- 91f489a Modify the blur bench to add tests for high-quality blurs, and large non-integer blurs. Change the Coarse flag (which just turns on separable blurs) to an #ifdef, since separable is no longer just Coarse. (This #ifdef will hopefully be shortlived, once Chrome has switched). by senorblanco@chromium.org · 12 years ago
- 9b0d4d7 Implement radius interpolation for separable blur. Unroll both separable implementations, which yields up to 2X perf improvement. by senorblanco@chromium.org · 12 years ago
- 336b4da Separable mask blurs: Add compile-time flag. Fix reversed offsets in asymmetrical blurs (this bug cancels itself out, but I thought it might be confusing for future readers). Use correct stride in asymmetrical blurs (this is a real bug). by senorblanco@chromium.org · 12 years ago
- c438130 Separable blur: subpixel blurring. We approximate intermediate values of blur for even kernel sizes by using a kernel of size N offset to the left in the first pass, a kernel of size N offset to the right in the second pass, and a centered kernel of size N + 1 in the third pass. This required adding support for asymmetrical radii to the box blur. by senorblanco@chromium.org · 12 years ago
- 884e60b Sanitizing source files in Skia_Periodic_House_Keeping by skia.committer@gmail.com · 12 years ago
- 908276b Reorder the separable blur passes into XXX/YYY, with an image transpose on the last pass of each group. This results in continuguous memory reads in all passes, giving a 22% speedup on theverge.skp over the previous separable implementation, and a 30%-50% improvement over the existing implementation (depending on platform). by senorblanco@chromium.org · 12 years ago
- 72fac35 Build fix. Unreviewed. by senorblanco@chromium.org · 12 years ago
- 71f0f34 A new implementation of mask blurs, using separable X/Y passes. Disabled for now. by senorblanco@chromium.org · 12 years ago
- fbfcd56 Result of running tools/sanitize_source_files.py (which was added in https://codereview.appspot.com/6465078/) by rmistry@google.com · 12 years ago
- 054ff1e Fix the string we're checking for in the #if defined(), so that loops actually do get unrolled on Windows (15% perf win claimed many months ago). by tomhudson@google.com · 13 years ago
- 0c00f21 Add SkDEBUGFAIL to clean up use of SkASSERT(!"text"); by tomhudson@google.com · 13 years ago
- 01224d5 Unroll loops in SkBlurMask for speedup on Windows (benchmarks should see by tomhudson@google.com · 13 years ago
- 8caac64 Add fractional-radius blurs to blur benchmark. by tomhudson@google.com · 13 years ago
- 889bd8b Make SkMath.h less visible in public header files. by tomhudson@google.com · 13 years ago
- 5af16f8 Reduce size of filter mask. http://codereview.appspot.com/4965057/ by bungeman@google.com · 13 years ago
- 03016a3 style cleanup by reed@google.com · 13 years ago
- ec3ed6a Automatic update of all copyright notices to reflect new license terms. by epoger@google.com · 13 years ago
- 4868e6b [Committed on behalf of junov@chromium.org] by senorblanco@chromium.org · 14 years ago
- 4560767 fix sumbuffer to initialize its top and left colum/row with zeros. Without this fix we were by reed@android.com · 15 years ago
- 0e3c664 use rowbytes instead of bounds.width() by reed@android.com · 15 years ago
- 543ed93 add ADD xfermode safely handle (don't draw) if the size of SkMask exceeds 31bits by reed@android.com · 15 years ago
- 49f0ff2 promote SkMask's rowBytes to 32bits, since its bounds is a 32bit rect by reed@android.com · 15 years ago
- 8a1c16f grab from latest android by reed@android.com · 16 years ago