The attached patch changes the normalization to the version I was
suggesting. I added some print statements and discovered that the "residual
error" was always zero, after some analysis I figured out that this version
will always produce a set of values that sum to pixman_fixed_1 (unless the
filter
Never mind, I should have just clicked on that rather than thinking it was
a repository name.
I am able to compile your version and run it and the output looks correct.
Changing the plot line to this:
printf ("plot [x=%g:%g] '-' with linespoints ls 1\n", -width*0.5,
width*0.5);
does make i
On 04/12/2016 01:55 PM, Søren Sandmann wrote:
This series is available as a git repository here:
https://cgit.freedesktop.org/~sandmann/pixman/log/?h=spitzak-for-master
Not having much luck with that url:
fatal: repository
'https://cgit.freedesktop.org/~sandmann/pixman/log/?h=spitzak-for-ma
Hi.
In aarch64, I cannot find corresponding instruction of this patch,
but
> movi v2.2s, 0xff, lsl #24
works well as your modification.
On 11 April 2016 at 21:26, Ben Avison wrote:
> A minor point, but 0xFF00 is already a valid immediate constant for
> NEON, there's no need to construct i