At 2013-11-19 04:42:21,"Steve Borho" <[email protected]> wrote:
>
>On Nov 18, 2013, at 9:23 AM, [email protected] wrote:
>
>> # HG changeset patch
>> # User Praveen Tiwari
>> # Date 1384788209 -19800
>> # Node ID 59646d515e79b4d0f9a3a72c77c7af17a83bf3d9
>> # Parent b353d170c54f0e33a8869c413be226a48deb1f5c
>> added csp support for blpckcopy_ps
>>
>> diff -r b353d170c54f -r 59646d515e79 source/common/pixel.cpp
>> --- a/source/common/pixel.cpp Mon Nov 18 19:34:07 2013 +0530
>> +++ b/source/common/pixel.cpp Mon Nov 18 20:53:29 2013 +0530
>> @@ -837,7 +837,7 @@
>> #define CHROMA(W, H) \
>> p.chroma_copy_pp[CSP_I420][CHROMA_ ## W ## x ## H] = blockcopy_pp_c<W, H>; \
>> p.chroma_copy_sp[CHROMA_ ## W ## x ## H] = blockcopy_sp_c<W, H>; \
>> - p.chroma_copy_ps[CHROMA_ ## W ## x ## H] = blockcopy_ps_c<W, H>;\
>> + p.chroma_copy_ps[CSP_I420][CHROMA_ ## W ## x ## H] = blockcopy_ps_c<W,
>> H>;\
>> p.chroma_sub_ps[CHROMA_ ## W ## x ## H] = pixel_sub_ps_c<W, H>;
>>
>> #define LUMA(W, H) \
>> diff -r b353d170c54f -r 59646d515e79 source/common/primitives.h
>> --- a/source/common/primitives.h Mon Nov 18 19:34:07 2013 +0530
>> +++ b/source/common/primitives.h Mon Nov 18 20:53:29 2013 +0530
>> @@ -247,7 +247,7 @@
>> copy_sp_t luma_copy_sp[NUM_LUMA_PARTITIONS];
>> copy_sp_t chroma_copy_sp[NUM_CHROMA_PARTITIONS];
>> copy_ps_t luma_copy_ps[NUM_LUMA_PARTITIONS];
>> - copy_ps_t chroma_copy_ps[NUM_CHROMA_PARTITIONS];
>> + copy_ps_t chroma_copy_ps[NUM_CSP][NUM_CHROMA_PARTITIONS];
>>
>> pixel_sub_ps_t luma_sub_ps[NUM_LUMA_PARTITIONS];
>> pixel_sub_ps_t chroma_sub_ps[NUM_CHROMA_PARTITIONS];
>> diff -r b353d170c54f -r 59646d515e79 source/common/x86/asm-primitives.cpp
>> --- a/source/common/x86/asm-primitives.cpp Mon Nov 18 19:34:07 2013 +0530
>> +++ b/source/common/x86/asm-primitives.cpp Mon Nov 18 20:53:29 2013 +0530
>> @@ -141,7 +141,6 @@
>> p.chroma_hps[CHROMA_ ## W ## x ## H] = x265_interp_4tap_horiz_ps_ ## W ## x
>> ## H ## cpu; \
>> p.chroma_vpp[CHROMA_ ## W ## x ## H] = x265_interp_4tap_vert_pp_ ## W ## x
>> ## H ## cpu; \
>> p.chroma_vps[CHROMA_ ## W ## x ## H] = x265_interp_4tap_vert_ps_ ## W ## x
>> ## H ## cpu; \
>> - p.chroma_copy_ps[CHROMA_ ## W ## x ## H] = x265_blockcopy_ps_ ## W ## x
>> ## H ## cpu; \
>> p.chroma_sub_ps[CHROMA_ ## W ## x ## H] = x265_pixel_sub_ps_ ## W ## x ## H
>> ## cpu;
>>
>> #define SETUP_CHROMA_SP_FUNC_DEF(W, H, cpu) \
>> @@ -380,6 +379,36 @@
>> SETUP_LUMA_BLOCKCOPY_FUNC_DEF(64, 16, cpu); \
>> SETUP_LUMA_BLOCKCOPY_FUNC_DEF(16, 64, cpu);
>>
>> +#define SETUP_CHROMA_FROM_LUMA_SSE4(W1, H1, W2, H2, cpu) \
>> + p.chroma_copy_ps[X265_CSP_I420][LUMA_ ## W1 ## x ## H1] =
>> x265_blockcopy_ps_ ## W2 ## x ## H2 ## cpu;
>> +
>> +// For X265_CSP_I420 chroma width and height will be half of luma width and
>> height
>> +#define CHROMA_BLOCKCOPY_SSE4(cpu) \
>
>When the macro accepts a cpu type argument, adding SSE4 to the name is
>redundant (and confusing)
>there should probably be a generic I420 macro that maps luma blocks to I420
>blocks so adding more color spaces does not multiply amount of code in this
>file
>
Steve give us a good idea, we may modify macro to SETUP_CHROMA_FROM_LUMA(W1,
H1, W2, H2, cpu, csp)
_______________________________________________
x265-devel mailing list
[email protected]
https://mailman.videolan.org/listinfo/x265-devel
