Commit Graph

2916 Commits

Author SHA1 Message Date
David Fort
c0f4b6bcad Merge pull request #3772 from akallabeth/sse_test_fixes
Sse test fixes
2017-02-16 16:37:27 +01:00
David Fort
be41769506 Add some useful debug info for the egfx channel 2017-02-16 14:49:08 +01:00
Armin Novak
c50e5ba57e Fixed warnings. 2017-02-16 14:34:43 +01:00
Armin Novak
c04918d6c2 Added value comparison for all YCbCr functions. 2017-02-16 14:00:25 +01:00
Armin Novak
4891ae62b1 Reduced test runtime. 2017-02-16 13:52:27 +01:00
Armin Novak
198bc6d9e1 Fixed compiler warnings. 2017-02-16 13:17:49 +01:00
Armin Novak
4d03a9aa6f Improved YCoCgToRGB tests. 2017-02-16 12:16:58 +01:00
Armin Novak
799afe0cca Fixed YUV test result print. 2017-02-16 12:16:58 +01:00
Armin Novak
dda6b14402 Fixed SSE yCbCrToRGB_16s8u_P3AC4R, improved tests. 2017-02-16 12:16:58 +01:00
Armin Novak
ee3a3a787b Fixed SSE checks and RGBToRGB_16s8u_P3AC4R 2017-02-16 09:57:13 +01:00
Martin Fleisz
0d43201e3c Merge pull request #3767 from akallabeth/ssse3_optimize
Ssse3 optimize and sanitize address
2017-02-16 09:51:28 +01:00
Armin Novak
6960ca209b Added GFX surface command profiler. 2017-02-15 15:42:14 +01:00
Martin Fleisz
68a9408249 core: Get rid of useless settings copy 2017-02-15 14:59:24 +01:00
Armin Novak
1b78c45eca Fixed alignment checks for SSE yCbCrToRGB 2017-02-15 12:39:57 +01:00
Armin Novak
01d24264d2 Using aligned RFX buffers. 2017-02-15 12:37:37 +01:00
Armin Novak
dce3604bcf Refined alignment checks for SSSE3 YUV444ToRGB 2017-02-15 12:08:32 +01:00
Armin Novak
f24b112514 Refactored AVC444 decoding to single YUV420 buffer
Since not all H264 decoders support multiple YUV420 output
buffers process H264 decoding and YUV to RGB conversion
sequentially to avoid overriding the input data.
2017-02-15 09:28:10 +01:00
Armin Novak
2c99c17078 Fixed libavcodec issues. 2017-02-14 16:41:07 +01:00
Armin Novak
e44158ee29 Reordered SSSE3 YUV444ToRGB 2017-02-14 16:02:12 +01:00
Armin Novak
d41d65575c Use bitmap color format in gdi_BitBlt 2017-02-14 16:02:12 +01:00
Armin Novak
2463575083 Print profiler header and footer. 2017-02-14 16:02:11 +01:00
Armin Novak
ef4515a674 Added FPS in profiler print. 2017-02-14 16:02:11 +01:00
Armin Novak
f324b0ef9c Use aligned malloc/free for GFX surfaces/buffers 2017-02-14 16:02:11 +01:00
Armin Novak
9dba98598f SSSE3 YUV444ToRGB process 16 pixel per loop. 2017-02-14 16:02:11 +01:00
akallabeth
0540e189b1 Merge pull request #3765 from mfleisz/gdi_resize
gdi: Allow changing buffer even if size remains the same
2017-02-14 15:05:36 +01:00
Armin Novak
94b9fef049 Updated LIBAVCODEC support for H264 decoder. 2017-02-14 10:34:27 +01:00
Martin Fleisz
5930d7e81b gdi: Allow changing buffer even if size remains the same 2017-02-14 09:52:44 +01:00
Martin Fleisz
e97c4b57a4 Merge pull request #3754 from akallabeth/asm3
NEON and SSSE3 decoder optimisations
2017-02-14 09:23:21 +01:00
Armin Novak
e3ed91ea7a Optimized clear codec line copy. 2017-02-13 17:29:29 +01:00
Armin Novak
3644d8dfbd Fixed YUV444ToRGB, do not discard chroma frame. 2017-02-13 16:11:11 +01:00
Armin Novak
4f705b99ec Fixed primitives sign test buffer size. 2017-02-13 16:09:57 +01:00
Armin Novak
e60cc9bdd9 NEON optimized YCoCg conversion. 2017-02-13 16:09:28 +01:00
Armin Novak
a791ab0b17 Added NEON and SSSE3 YCbCr conversion functions. 2017-02-13 16:08:37 +01:00
Armin Novak
b1e3bab8ef Added NEON and SSSE3 YUV conversion optimisations. 2017-02-13 16:00:12 +01:00
Armin Novak
85d8157b41 Limit RFX update region to screen. 2017-02-13 15:26:21 +01:00
Armin Novak
57db52237c Fixed AVC444 luma/chroma checks. 2017-02-10 08:31:32 +01:00
Armin Novak
ef0a7cf5d1 Fixed missing chroma data for AVC444 2017-02-09 19:55:40 +01:00
Armin Novak
7ef90497c7 Respect width and height in rfx_process_message 2017-02-09 12:36:54 +01:00
David Fort
4e0003533e Parses the SupportStatusInfoPdu early capability and send it to clients if supported 2017-02-09 11:50:46 +01:00
David Fort
6d8969ece6 Merge pull request #3747 from nfedera/improve-generic-rgb-to-yuv420
primitives: RGB to YUV420 (non SIMD) optimizations
2017-02-07 11:19:25 +01:00
Norbert Federa
490473d26f primitives: RGB to YUV420 (non SIMD) optimizations
- added a dedicated non-simd encoder for BGR[AX] source formats
- simplified the generic RgbToYuv420 encoder and fixed the crash
  with odd dimensions
2017-02-06 15:48:52 +01:00
Norbert Federa
246801ec11 Merge pull request #3740 from akallabeth/gdi_color_conversion
Fix for GDI color decoding issues.
2017-02-06 15:13:56 +01:00
Armin Novak
8845b3ffa2 Added warning for unsupported color depth. 2017-02-06 10:31:43 +01:00
Norbert Federa
cb5dfd831f primitives: added SSSE3 rgb to yuv420 encoder
Currently supported source pixel formats are:
- PIXEL_FORMAT_BGRA32
- PIXEL_FORMAT_BGRX32

Support for PIXEL_FORMAT_RGB[XA]32 can simply be added if
required (see the comment in prim_YUV_opt.c).

On my old 3.1 GHz Core i5-2400 the new SSSE3 function can convert
over 900 1080p BGRX frames per second.

The current non-optimized C version (which supports all pixel formats)
can't do more than 40 yuv conversions per second on this cpu.

---------------------------+---------+-------------+-----------+-------
RGB TO YUV420 1080p 32bit  |   COUNT |       TOTAL |       AVG |    FPS
---------------------------+---------+-------------+-----------+-------
general_RGBToYUV420        |     500 |    13.1776s | 0.026355s |     38
ssse3_RGBToYUV420          |     500 |     0.5320s | 0.001064s |    940

Also fixed an error in TestPrimitivesYUV which generated resolutions
with height or width set to zero
2017-02-03 13:56:50 +01:00
Norbert Federa
6001cb710d Merge pull request #3717 from akallabeth/prim_fixes
Fixed primitives.
2017-02-01 11:11:27 +01:00
Armin Novak
df764f5aad Fixed GDI color decoding issues. 2017-02-01 11:02:23 +01:00
Armin Novak
16de1bca2b Removed unused profiler. 2017-01-31 14:18:28 +01:00
Martin Fleisz
c2eacd3b31 Build: Use correct pdb names when installing with symbols 2017-01-31 13:32:51 +01:00
Armin Novak
dbfbd5b086 Fixed argument order of yCbCrToRGB_16s8u_P3AC4R 2017-01-31 10:49:59 +01:00
Armin Novak
c4122d342c Added performance profiler to YUV test. 2017-01-31 10:31:34 +01:00