On 05/10/2015 18:25, Stroller wrote:
>
>> On Sun, 4 October 2015, at 11:38 p.m., Alan McKinnon
>> wrote:
>>
>>> I see that the ffmpeg ebuild requires that mmxext is set if sse is set.
>>>
>>> Isn't mmxext amd-only?
>>
>> Definitely not. Intel designed mmx
>
> The ffmpeg has both "mmx" and "mmxe
Stroller wrote:
>> On Sun, 4 October 2015, at 11:38 p.m., Alan McKinnon
>> wrote:
>>
>>> I see that the ffmpeg ebuild requires that mmxext is set if sse is set.
>>>
>>> Isn't mmxext amd-only?
>> Definitely not. Intel designed mmx
> The ffmpeg has both "mmx" and "mmxext" USE flags - presumably the
> On Sun, 4 October 2015, at 11:38 p.m., Alan McKinnon
> wrote:
>
>> I see that the ffmpeg ebuild requires that mmxext is set if sse is set.
>>
>> Isn't mmxext amd-only?
>
> Definitely not. Intel designed mmx
The ffmpeg has both "mmx" and "mmxext" USE flags - presumably they're not the
same
On Mon, Oct 05, 2015 at 12:38:15AM +0200, Alan McKinnon wrote:
> On 04/10/2015 22:49, James Cloos wrote:
> >
> > And is there really any value from micromanaging things like that in an
> > ebuild?
>
> For ffmpeg? Yes.
Actually, I think modern versions of ffmpeg probably do not benefit as
much an
On 04/10/2015 22:49, James Cloos wrote:
> I see that the ffmpeg ebuild requires that mmxext is set if sse is set.
>
> Isn't mmxext amd-only?
Definitely not. Intel designed mmx
>
> And is there really any value from micromanaging things like that in an
> ebuild?
For ffmpeg? Yes. I imagine users
I see that the ffmpeg ebuild requires that mmxext is set if sse is set.
Isn't mmxext amd-only?
And is there really any value from micromanaging things like that in an
ebuild?
-JimC
--
James Cloos OpenPGP: 0x997A9F17ED7DAEA6
6 matches
Mail list logo