| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Avoids duplication of code
|
|
It's basically the 32bit version of pixman_image_fill_rectangles(), just
with a saner data type.
|
|
This is equal to pixman_image_composite(), just with 32bit parameters.
pixman_image_composite() now just calls pixman_image_composite32()
|
|
No indenting of the header to keep git blame working
|
|
|
|
There is a couple of bugs in bugzilla where bugs in the X server
triggered asserts in the pixman region code. It is probably better to
let the X server survive this. (In fact, I thought I had disabled them
for 0.16.0, but apparently not).
The patch below uses these rules:
- In _stable_ pixman releases, assertions and selfchecks are turned
off. Assertions, so that the X server doesn't die. Selfchecks,
for performance reasons.
- In _unstable_ pixman releases, both assertions and selfcheck are
turned on. These releases are what get added to development
distributions such as rawhide, so we want as much self-checking
as possible.
- In _random git checkouts_, assertions are enabled, so that bugs
are caught, but selfchecks are disabled so that you can use them
for performance work without having to fiddle with turning
selfchecks off.
|
|
|
|
|
|
|
|
|
|
|
|
Setting TESTS will run the tests on `make check`
Bug 25131
|
|
|
|
|
|
Existing template already supports 2D images processing,
but pixman also needs some NEON optimized functions for
improving performance when compositing is decoupled
into "fetch -> process -> store" stages and done via
temporary scanline buffer. That's why a new simplified
template which deals only with the generation of single
scanline processing functions is handy.
|
|
|
|
This is ARM NEON optimized conversion of native RGBA format used by
GTK/GDK into native 32bpp RGBA format used by cairo/pixman.
|
|
This is ARM NEON optimized conversion of native RGB format used by
GTK/GDK into r5g6b5 format.
|
|
This is ARM NEON optimized conversion of native RGB format used by
GTK/GDK into native 32bpp RGB format used by cairo/pixman.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Now it's a bit harder to exceed 80 characters line limit
when binding assembly functions.
|
|
This fix prevents build failure due to not accepting PLD instruction when
compiling for armv4 cpu with the relevant -mcpu/-march options set in CFLAGS.
|
|
|
|
|
|
NEON unit has fast access to L1/L2 caches and even simple
copy of memory buffers using NEON provides more than 1.5x
performance improvement on ARM Cortex-A8.
|
|
|
|
|
|
|
|
|
|
Pointed out by scottmc2@gmail.com in bug 25137.
|
|
This function is needed to improve performance of xfce4 terminal when
using bitmap fonts and running with 16bpp desktop. Some other applications
may potentially benefit too.
After applying this patch, top functions from Xorg process in
oprofile log change from
samples % image name symbol name
13296 29.1528 libpixman-1.so.0.17.1 combine_over_u
6452 14.1466 libpixman-1.so.0.17.1 fetch_scanline_r5g6b5
5516 12.0944 libpixman-1.so.0.17.1 fetch_scanline_a1
2273 4.9838 libpixman-1.so.0.17.1 store_scanline_r5g6b5
1741 3.8173 libpixman-1.so.0.17.1 fast_composite_add_1000_1000
1718 3.7669 libc-2.9.so memcpy
to
samples % image name symbol name
5594 14.7033 libpixman-1.so.0.17.1 fast_composite_over_n_1_0565
4323 11.3626 libc-2.9.so memcpy
3695 9.7119 libpixman-1.so.0.17.1 fast_composite_add_1000_1000
when scrolling text in terminal (reading man page).
|
|
|
|
Instead introduce two new fake formats
PIXMAN_pixbuf
PIXMAN_rpixbuf
and compute whether the source and mask have them in
find_fast_path(). This lead to some duplicate entries in the fast path
tables that could then be removed.
|
|
Inside the loop all we have to do is check that the formats match.
|
|
Instead introduce two new fake formats
PIXMAN_a8r8g8b8_ca
PIXMAN_a8b8g8r8_ca
that are used in the fast path tables for this case.
|
|
This flag was used to indicate that the mask was solid while still
allowing a specific format to be required. However, there is not
actually any need for this because the fast paths all used
_pixman_image_get_solid() which already allowed arbitrary formats.
The one thing that had to be dealt with was component alpha. In
addition to interpreting the presence of the NEED_COMPONENT_ALPHA
flag, we now also interprete the *absence* of this flag as a
requirement that the mask does *not* have component alpha.
Siarhei Siamashka pointed out that the first version of this commit
had a bug, in which a NEED_SOLID_MASK was accidentally not turned into
a PIXMAN_solid in the ARM NEON implementation.
|
|
When the destination buffer is either a8r8g8b8 or x8r8g8b8, we can use
it directly instead of fetching into a temporary buffer. When the
format is x8r8g8b8, we require the operator to not make use of
destination alpha, but when it is a8r8g8b8, there are no restrictions.
This is approximately a 5% speedup on the poppler cairo benchmark:
[ # ] backend test min(s) median(s) stddev. count
Before:
[ 0] image poppler 6.661 6.709 0.59% 6/6
After:
[ 0] image poppler 6.307 6.320 0.12% 5/6
|
|
|
|
|
|
|
|
|
|
|
|
This is a small speedup on the swfdec-youtube benchmark:
Before:
[ 0] image swfdec-youtube 5.789 5.806 0.20% 6/6
After:
[ 0] image swfdec-youtube 5.489 5.524 0.27% 6/6
Ie., approximately 5% faster.
|
|
|
