J. Lu
  Cc: Jakub Jelinek, Kenneth Zadeck, Joseph S. Myers, ramrad01,
	gcc-patches, Mike Stump, Richard Henderson

"H.J. Lu" <hjl.tools@gmail.com> writes:
> On Thu, May 8, 2014 at 12:42 PM, Jakub Jelinek <jakub@redhat.com> wrote:
>> On Thu, May 08, 2014 at 12:34:28PM -0700, H.J. Lu wrote:
>>> On Thu, May 8, 2014 at 12:18 PM, Richard Sandiford
>>> <rdsandiford@googlemail.com> wrote:
>>> > Kenneth Zadeck <zadeck@naturalbridge.com> writes:
>>> >> everyone who has a private port will hate you forever.   note that i
>>> >> have 2 of them.
>>> >
>>> > Got any other ideas though?  I suppose if we're prepared to break
>>> > compatibility with whatever the upstream of longlong.h is, we could
>>> > make more use of intN_t and uintN_t.
>>> >
>>> > Having a whitelist of hosts seems like the best fix though.
>>> > I'm not sure the default umul_ppmm is going to be any better
>>> > than not defining it.
>>> >
>>>
>>> Can you add a configure time check if
>>>
>>> typedef unsigned int UTItype __attribute__ ((mode (TI)));
>>>
>>> is supported?
>>
>> Why?  Isn't that #ifdef __SIZEOF_INT128__ ?

Read this just after getting the configure test to work...

> Yes, we can use that.  Will it work?

Seems to.  How does this look?

Thanks,
Richard


gcc/
	* wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
	__SIZEOF_INT128__ is defined.

Index: gcc/wide-int.cc
===================================================================
--- gcc/wide-int.cc	2014-05-08 20:48:25.341583885 +0100
+++ gcc/wide-int.cc	2014-05-08 21:09:29.324386217 +0100
@@ -27,18 +27,17 @@ along with GCC; see the file COPYING3.
 #include "tree.h"
 #include "dumpfile.h"
 
-#if GCC_VERSION >= 3000
 #define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
+#if GCC_VERSION >= 3000 && (W_TYPE_SIZE == 32 || defined (__SIZEOF_INT128__))
 typedef unsigned HOST_HALF_WIDE_INT UHWtype;
 typedef unsigned HOST_WIDE_INT UWtype;
 typedef unsigned int UQItype __attribute__ ((mode (QI)));
 typedef unsigned int USItype __attribute__ ((mode (SI)));
 typedef unsigned int UDItype __attribute__ ((mode (DI)));
-typedef unsigned int UTItype __attribute__ ((mode (TI)));
 #if W_TYPE_SIZE == 32
-# define UDWtype       UDItype
-#elif W_TYPE_SIZE == 64
-# define UDWtype       UTItype
+typedef unsigned int UDWtype __attribute__ ((mode (DI)));
+#else
+typedef unsigned int UDWtype __attribute__ ((mode (TI)));
 #endif
 #include "longlong.h"
 #endif

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 20:12                   ` Richard Sandiford
@ 2014-05-08 20:48                     ` H.J. Lu
  2014-05-08 21:04                       ` Richard Sandiford
  2014-05-12 18:23                     ` Richard Henderson
  1 sibling, 1 reply; 24+ messages in thread
From: H.J. Lu @ 2014-05-08 20:48 UTC (permalink / raw)
  To: Jakub Jelinek, Kenneth Zadeck, Joseph S. Myers, ramrad01,
	gcc-patches, Mike Stump, Richard Henderson, Richard Sandiford

On Thu, May 8, 2014 at 1:12 PM, Richard Sandiford
<rdsandiford@googlemail.com> wrote:
> "H.J. Lu" <hjl.tools@gmail.com> writes:
>> On Thu, May 8, 2014 at 12:42 PM, Jakub Jelinek <jakub@redhat.com> wrote:
>>> On Thu, May 08, 2014 at 12:34:28PM -0700, H.J. Lu wrote:
>>>> On Thu, May 8, 2014 at 12:18 PM, Richard Sandiford
>>>> <rdsandiford@googlemail.com> wrote:
>>>> > Kenneth Zadeck <zadeck@naturalbridge.com> writes:
>>>> >> everyone who has a private port will hate you forever.   note that i
>>>> >> have 2 of them.
>>>> >
>>>> > Got any other ideas though?  I suppose if we're prepared to break
>>>> > compatibility with whatever the upstream of longlong.h is, we could
>>>> > make more use of intN_t and uintN_t.
>>>> >
>>>> > Having a whitelist of hosts seems like the best fix though.
>>>> > I'm not sure the default umul_ppmm is going to be any better
>>>> > than not defining it.
>>>> >
>>>>
>>>> Can you add a configure time check if
>>>>
>>>> typedef unsigned int UTItype __attribute__ ((mode (TI)));
>>>>
>>>> is supported?
>>>
>>> Why?  Isn't that #ifdef __SIZEOF_INT128__ ?
>
> Read this just after getting the configure test to work...
>
>> Yes, we can use that.  Will it work?
>
> Seems to.  How does this look?
>
> Thanks,
> Richard
>
>
> gcc/
>         * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
>         __SIZEOF_INT128__ is defined.
>
> Index: gcc/wide-int.cc
> ===================================================================
> --- gcc/wide-int.cc     2014-05-08 20:48:25.341583885 +0100
> +++ gcc/wide-int.cc     2014-05-08 21:09:29.324386217 +0100
> @@ -27,18 +27,17 @@ along with GCC; see the file COPYING3.
>  #include "tree.h"
>  #include "dumpfile.h"
>
> -#if GCC_VERSION >= 3000
>  #define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT

Isn't  HOST_BITS_PER_WIDE_INT always 64 now?

> +#if GCC_VERSION >= 3000 && (W_TYPE_SIZE == 32 || defined (__SIZEOF_INT128__))

W_TYPE_SIZE == 32 is always false and on 32-bit hosts,
__SIZEOF_INT128__ won't be defined.

>  typedef unsigned HOST_HALF_WIDE_INT UHWtype;
>  typedef unsigned HOST_WIDE_INT UWtype;
>  typedef unsigned int UQItype __attribute__ ((mode (QI)));
>  typedef unsigned int USItype __attribute__ ((mode (SI)));
>  typedef unsigned int UDItype __attribute__ ((mode (DI)));
> -typedef unsigned int UTItype __attribute__ ((mode (TI)));
>  #if W_TYPE_SIZE == 32
> -# define UDWtype       UDItype
> -#elif W_TYPE_SIZE == 64
> -# define UDWtype       UTItype
> +typedef unsigned int UDWtype __attribute__ ((mode (DI)));

Can't we use

#ifndef __SIZEOF_INT128__

here?

> +#else
> +typedef unsigned int UDWtype __attribute__ ((mode (TI)));
>  #endif
>  #include "longlong.h"
>  #endif



-- 
H.J.

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 20:48                     ` H.J. Lu
@ 2014-05-08 21:04                       ` Richard Sandiford
  2014-05-09  8:34                         ` Ramana Radhakrishnan
  0 siblings, 1 reply; 24+ messages in thread
From: Richard Sandiford @ 2014-05-08 21:04 UTC (permalink / raw)
  To: H.J. Lu
  Cc: Jakub Jelinek, Kenneth Zadeck, Joseph S. Myers, ramrad01,
	gcc-patches, Mike Stump, Richard Henderson

"H.J. Lu" <hjl.tools@gmail.com> writes:
> On Thu, May 8, 2014 at 1:12 PM, Richard Sandiford
> <rdsandiford@googlemail.com> wrote:
>> "H.J. Lu" <hjl.tools@gmail.com> writes:
>>> On Thu, May 8, 2014 at 12:42 PM, Jakub Jelinek <jakub@redhat.com> wrote:
>>>> On Thu, May 08, 2014 at 12:34:28PM -0700, H.J. Lu wrote:
>>>>> On Thu, May 8, 2014 at 12:18 PM, Richard Sandiford
>>>>> <rdsandiford@googlemail.com> wrote:
>>>>> > Kenneth Zadeck <zadeck@naturalbridge.com> writes:
>>>>> >> everyone who has a private port will hate you forever.   note that i
>>>>> >> have 2 of them.
>>>>> >
>>>>> > Got any other ideas though?  I suppose if we're prepared to break
>>>>> > compatibility with whatever the upstream of longlong.h is, we could
>>>>> > make more use of intN_t and uintN_t.
>>>>> >
>>>>> > Having a whitelist of hosts seems like the best fix though.
>>>>> > I'm not sure the default umul_ppmm is going to be any better
>>>>> > than not defining it.
>>>>> >
>>>>>
>>>>> Can you add a configure time check if
>>>>>
>>>>> typedef unsigned int UTItype __attribute__ ((mode (TI)));
>>>>>
>>>>> is supported?
>>>>
>>>> Why?  Isn't that #ifdef __SIZEOF_INT128__ ?
>>
>> Read this just after getting the configure test to work...
>>
>>> Yes, we can use that.  Will it work?
>>
>> Seems to.  How does this look?
>>
>> Thanks,
>> Richard
>>
>>
>> gcc/
>>         * wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
>>         __SIZEOF_INT128__ is defined.
>>
>> Index: gcc/wide-int.cc
>> ===================================================================
>> --- gcc/wide-int.cc     2014-05-08 20:48:25.341583885 +0100
>> +++ gcc/wide-int.cc     2014-05-08 21:09:29.324386217 +0100
>> @@ -27,18 +27,17 @@ along with GCC; see the file COPYING3.
>>  #include "tree.h"
>>  #include "dumpfile.h"
>>
>> -#if GCC_VERSION >= 3000
>>  #define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
>
> Isn't  HOST_BITS_PER_WIDE_INT always 64 now?

Right, but like I said in reply to Ramana's email, there's no guarantee
that we'll continue to use HOST_WIDE_INT for wide-int.h.  I'd rather
keep this code parameterised on W_TYPE_SIZE rather than hard-code
W_TYPE_SIZE==64.

>> +#if GCC_VERSION >= 3000 && (W_TYPE_SIZE == 32 || defined (__SIZEOF_INT128__))
>
> W_TYPE_SIZE == 32 is always false and on 32-bit hosts,
> __SIZEOF_INT128__ won't be defined.

Right, but we won't try to use TImode if in future we do revert to using
32-bit types for 32-bit hosts.

Thanks,
Richard

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 21:04                       ` Richard Sandiford
@ 2014-05-09  8:34                         ` Ramana Radhakrishnan
  0 siblings, 0 replies; 24+ messages in thread
From: Ramana Radhakrishnan @ 2014-05-09  8:34 UTC (permalink / raw)
  To: H.J. Lu, Jakub Jelinek, Kenneth Zadeck, Joseph S. Myers,
	ramrad01, gcc-patches, Mike Stump, Richard Henderson,
	rdsandiford


>>> +#if GCC_VERSION >= 3000 && (W_TYPE_SIZE == 32 || defined (__SIZEOF_INT128__))
>>
>> W_TYPE_SIZE == 32 is always false and on 32-bit hosts,
>> __SIZEOF_INT128__ won't be defined.
>
> Right, but we won't try to use TImode if in future we do revert to using
> 32-bit types for 32-bit hosts.

For now, I've reverted my commit until we agree on the best approach 
here as the fallout has been greater than I expected.

I wasn't looking at email last night or would have done this sooner.

Apologies for the breakage.

Ramana

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 18:31       ` Richard Sandiford
  2014-05-08 18:37         ` Kenneth Zadeck
@ 2014-05-09  9:01         ` Richard Earnshaw
  1 sibling, 0 replies; 24+ messages in thread
From: Richard Earnshaw @ 2014-05-09  9:01 UTC (permalink / raw)
  To: Joseph S. Myers, ramrad01, gcc-patches, zadeck, Mike Stump,
	Richard Henderson, rdsandiford

On 08/05/14 19:31, Richard Sandiford wrote:
> "Joseph S. Myers" <joseph@codesourcery.com> writes:
>> On Thu, 8 May 2014, Ramana Radhakrishnan wrote:
>>
>>> <DATE>  Ramana Radhakrishnan  <ramana.radhakrishnan@arm.com>
>>>
>>>        * wide-int.cc (UTItype): Define.
>>>        (UDWtype): Define for appropriate W_TYPE_SIZE.
>>
>> This breaks builds for 32-bit hosts, where TImode isn't supported.  You 
>> can only use TImode on the host if it's 64-bit.
>>
>> wide-int.cc:37:56: error: unable to emulate 'TI'
> 
> The longlong.h interface seems to be designed to be as difficult to use
> as possible :-(  So maybe we really do need to limit it to hosts that are
> known to work and benefit from it.
> 
> How about the following?  I tested that it produces identical
> wide-int.o .text for x86_64.
> 
> I think additions to or removals from the list should be treated as
> pre-approved.
> 
> Thanks,
> Richard
> 
> 
> gcc/
> 	* wide-int.cc: Only include longlong.h for certain targets.
> 
> Index: gcc/wide-int.cc
> ===================================================================
> --- gcc/wide-int.cc	2014-05-08 19:13:15.782158808 +0100
> +++ gcc/wide-int.cc	2014-05-08 19:28:52.880742385 +0100
> @@ -27,19 +27,20 @@ along with GCC; see the file COPYING3.
>  #include "tree.h"
>  #include "dumpfile.h"
>  
> -#if GCC_VERSION >= 3000
> +#if (GCC_VERSION >= 3000 \
> +     && (defined __aarch64 \

Needs tailing underscores: __aarch64__

R.

> +         || defined __alpha \
> +         || defined __ia64 \
> +         || defined __powerpc64__ \
> +         || defined __sparcv9 \
> +         || defined __x86_64__))
>  #define W_TYPE_SIZE HOST_BITS_PER_WIDE_INT
> -typedef unsigned HOST_HALF_WIDE_INT UHWtype;
> -typedef unsigned HOST_WIDE_INT UWtype;
>  typedef unsigned int UQItype __attribute__ ((mode (QI)));
>  typedef unsigned int USItype __attribute__ ((mode (SI)));
>  typedef unsigned int UDItype __attribute__ ((mode (DI)));
> -typedef unsigned int UTItype __attribute__ ((mode (TI)));
> -#if W_TYPE_SIZE == 32
> -# define UDWtype       UDItype
> -#elif W_TYPE_SIZE == 64
> -# define UDWtype       UTItype
> -#endif
> +typedef unsigned HOST_HALF_WIDE_INT UHWtype;
> +typedef unsigned HOST_WIDE_INT UWtype;
> +typedef unsigned int UDWtype __attribute__ ((mode (TI)));
>  #include "longlong.h"
>  #endif
>  
> 


^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-08 20:12                   ` Richard Sandiford
  2014-05-08 20:48                     ` H.J. Lu
@ 2014-05-12 18:23                     ` Richard Henderson
  2014-05-17  8:30                       ` Richard Sandiford
  1 sibling, 1 reply; 24+ messages in thread
From: Richard Henderson @ 2014-05-12 18:23 UTC (permalink / raw)
  To: H.J. Lu, Jakub Jelinek, Kenneth Zadeck, Joseph S. Myers,
	ramrad01, gcc-patches, Mike Stump, rdsandiford

On 05/08/2014 01:12 PM, Richard Sandiford wrote:
> 	* wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
> 	__SIZEOF_INT128__ is defined.

FWIW, this looks pretty good to me.


r~

^ permalink raw reply	[flat|nested] 24+ messages in thread

* Re: [wide-int] Add fast path for hosts with HWI widening multiplication
  2014-05-12 18:23                     ` Richard Henderson
@ 2014-05-17  8:30                       ` Richard Sandiford
  0 siblings, 0 replies; 24+ messages in thread
From: Richard Sandiford @ 2014-05-17  8:30 UTC (permalink / raw)
  To: Richard Henderson
  Cc: H.J. Lu, Jakub Jelinek, Kenneth Zadeck, Joseph S. Myers,
	ramrad01, gcc-patches, Mike Stump

Richard Henderson <rth@redhat.com> writes:
> On 05/08/2014 01:12 PM, Richard Sandiford wrote:
>> 	* wide-int.cc: Only include longlong.h if W_TYPE_SIZE==32 or
>> 	__SIZEOF_INT128__ is defined.
>
> FWIW, this looks pretty good to me.

Thanks, belatedly applied.

Richard

^ permalink raw reply	[flat|nested] 24+ messages in thread

end of thread, other threads:[~2014-05-17  8:30 UTC | newest]

Thread overview: 24+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-12-01 10:29 [wide-int] Add fast path for hosts with HWI widening multiplication Richard Sandiford
2013-12-01 18:48 ` Ian Lance Taylor
2013-12-02 19:35   ` Richard Sandiford
2013-12-02 19:48     ` Paolo Bonzini
2013-12-03 23:55       ` Richard Sandiford
2013-12-04  0:18         ` Ian Lance Taylor
2013-12-04 12:56 ` Richard Sandiford
2013-12-04 14:04   ` Kenneth Zadeck
2014-05-08 12:49   ` Ramana Radhakrishnan
2014-05-08 13:13     ` Richard Sandiford
2014-05-08 16:38     ` Joseph S. Myers
2014-05-08 18:31       ` Richard Sandiford
2014-05-08 18:37         ` Kenneth Zadeck
2014-05-08 19:18           ` Richard Sandiford
2014-05-08 19:34             ` H.J. Lu
2014-05-08 19:42               ` Jakub Jelinek
2014-05-08 19:48                 ` H.J. Lu
2014-05-08 20:12                   ` Richard Sandiford
2014-05-08 20:48                     ` H.J. Lu
2014-05-08 21:04                       ` Richard Sandiford
2014-05-09  8:34                         ` Ramana Radhakrishnan
2014-05-12 18:23                     ` Richard Henderson
2014-05-17  8:30                       ` Richard Sandiford
2014-05-09  9:01         ` Richard Earnshaw

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).