diff options
author | Sergey Sharybin <sergey.vfx@gmail.com> | 2014-06-13 22:30:13 +0400 |
---|---|---|
committer | Sergey Sharybin <sergey.vfx@gmail.com> | 2014-06-13 22:38:07 +0400 |
commit | a87fb34edaf1a10f5527b6dc8a506a1c9ecbc683 (patch) | |
tree | 06386145cbf7f9dcf6684b3a39722ed4d4e62c4d /source/blender/compositor/operations/COM_BlurBaseOperation.h | |
parent | b0708dd7189dfef21f7f9af5e98b0a7e1369e507 (diff) |
Use advantage of SSE2 instructions in gaussian blur node
This gives around 30% of speedup for gaussian blur node.
Pretty much straightforward implementation inside the node
itself, but needed to implement some additional things:
- Aligned malloc. It's needed to load data onto SSE registers
faster. based on the aligned_malloc() from Libmv with
some additional trickery going on to support arbitrary
alignment (this magic is needed because of MemHead).
In the practice only 16bit alignment is supported because
of the lack of aligned malloc with arbitrary alignment
for OSX. Not a bit deal for now because we need 16 bytes
alignment at this moment only. Could be tweaked further
later.
- Memory buffers in compositor are now aligned to 16 bytes.
Should be harmless for non-SSE cases too. just mentioning.
Reviewers: campbellbarton, lukastoenne, jbakker
Reviewed By: campbellbarton
CC: lockal
Differential Revision: https://developer.blender.org/D564
Diffstat (limited to 'source/blender/compositor/operations/COM_BlurBaseOperation.h')
-rw-r--r-- | source/blender/compositor/operations/COM_BlurBaseOperation.h | 7 |
1 files changed, 7 insertions, 0 deletions
diff --git a/source/blender/compositor/operations/COM_BlurBaseOperation.h b/source/blender/compositor/operations/COM_BlurBaseOperation.h index 052a525ef2c..e97dd4d766d 100644 --- a/source/blender/compositor/operations/COM_BlurBaseOperation.h +++ b/source/blender/compositor/operations/COM_BlurBaseOperation.h @@ -27,6 +27,10 @@ #define MAX_GAUSSTAB_RADIUS 30000 +#ifdef __SSE2__ +# include <emmintrin.h> +#endif + class BlurBaseOperation : public NodeOperation, public QualityStepHelper { private: @@ -34,6 +38,9 @@ protected: BlurBaseOperation(DataType data_type); float *make_gausstab(float rad, int size); +#ifdef __SSE2__ + __m128 *convert_gausstab_sse(const float *gaustab, float rad, int size); +#endif float *make_dist_fac_inverse(float rad, int size, int falloff); void updateSize(); |