[Development] QTBUG-30440: restricting the SIMD files

Knoll Lars Lars.Knoll at digia.com
Mon Aug 19 13:56:26 CEST 2013


On 19.08.13 09:56, "Koehne Kai" <Kai.Koehne at digia.com> wrote:

>> -----Original Message-----
>> From: development-bounces+kai.koehne=digia.com at qt-project.org
>> [mailto:development-bounces+kai.koehne=digia.com at qt-project.org] On
>> Behalf Of Thiago Macieira
>> Sent: Monday, August 19, 2013 9:42 AM
>> To: development at qt-project.org
>> Subject: Re: [Development] QTBUG-30440: restricting the SIMD files
>> 
>> On segunda-feira, 19 de agosto de 2013 07:37:58, Koehne Kai wrote:
>> > I don't know how big the performance gains really are, but if it's
>> > noticeable, why not switch the default for everyone using the default
>> > mkspec?
>> 
>> Switching the default means making it difficult to unset for those who
>>want it
>> unset.
>
>Well, it's just copying /editing the mkspec ... Arguably not the most
>obvious way to do it, but maybe we should just document it a bit better,
>then :)
>
>> We recommend people set the environment if they want different flags,
>> besides the stock from their compilers.
>
>Where do we recommend this? E.g.
>
>http://qt-project.org/doc/qt-5.0/qtdoc/install-x11.html
>
>just says to run configure ... Actually it doesn't look like CFLAGS,
>LFLAGS is mentioned anywhere in the Qt documentation.
>
>> If we recommend that, why not use
>> that ourselves?
>
>Well, I just think most people compiling on their own will miss this
>optimization then. If we think it's a useful optimization that helps >95%
>of our customers, it should IMO be the default . If not, I'm not sure we
>should apply it to our binary installers either.

I tend to agree with Kai. Why should we penalize 99% of our users to
support a 15 year old CPU architecture? We don't do this with anything
else. Ie. Qt won't compile on a 15 year old Linux distribution, and we
don't support Win98 anymore neither.

Cheers,
Lars





More information about the Development mailing list