try to speed-up maprect + round2i + contains

gyp/core.gypi

Index: gyp/core.gypi
diff --git a/gyp/core.gypi b/gyp/core.gypi
index 97e38c2cc6b5149a2b072dfac9f7bcb6b6190c53..06738f172eec8431e1521aed1a9aac94ea0ba8d0 100644
--- a/gyp/core.gypi
+++ b/gyp/core.gypi
@@ -86,7 +86,6 @@
         '<(skia_src_path)/core/SkConvolver.h',
         '<(skia_src_path)/core/SkCoreBlitters.h',
         '<(skia_src_path)/core/SkCpu.cpp',
-        '<(skia_src_path)/core/SkCpu.h',
         '<(skia_src_path)/core/SkCubicClipper.cpp',
         '<(skia_src_path)/core/SkCubicClipper.h',
         '<(skia_src_path)/core/SkData.cpp',
@@ -192,7 +191,6 @@
         '<(skia_src_path)/core/SkNinePatchIter.h',
         '<(skia_src_path)/core/SkNormalSource.cpp',
         '<(skia_src_path)/core/SkNormalSource.h',
-        '<(skia_src_path)/core/SkNx.h',
         '<(skia_src_path)/core/SkOpts.cpp',
         '<(skia_src_path)/core/SkOpts.h',
         '<(skia_src_path)/core/SkOrderedReadBuffer.h',
@@ -420,12 +418,14 @@
         # private
         '<(skia_include_path)/private/SkAtomics.h',
         '<(skia_include_path)/private/SkChecksum.h',
+        '<(skia_include_path)/private/SkCpu.h',
         '<(skia_include_path)/private/SkFixed.h',
         '<(skia_include_path)/private/SkFloatBits.h',
         '<(skia_include_path)/private/SkFloatingPoint.h',
         '<(skia_include_path)/private/SkGpuFenceSync.h',
         '<(skia_include_path)/private/SkMiniRecorder.h',
         '<(skia_include_path)/private/SkMutex.h',
+        '<(skia_include_path)/private/SkNx.h',
         '<(skia_include_path)/private/SkOnce.h',
         '<(skia_include_path)/private/SkRecords.h',
         '<(skia_include_path)/private/SkSemaphore.h',