diff options
| author | Alexei Podtelezhnikov <apodtele@gmail.com> | 2021-08-30 07:09:53 -0400 |
|---|---|---|
| committer | Alexei Podtelezhnikov <apodtele@gmail.com> | 2021-08-30 07:09:53 -0400 |
| commit | b6c11d494e9b5ceaab75a5858b836e8a8ede41c0 (patch) | |
| tree | ee640c857ecf7efaaea7d80378a4acd1fbce7868 | |
| parent | d1c20005042a31fba7e86603b0f65fcbf21fe51c (diff) | |
| download | freetype2-b6c11d494e9b5ceaab75a5858b836e8a8ede41c0.tar.gz | |
[smooth] Reduce shift in multiply-shift optimization.
* src/smooth/ftgrays.c (FT_UDIVPREP, FT_UDIV): Reduce shift.
Smaller shifts that keep the division operands of FT_UDIVPREP within
32 bits result in slightly faster divisions, which is noticeable in
the overall performance. The loss of precision is tolerable until the
divisors (the components dx and dy) approach 32 - PIXEL_BITS. With
PIXEL_BITS = 8, this corresponds to 65,000 pixels or the bitmap size
that we refuse to render anyway.
Using `ftbench -p -s60 -t5 -bc timesi.ttf`,
Before: 8.52 us/op
After: 8.32 us/op
| -rw-r--r-- | src/smooth/ftgrays.c | 10 |
1 files changed, 4 insertions, 6 deletions
diff --git a/src/smooth/ftgrays.c b/src/smooth/ftgrays.c index 576dbb325..b8d680160 100644 --- a/src/smooth/ftgrays.c +++ b/src/smooth/ftgrays.c @@ -386,12 +386,10 @@ typedef ptrdiff_t FT_PtrDist; /* divisors to provide sufficient accuracy of the multiply-shift. */ /* It should not exceed (64 - PIXEL_BITS) to prevent overflowing and */ /* leave enough room for 64-bit unsigned multiplication however. */ -#define FT_UDIVPREP( c, b ) \ - FT_Int64 b ## _r = c ? (FT_Int64)( ~(FT_UInt64)0 >> PIXEL_BITS ) / ( b ) \ - : 0 -#define FT_UDIV( a, b ) \ - (TCoord)( ( (FT_UInt64)( a ) * (FT_UInt64)( b ## _r ) ) >> \ - ( sizeof( FT_UInt64 ) * FT_CHAR_BIT - PIXEL_BITS ) ) +#define FT_UDIVPREP( c, b ) \ + FT_Int64 b ## _r = c ? (FT_Int64)0xFFFFFFFF / ( b ) : 0 +#define FT_UDIV( a, b ) \ + (TCoord)( ( (FT_UInt64)( a ) * (FT_UInt64)( b ## _r ) ) >> 32 ) /* Scale area and apply fill rule to calculate the coverage byte. */ |