This is the mail archive of the cygwin mailing list for the Cygwin 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: i686-pc-cygwin on an i586

Corinna wrote:

> On Jan 28 01:33, Christopher Faylor wrote:
>> On Fri, Jan 27, 2006 at 03:06:44PM -0700, Eric Blake wrote:
>> >According to Dave Korn on 1/27/2006 9:34 AM:
>> >>Nope, don't worry about it, that's a bit of a red-herring.  By default,
>> >>the code gcc generates is good for everything from '486 up.  The
>> >>instruction scheduling and choice of which instructions to use may be
>> >>tuned to be optimal for a 686 and so may be less-than-optimal on a
>> >>'586, but there should not be any actual backward-compatibility issues.
>> >
>> >Speaking of which, should the next release of cygwin gcc be configured
>> >to generate code tuned for 686, rather than penalizing most modern CPUs
>> >with 386-compatible but slower code sequences?
>> Why do you assume that this is not already the case?  I use i686-pc-cygwin
>> as the target for everything that I build and I use a i686-pc-cygwin-gcc
>> cross compiler.

> Same here, same for the net distro itself.  AFAIK we're generating
> i686-opimized code by default for at least three years, don't we?

The GCC release is compiled to support 486 and up targets.  386 is not
supported anymore.  Building binaries optimized for i686 may create not
working binaries if you have eg. an i586 compatible AMD processor,
howwever, the only time I had problems was with GCC itself optimized for
i686 which was crashing on my old AMD box.


Unsubscribe info:
Problem reports:

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