- 02 May, 2013 11 commits
-
-
Behdad Esfahbod authored
If we need to apply many many lookups, we can fasten that up by applying them in batches. For each batch we keep the union of the coverage of the lookups participating. We can then skip glyph ranges that do NOT participate in any lookup in the batch. The batch partition is determined optimally by a mathematical probability model on the glyphs and a dynamic-program to optimize the partition. The net effect is 30% speedup on Amiri. the downside is more memory consuption as each batch will keep an hb_set_t of its coverage. I'm not yet convinced that the tradeoff is worth pursuing. I'm trying to find out ways to optimized this more, with less memory overhead. This work also ignores the number of subtables per lookup. That may prove to be very important for the performance numbers from here on.
-
Behdad Esfahbod authored
No functional change.
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
- 30 Apr, 2013 1 commit
-
-
Behdad Esfahbod authored
Some infrastructures use DEBUG as a generic symbol.
-
- 29 Apr, 2013 2 commits
-
-
Behdad Esfahbod authored
This is not ideal as we don't like -L/usr/lib in our linker line. But this is only relevant to environments that don't have pkgconfig files for ICU... https://github.com/behdad/harfbuzz/pull/2
-
Behdad Esfahbod authored
Sigh.. Bug 64039 - undefined macro AM_PROG_AR
-
- 21 Apr, 2013 5 commits
-
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
The compile() function is starting to become illegible...
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
We always push a pause at the end such that each lookup falls in exactly one pause_map_t. Now, only if I can find a better name for that...
-
- 19 Apr, 2013 2 commits
-
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
Saves another 3 / 4 percent with Amiri.
-
- 18 Apr, 2013 3 commits
-
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
Not measurable by any means, but conceptually this is faster since the mask matches more often than the digest.
-
- 17 Apr, 2013 10 commits
-
-
Behdad Esfahbod authored
Wasn't correct with the new combiner. I should add it back somehow, but for now the digests seem to be working very well...
-
Behdad Esfahbod authored
Was unused.
-
Behdad Esfahbod authored
Make Amiri rendering faster a whopping 45% again! Speends up pretty much anything I tested.
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
Correctly this time.
-
Behdad Esfahbod authored
We were not initializing the digests properly and as a result they were being initialized to zero, making digest1 to never do any useful work. Speeds up Amiri shaping significantly.
-
Behdad Esfahbod authored
-
Behdad Esfahbod authored
This reverts commit 3d1b66a3. The calculations were buggy. It's not worth optimizing right now.
-
- 16 Apr, 2013 1 commit
-
-
Behdad Esfahbod authored
-
- 11 Apr, 2013 3 commits
-
-
Behdad Esfahbod authored
Useful for profiling shapers.
-
Behdad Esfahbod authored
Just to be sure it's not punishing the old shaper performance numbers. Doesn't seem to have a measurable effect.
-
Behdad Esfahbod authored
-
- 09 Apr, 2013 1 commit
-
-
Behdad Esfahbod authored
-
- 05 Apr, 2013 1 commit
-
-
Behdad Esfahbod authored
-