Xfermode_SrcOver:
SSE: 2.08ms -> 2.03ms (~2% faster)
NEON: my N5 is noisy, but there appears to be no perf change
BUG=skia:
Review URL: https://codereview.chromium.org/1132273004
This replacement shoots axis-aligned rays through all intersecting edges to find the outermost one either horizontally or vertically. The resulting code is smaller and twice as fast.
To support this, most of the horizontal / vertical intersection code was rewritten and standardized, and old code supporting the top-directed winding was deleted.
Contours were pointed to by an SkTDArray. Instead, put them in a linked list, and designate the list head with its own class to ensure that methods that take lists of contours start at the top. This change removed a large percentage of memory allocations used by path ops.
TBR=reed@google.com
BUG=skia:3588
Review URL: https://codereview.chromium.org/1111333002
This is a temporary disabling of advanced blend equations till we can fix
the various bugs associated with it.
BUG=skia:3822
Review URL: https://codereview.chromium.org/1130003003
The entire text content of the 'font' element is currently used as the
file name. This wasn't an issue in earlier versions, as the 'file'
element was dedicated to only containing the file name. The new 'font'
element also contains a number of attributes and potentially other tags.
This means that a 'font' element can become quite long, making it
desireable to be able to split it across multiple lines.
However, splitting the 'font' element across multiple lines is currently
difficult and awkward as any whitespace outside of tags will be
considered part of the file name. This change means that any leading or
trailing whitespace will not be considered part of the file name.
This only applies to v21 and later files, so while this restricts font
file names from beginning and ending with whitespace, it is unlikely to
break any users in practice. It is probably also undesireable to have
font files with names that begin or end with whitespace in any event.
Review URL: https://codereview.chromium.org/1125413003
Reason for revert:
speculative revert to fix failures in DEPS roll
Original issue's description:
> Sketch splitting SkPicture into an interface and SkBigPicture.
>
> Adds small pictures for drawRect(), drawTextBlob(), and drawPath().
> These cover about 89% of draw calls from Blink SKPs,
> and about 25% of draw calls from our GMs.
>
> SkPicture handles:
> - serialization and deserialization
> - unique IDs
>
> Everything else is left to the subclasses:
> - playback(), cullRect()
> - hasBitmap(), hasText(), suitableForGPU(), etc.
> - LayerInfo / AccelData if applicable.
>
> The time to record a 1-op picture improves a good chunk
> (2 mallocs to 1), and the time to record a 0-op picture
> greatly improves (2 mallocs to none):
>
> picture_overhead_draw: 450ns -> 350ns
> picture_overhead_nodraw: 300ns -> 90ns
>
> BUG=skia:
>
> Committed: https://skia.googlesource.com/skia/+/c92c129ff85b05a714bd1bf921c02d5e14651f8bTBR=reed@google.com,robertphillips@google.com,mtklein@google.com,mtklein@chromium.org
NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG=skia:
Review URL: https://codereview.chromium.org/1130333002
Adds small pictures for drawRect(), drawTextBlob(), and drawPath().
These cover about 89% of draw calls from Blink SKPs,
and about 25% of draw calls from our GMs.
SkPicture handles:
- serialization and deserialization
- unique IDs
Everything else is left to the subclasses:
- playback(), cullRect()
- hasBitmap(), hasText(), suitableForGPU(), etc.
- LayerInfo / AccelData if applicable.
The time to record a 1-op picture improves a good chunk
(2 mallocs to 1), and the time to record a 0-op picture
greatly improves (2 mallocs to none):
picture_overhead_draw: 450ns -> 350ns
picture_overhead_nodraw: 300ns -> 90ns
BUG=skia:
Review URL: https://codereview.chromium.org/1112523006
This CL refactors the location of the GrBufferAllocPools so they reside entirely inside the GrBatchTarget. This is in preparation for making them use scratch resources.
Review URL: https://codereview.chromium.org/1131553002
We don't seem to be making good use of the available instruction set.
SSE4.1 gives us an easy way to unpack a pixel into an __m128i, and
SSSE3 gave us an easy way to do the reverse.
This should be bit-perfect and about a 10% speedup.
BUG=skia:
Review URL: https://codereview.chromium.org/1123263003
This totally overhauls the implementation to use ordinary inheritance-based type erasure. I give up for now getting my manual vtable shenanigans to work with MSVC. Still those same "expected ; before ), also expected ) before ;" errors.
I added support for uninitialized SkFunctions and operator=(), because it was fairly straightforward with this implementation.
The main downside here is that I've removed the inline implementation. All SkFunctions involve a heap allocation, even when just wrapping function pointers.
BUG=skia:
Review URL: https://codereview.chromium.org/1056673002
