diff options
| author | Sami Liedes <sami.liedes@iki.fi> | 2013-07-29 13:40:08 +0300 |
|---|---|---|
| committer | The Qt Project <gerrit-noreply@qt-project.org> | 2014-01-18 18:30:43 +0100 |
| commit | 0e65cec6b4f5ca21f3964a18b4a3f35a32a6762f (patch) | |
| tree | e8d3901b4ca4ebef77c4db24a095eafcb39e98e5 /src/gui | |
| parent | 18c04d0ab6debceea406a77676913f826a093153 (diff) | |
Optimize blend_transformed_tiled_argb().
Profiling indicates that this function is one of two hot spots causing
a noticeable latency when changing KDE virtual desktops.
Instead of computing two modulos per pixel in the inner loop, it is
possible to compute the modulos outside the loop and compute a modulo
sum in the inner loop for a reasonable speedup.
Change-Id: Ic4217b7686e031d7673b3e10aa977dae263096dc
Reviewed-by: Gunnar Sletta <gunnar.sletta@jollamobile.com>
Diffstat (limited to 'src/gui')
| -rw-r--r-- | src/gui/painting/qdrawhelper.cpp | 20 |
1 files changed, 14 insertions, 6 deletions
diff --git a/src/gui/painting/qdrawhelper.cpp b/src/gui/painting/qdrawhelper.cpp index 75e5f1eae3..7750d22a91 100644 --- a/src/gui/painting/qdrawhelper.cpp +++ b/src/gui/painting/qdrawhelper.cpp @@ -5116,13 +5116,13 @@ static void blend_transformed_tiled_argb(int count, const QSpan *spans, void *us int l = qMin(length, buffer_size); const uint *end = buffer + l; uint *b = buffer; + int px16 = x % (image_width << 16); + int py16 = y % (image_height << 16); + int px_delta = fdx % (image_width << 16); + int py_delta = fdy % (image_height << 16); while (b < end) { - int px = x >> 16; - int py = y >> 16; - px %= image_width; - py %= image_height; - if (px < 0) px += image_width; - if (py < 0) py += image_height; + int px = px16 >> 16; + int py = py16 >> 16; int y_offset = py * scanline_offset; Q_ASSERT(px >= 0 && px < image_width); @@ -5131,6 +5131,14 @@ static void blend_transformed_tiled_argb(int count, const QSpan *spans, void *us *b = image_bits[y_offset + px]; x += fdx; y += fdy; + px16 += px_delta; + if (px16 >= image_width << 16) + px16 -= image_width << 16; + py16 += py_delta; + if (py16 >= image_height << 16) + py16 -= image_height << 16; + if (px16 < 0) px16 += image_width << 16; + if (py16 < 0) py16 += image_height << 16; ++b; } func(target, buffer, l, coverage); |