[MPlayer-users] Building with SSE3 support

[Jason Garrett-Glaser]

On Sat, Oct 18, 2008 at 6:22 AM, Andrew Berg <bahamutzero8825 at gmail.com> wrote:
> Guillaume POIRIER wrote:
>> On Sat, Oct 18, 2008 at 1:40 PM, Andrew Berg <bahamutzero8825 at gmail.com> wrote:
>> > I'm building MPlayer 1.0rc2 under Vista with a MinGW environment and I
>> > have an Intel Core 2 Duo processor. The configure script seems to ignore
>> > the --enables-ssse3 switch (why 3 's'es?) and doesn't check for SSE3
>> > support (and builds for pentium-m instead of prescott). It does, however
>> > correctly detect my processor.
>>
>> SSSE3 and SSE3 are 2 different things:
>> http://en.wikipedia.org/wiki/SSE3
>> http://en.wikipedia.org/wiki/SSSE3
>>
>> Yes, the naming scheme is ugly.
>>
> Thanks. That explains it. It's too bad that there isn't a GCC for
> Windows (out of alpha at least) that supports Core 2 optimizations. :-(

As far as I've experienced, there is zero measurable benefit from
-march higher than pentium2 (except for SSE float math maybe), and
GCC's autovectorization is patently worthless.

Dark Shikari