On Wed, Sep 26, 2018 at 05:06:46PM +0200, Nicolas George wrote:
> Michael Niedermayer (2018-09-24):
> > CC-ing you so you dont miss this (though its not important)
> >
> > if i hear nothing from anyone i will apply the patchset as is.
> > IIUC james is ok with that.
> > But as said i can redo it
Michael Niedermayer (2018-09-24):
> CC-ing you so you dont miss this (though its not important)
>
> if i hear nothing from anyone i will apply the patchset as is.
> IIUC james is ok with that.
> But as said i can redo it with unsigned if people want.
> Just dont want to spend time redoing a patch
On Mon, Aug 20, 2018 at 12:42:29AM +0200, Michael Niedermayer wrote:
> On Sun, Aug 05, 2018 at 11:55:06PM +0200, Nicolas George wrote:
> > Michael Niedermayer (2018-08-05):
> > > thats done in the next patch
> >
> > Indeed, except for the part where the variable are turned to unsigned to
> > let t
On Sun, Aug 05, 2018 at 11:55:06PM +0200, Nicolas George wrote:
> Michael Niedermayer (2018-08-05):
> > thats done in the next patch
>
> Indeed, except for the part where the variable are turned to unsigned to
> let the compiler optimize the division. Sorry to have missed that.
are you ok with le
On 8/5/2018 6:58 PM, Michael Niedermayer wrote:
> On Sun, Aug 05, 2018 at 06:19:01PM -0300, James Almer wrote:
>> On 8/5/2018 5:29 PM, Michael Niedermayer wrote:
>>> Divisions tend to be slower than shifts unless the compiler optimizes them
>>> out.
>>> And some of these are in inner loops.
>>
>>
On Sun, Aug 05, 2018 at 06:19:01PM -0300, James Almer wrote:
> On 8/5/2018 5:29 PM, Michael Niedermayer wrote:
> > Divisions tend to be slower than shifts unless the compiler optimizes them
> > out.
> > And some of these are in inner loops.
>
> This patch makes the code slightly less readable, IM
Michael Niedermayer (2018-08-05):
> thats done in the next patch
Indeed, except for the part where the variable are turned to unsigned to
let the compiler optimize the division. Sorry to have missed that.
Regards,
--
Nicolas George
signature.asc
Description: Digital signature
__
On Sun, Aug 05, 2018 at 11:37:24PM +0200, Nicolas George wrote:
> James Almer (2018-08-05):
> > This patch makes the code slightly less readable, IMO. What compiler
> > doesn't optimize out an integer division by 2 into a right shift?
>
> Any compiler that cannot figure out that x is positive in t
James Almer (2018-08-05):
> This patch makes the code slightly less readable, IMO. What compiler
> doesn't optimize out an integer division by 2 into a right shift?
Any compiler that cannot figure out that x is positive in this
convoluted loop. Signed divisions cannot be optimized as shifts. Make
On 8/5/2018 5:29 PM, Michael Niedermayer wrote:
> Divisions tend to be slower than shifts unless the compiler optimizes them
> out.
> And some of these are in inner loops.
This patch makes the code slightly less readable, IMO. What compiler
doesn't optimize out an integer division by 2 into a rig
Divisions tend to be slower than shifts unless the compiler optimizes them out.
And some of these are in inner loops.
Signed-off-by: Michael Niedermayer
---
libavcodec/gdv.c | 16
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/libavcodec/gdv.c b/libavcodec/gdv.c
i
11 matches
Mail list logo