This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [wide-int] Add fast path for hosts with HWI widening multiplication


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.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]