This supports a1r5g5b5 source images.
lowlevel-blt-bench results for src_1555_, which does not yet have a
dedicated fast path:
Before After
Mean StdDev Mean StdDev Confidence Change
L1 24.5 0.2 57.0 1.1 100.0% +132.2%
L2 19.3 0.4 41.4 1.0
Perhaps this was considered too obvious to check, but it's possible to
add a test without imposing any run-time overhead by adding
FAST_PATH_BITS_IMAGE to FAST_PATH_STD_DEST_FLAGS.
---
pixman/pixman-private.h |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/pixman/pixman-
This supports r5g6b5 source and desitination images, and a8 source images.
lowlevel-blt-bench results for example operations which use these because
they lack a dedicated fast path at the time of writing:
in_reverse_8_
Before After
Mean StdDev Mean StdDev Confidence
On Sun, 05 Oct 2014 20:03:42 +0100, Søren Sandmann
wrote:
Don't you need "FAST_PATH_STANDARD_DST_FLAGS" in the ITER_DEST case?
The selection of the correct flags in the iter_info_t struct (and in the
pixman_fast_path_t struct too for that matter) always seemed to be
something of a black art t
This is a C fast path, useful for reference or for platforms that don't
have their own fast path for this operation.
lowlevel-blt-bench results on ARMv6:
Before After
Mean StdDev Mean StdDev Confidence Change
L1 12.4 0.1 24.4 0.3 100.0% +97.8%
L2 9.5
This is used instead of the equivalent C fast path.
lowlevel-blt-bench results, compared to no fast path at all:
Before After
Mean StdDev Mean StdDev Confidence Change
L1 12.4 0.1 117.5 2.3 100.0% +851.2%
L2 9.50.1 46.9 2.4 100.0% +3
This is tuned for the Cortex-A7 (Raspberry Pi 2).
lowlevel-blt-bench results, compared to the ARMv6 fast path:
Before After
Mean StdDev Mean StdDev Confidence Change
L1 146.0 0.7 231.4 1.2 100.0% +58.5%
L2 143.1 0.9 222.1 1.7 100.0% +55.3%
On Sun, 05 Oct 2014 20:03:42 +0100, Søren Sandmann
wrote:
These can presumably also be used for the a8r8g8b8_sRGB format since
all we care about is the alpha channel.
Looks like it, yes. Updated versions of the three in__8 patches
coming up.
Also, is this actually showing up in the wild
This places a heavier emphasis on solid images than the other fuzz testers,
and tests both single-pixel repeating bitmap images as well as those created
using pixman_image_create_solid_fill(). In the former case, it also
exercises the case where the bitmap contents are written to after the
image's
Doesn't handle every pixel format, but for those that it does, enables early
conversion of OVER to SRC, for example.
---
pixman/pixman.c | 28 +---
1 files changed, 25 insertions(+), 3 deletions(-)
diff --git a/pixman/pixman.c b/pixman/pixman.c
index 9555cea..08293c8 100
On Thu, 07 Aug 2014 17:50:01 +0100, I wrote:
Doesn't handle every pixel format, but for those that it does, enables
early conversion of OVER to SRC, for example.
---
pixman/pixman-image.c |6 ++
1 files changed, 6 insertions(+), 0 deletions(-)
diff --git a/pixman/pixman-image.c b/pixman
On Tue, 21 Apr 2015 01:14:09 +0100
Ben Avison wrote:
> There are a group of combiner types - SRC, OVER_REVERSE, IN, OUT and ADD -
> where the source alpha affects only the destination alpha component. This
> means that any fast path with a8r8g8b8 source and destination can also be
> applied to an
12 matches
Mail list logo
