|
|
|
@ -127,7 +127,7 @@ __kernel void arithm_add_D2 (__global ushort *src1, int src1_step, int src1_offs |
|
|
|
|
#ifdef dst_align |
|
|
|
|
#undef dst_align |
|
|
|
|
#endif |
|
|
|
|
#define dst_align ((dst_offset >> 1) & 3) |
|
|
|
|
#define dst_align ((dst_offset / 2) & 3) |
|
|
|
|
int src1_index = mad24(y, src1_step, (x << 1) + src1_offset - (dst_align << 1)); |
|
|
|
|
int src2_index = mad24(y, src2_step, (x << 1) + src2_offset - (dst_align << 1)); |
|
|
|
|
|
|
|
|
@ -165,7 +165,7 @@ __kernel void arithm_add_D3 (__global short *src1, int src1_step, int src1_offse |
|
|
|
|
#ifdef dst_align |
|
|
|
|
#undef dst_align |
|
|
|
|
#endif |
|
|
|
|
#define dst_align ((dst_offset >> 1) & 3) |
|
|
|
|
#define dst_align ((dst_offset / 2) & 3) |
|
|
|
|
int src1_index = mad24(y, src1_step, (x << 1) + src1_offset - (dst_align << 1)); |
|
|
|
|
int src2_index = mad24(y, src2_step, (x << 1) + src2_offset - (dst_align << 1)); |
|
|
|
|
|
|
|
|
@ -335,7 +335,7 @@ __kernel void arithm_add_with_mask_C1_D2 (__global ushort *src1, int src1_step, |
|
|
|
|
#ifdef dst_align |
|
|
|
|
#undef dst_align |
|
|
|
|
#endif |
|
|
|
|
#define dst_align ((dst_offset >> 1) & 1) |
|
|
|
|
#define dst_align ((dst_offset / 2) & 1) |
|
|
|
|
int src1_index = mad24(y, src1_step, (x << 1) + src1_offset - (dst_align << 1)); |
|
|
|
|
int src2_index = mad24(y, src2_step, (x << 1) + src2_offset - (dst_align << 1)); |
|
|
|
|
int mask_index = mad24(y, mask_step, x + mask_offset - dst_align); |
|
|
|
@ -375,7 +375,7 @@ __kernel void arithm_add_with_mask_C1_D3 (__global short *src1, int src1_step, i |
|
|
|
|
#ifdef dst_align |
|
|
|
|
#undef dst_align |
|
|
|
|
#endif |
|
|
|
|
#define dst_align ((dst_offset >> 1) & 1) |
|
|
|
|
#define dst_align ((dst_offset / 2) & 1) |
|
|
|
|
int src1_index = mad24(y, src1_step, (x << 1) + src1_offset - (dst_align << 1)); |
|
|
|
|
int src2_index = mad24(y, src2_step, (x << 1) + src2_offset - (dst_align << 1)); |
|
|
|
|
int mask_index = mad24(y, mask_step, x + mask_offset - dst_align); |
|
|
|
@ -507,7 +507,7 @@ __kernel void arithm_add_with_mask_C2_D0 (__global uchar *src1, int src1_step, i |
|
|
|
|
#ifdef dst_align |
|
|
|
|
#undef dst_align |
|
|
|
|
#endif |
|
|
|
|
#define dst_align ((dst_offset >> 1) & 1) |
|
|
|
|
#define dst_align ((dst_offset / 2) & 1) |
|
|
|
|
int src1_index = mad24(y, src1_step, (x << 1) + src1_offset - (dst_align << 1)); |
|
|
|
|
int src2_index = mad24(y, src2_step, (x << 1) + src2_offset - (dst_align << 1)); |
|
|
|
|
int mask_index = mad24(y, mask_step, x + mask_offset - dst_align); |
|
|
|
|