i3lock-color/blur.h

#ifndef _BLUR_H
#define _BLUR_H

#include <stdint.h>
#include <cairo.h>

void blur_image_surface (cairo_surface_t *surface, int radius);
void blur_impl_naive(uint32_t* src, uint32_t* dst, int width, int height, int src_stride, int dst_stride, int radius);

void blur_impl_sse2(uint32_t* src, uint32_t* dst, int width, int height, float sigma);
void blur_impl_horizontal_pass_sse2(uint32_t *src, uint32_t *dst, float *kernel, int width, int height)
    __attribute__ ((__target__ ("no-avx")));

void blur_impl_avx(uint32_t* src, uint32_t* dst, int width, int height, float sigma);
void blur_impl_horizontal_pass_avx(uint32_t *src, uint32_t *dst, float *kernel, int width, int height);

void blur_impl_ssse3(uint32_t* src, uint32_t* dst, int width, int height, float sigma);
void blur_impl_horizontal_pass_ssse3(uint32_t *src, uint32_t *dst, int16_t *kernel, int width, int height);

#endif
add initial blur support thanks to https://github.com/shiver/i3lock 8 years ago			`#ifndef _BLUR_H`
			`#define _BLUR_H`

Isolate blur implementation to a function. This will allow easier switching between naive/SSE2/AVX implementations. 8 years ago			`#include <stdint.h>`
			`#include <cairo.h>`

add initial blur support thanks to https://github.com/shiver/i3lock 8 years ago			`void blur_image_surface (cairo_surface_t *surface, int radius);`
Isolate blur implementation to a function. This will allow easier switching between naive/SSE2/AVX implementations. 8 years ago			`void blur_impl_naive(uint32_t* src, uint32_t* dst, int width, int height, int src_stride, int dst_stride, int radius);`
Add AVX version. It relies on some SSE2 instructions, so performance gain is not that huge (about 1.4x). I experimented with 256-bit loads, but they turned out to be slower (at least on Sandy Bridge). 8 years ago
Add SSE2-optimized blur. About 4-6 times faster than naive implementation. 8 years ago			`void blur_impl_sse2(uint32_t* src, uint32_t* dst, int width, int height, float sigma);`
Add AVX version. It relies on some SSE2 instructions, so performance gain is not that huge (about 1.4x). I experimented with 256-bit loads, but they turned out to be slower (at least on Sandy Bridge). 8 years ago			`void blur_impl_horizontal_pass_sse2(uint32_t src, uint32_t dst, float *kernel, int width, int height)`
			`__attribute__ ((__target__ ("no-avx")));`

			`void blur_impl_avx(uint32_t* src, uint32_t* dst, int width, int height, float sigma);`
			`void blur_impl_horizontal_pass_avx(uint32_t src, uint32_t dst, float *kernel, int width, int height);`

Add SSSE3-based blur implementation. Calculations are done on integer, rather than floating point numbers, so this implementation is not as accurate (but when scale factor is reasonable enough, no artifacs are visible). It is, however, faster by a factor of ~3. 8 years ago			`void blur_impl_ssse3(uint32_t* src, uint32_t* dst, int width, int height, float sigma);`
SSSE3: use 16-bit weights. Overall, I'm very happy with performance of this code, but not so much with resulting image. It seems like integer approximations won't do. I might remove this code altogether, so I didn't update comments. 8 years ago			`void blur_impl_horizontal_pass_ssse3(uint32_t src, uint32_t dst, int16_t *kernel, int width, int height);`
add initial blur support thanks to https://github.com/shiver/i3lock 8 years ago
			`#endif`