Neon move prfm after loads for all functions. Example performance improvement
Was
I444ToARGB_Opt (3275 ms)
I444ToNV12_Opt (1509 ms)
Now
I444ToARGB_Opt (2751 ms)
I444ToNV12_Opt (1367 ms)
Bug: libyuv:447
Change-Id: I78bf797b3600084c1eceb0be44cdbc9a575de803
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/2189559
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: richard winterton <rrwinterton@gmail.com>
This patch is a complement for commit bed9292f2cbba2f8f9ff0f1635a8aa17a311f2f9.
1. Supplement inspection for macro HAS_***TOUV*ROW_MMI/MSA.
2. Reduce calls to function TestCpuFlag().
3. Fix a mistake in source/convert.cc: line 1105.
Change-Id: I5e7f9fe367fa0f6d1db6f7644c5b48d4ad85fedb
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/2169342
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Some processors support both MSA and MMI.
when they are enabled together, MSA will be preferd.
This patch move MSA initialization after MMI, so that
MSA can overide MMI and be setted to effective.
Change-Id: I8a52cce83ee4ec9727d47c99b287c9580329b149
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/2155944
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
1. Switch to 8 bit precision.
2. Fix an error in the implementation of MMI and MSA.
About the error:
MMI and MSA implementation for RGBtoY and RGBToYJ used different
precision according to the C implementation( The C version has been
unified in commit fce0fed542001577e6b10f4cf859e0fa1774974e). This
patch unifies the precision to 8 bit for RGBToYJ in MMI and MSA.
Change-Id: Ic6a6e424d27a2f049b0c954f03174192d2beb091
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/2155608
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
This unittest help to test MipsCpuCaps.
Change-Id: I9e0ceeed0e5243446eaafa27e8de4c5f8163b09e
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/2133314
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
1. Refactored function MipsCpuCaps.
2. allow msa and mmi can be enabled together.
Signed-off-by: Jiaxun Yang <jiaxun.yang@flygoat.com>
Signed-off-by: Shiyou Yin <yinshiyou-hf@loongson.cn>
Change-Id: I7330d0551a6a167e4c76d37e4defcc20783f5815
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/2131145
Reviewed-by: Hsiu Wang <hsiu@google.com>
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Call a row function for each row, based on ARGBToI400 code.
But implement row functions as 2 step conversion. Adds the
row functions:
RAWToYJ, RGBToYJ, SSSE3 and AVX2 versions, and Any versions.
The smaller row buffer is more cache friendly on large images.
The max cache size can be configured, and is currently:
// Maximum temporary width for wrappers to process at a time, in pixels.
And the row buffer is
SIMD_ALIGNED(uint8_t row[MAXTWIDTH * 4]);
So 8192 bytes are used for the row buffer, leaving the rest for source
and destination buffers.
blaze-bin/third_party/libyuv/libyuv_test '--gunit_filter=*R*To?400_Opt' --libyuv_width=3600 --libyuv_height=2500 --libyuv_repeat=1000 --libyuv_flags=-1 --libyuv_cpu_info=-1 | sortms
Was
RAWToJ400_Opt (3996 ms)
ARGBToI400_Opt (3964 ms)
RGB24ToJ400_Opt (3960 ms)
ARGBToJ400_Opt (3909 ms)
RGBAToJ400_Opt (3885 ms)
Now
ARGBToJ400_Opt (4091 ms)
ARGBToI400_Opt (3936 ms)
RGBAToJ400_Opt (3428 ms)
RGB24ToJ400_Opt (3324 ms)
RAWToJ400_Opt (3309 ms)
Bug: libyuv:854, b/147753855
Change-Id: Ieb65fbda94e812c737f4c3c74107354b73c4bcd2
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/2016203
Reviewed-by: richard winterton <rrwinterton@gmail.com>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
This adds some missing prototypes from the BT.2020 CL as well as expands
the H444 and J444 results.
BUG=960620, libyuv:845, b/129864744
Change-Id: I8ea3959379f1bb2edb857d4eb90fb9a1f6aa4e03
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1899093
Reviewed-by: Dale Curtis <dalecurtis@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
This pulls in the changes that Firefox made to add BT.2020 support as well
as expands them to the existing 10-bit support. So we now have the following
input formats: U420, U422, U444, U010.
BUG=960620, libyuv:845
Change-Id: If0c47853a465d0ed660f849db08e71489fe1b9c2
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1884468
Commit-Queue: Dale Curtis <dalecurtis@chromium.org>
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Recent versions of Clang started warning when the loop doesn't get
vectorized, such as when compiling with -Oz (see bug).
To fix the build, remove the pragma and let the compiler decide on its
own when to vectorize.
Bug: chromium:1015665
Change-Id: I40a610c9e0d94cfd577a6cd2b01e6fdaa08bef7d
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1872580
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Hans Wennborg <hans@chromium.org>
Replace ARM64 only row function with high level function
that implements SSSE3, 32 bit Neon and C.
Compared to 2 step RAWToARGB + ARGBToRGBA on row level:
3.1x faster on ARM
6.2% faster on Intel
BUG=b/140748379
Change-Id: Ia8636d9e4fcdbe10b8c2e81610a54728e29845cd
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1860914
Reviewed-by: richard winterton <rrwinterton@gmail.com>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Neon and GCC Intel optimized, but win32 and mips not optimized.
BUG=libyuv:842, b/141482243
Change-Id: Ia56fa85c8cc1db51f374bd0c89b56d21ec94afa7
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1825642
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: Miguel Casas <mcasas@chromium.org>
Reviewed-by: richard winterton <rrwinterton@gmail.com>
Based on ARGBShuffle but with count adjusted and new shuffle mask
BUG=libyuv:809
Change-Id: Idd936ee6bedcf285607a68c2fc54d876b4becc01
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1711882
Reviewed-by: richard winterton <rrwinterton@gmail.com>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Apply clang-format to fix jpeg if() for lint fix.
Change comments about 4th pixel for open source compliance.
Rename UVToVU to SwapUV for consistency with MergeUV.
BUG=b/135532289, b/136515133
Change-Id: I9ce377c57b1d4d8f8b373c4cb44cd3f836300f79
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1685936
Reviewed-by: Chong Zhang <chz@google.com>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Includes a rounding change for neon.
BUG=b/135532289
Change-Id: I36ffb57b55db6c64804ad169def865be1ac6d66e
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1684439
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: Chong Zhang <chz@google.com>
Gaussian blur low levels ported to 32 bit neon.
But they are not hooked up to anything but a unittest.
Bug:b/248041731, b/132108021, b/129908793
Change-Id: Iccebb8ffd6b719810aa11dd770a525227da4c357
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1611206
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: Chong Zhang <chz@google.com>
Alternatives to RGB24 and AYUV for working with GPU.
BUG=libyuv:832
TESTED=out/Release/libyuv_unittest --gtest_filter=*NV21To???24* --libyuv_width=1280 --libyuv_height=720 --libyuv_repeat=1000 --libyuv_flags=-1 --libyuv_cpu_info=-1
R=rrwinterton@gmail.com
Change-Id: I5559c63f4bd4c847492fcb1571f7b03c58146689
Reviewed-on: https://chromium-review.googlesource.com/c/libyuv/libyuv/+/1501735
Reviewed-by: richard winterton <rrwinterton@gmail.com>
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
I422ToRGB24 is implemented as a C wrapper for Intel, calling
I422ToARGB and ARGBToRGB24. The ARGBToRGB24 for AVX2 requires 32
pixels.
This CL increases the width alignment required to use I422ToRGB24_AVX2
TBR=rrwinterton0gmail.com
Bug: libyuv:822, b:118386049
Change-Id: I4454f4eece33fbd5f593655f577c9ef5c00d1f63
Tested: locally tested with app that crashed using this function.
Reviewed-on: https://chromium-review.googlesource.com/c/1299931
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Add jpeg to NV21 conversions, unittests and conversions
for I444, I422, I420 and I420 to NV21 needed for internals.
Bug: libyuv:820
Change-Id: Idf0f15f91307e80a82cd23943f6eed5508f13fe2
Tested: out/Release/libyuv_unittest --sandbox_unittests --gtest_filter=*MJ*
Reviewed-on: https://chromium-review.googlesource.com/c/1297710
Reviewed-by: Johann Koenig <johannkoenig@google.com>
RAW is a big endian style RGB buffer with R first in memory, then G and B.
Convert NV21 and NV12 to RAW format.
Performance on SkylakeX for 720p with AVX2
I420ToRAW_Opt (388 ms)
H420ToRAW_Opt (371 ms)
NV12ToRAW_Opt (341 ms)
NV21ToRAW_Opt (339 ms)
SSSE3
I420ToRAW_Opt (507 ms)
H420ToRAW_Opt (481 ms)
NV12ToRAW_Opt (498 ms)
NV21ToRAW_Opt (493 ms)
C
I420ToRAW_Opt (2287 ms)
H420ToRAW_Opt (2246 ms)
NV12ToRAW_Opt (2191 ms)
NV21ToRAW_Opt (2204 ms)
Performance on Pixel 2 for 720p
out/Release/bin/run_libyuv_unittest -v -t 7200 --gtest_filter=*NV??ToR*Opt --libyuv_repeat=1000 --libyuv_width=1280 --libyuv_height=720
LibYUVConvertTest.NV12ToRGB24_Opt (1739 ms)
LibYUVConvertTest.NV21ToRGB24_Opt (1734 ms)
LibYUVConvertTest.NV12ToRAW_Opt (1719 ms)
LibYUVConvertTest.NV21ToRAW_Opt (1691 ms)
LibYUVConvertTest.NV12ToRGB565_Opt (2152 ms)
Bug: libyuv:778, b:117522975
Test: add new NV21ToRAW and NV12ToRAW tests
Change-Id: Ieabb68a2c6d8c26743e609c5696c81bb14fb253f
Reviewed-on: https://chromium-review.googlesource.com/c/1272615
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Reviewed-by: Mirko Bonadei <mbonadei@chromium.org>
The original src_u calculation of FOURCC_I420 shifted half width if
crop_y is odd.
This CL fixs the problem and also add a test case for it.
Bug: b:115278653
Test: pass libyuv_unittest
Change-Id: Ia9732d22e64e13de26df47726ba44ad1c5a06484
Reviewed-on: https://chromium-review.googlesource.com/c/1258743
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
When loading or storing the data, the unaligned address will greatly degrade
the optimization performance, so non-aligned access instructions are required
on the loongson platform.
Also delete the optimization function:ScaleARGBFilterCols_MMI,
because it degraded the performance.
BUG=libyuv:804
R=fbarchard@chromium.org
Change-Id: If4c15886a21cdcbac7ae8b336292e4549acf1e47
Reviewed-on: https://chromium-review.googlesource.com/1164627
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
This was changed in 21be9122aadf7824efe3fc19b2a09ff253a688e1.
Change-Id: I6c04dc92f673557e10c231bd090ec8aa88b6bee4
Reviewed-on: https://chromium-review.googlesource.com/1146183
Reviewed-by: Frank Barchard <fbarchard@chromium.org>
Commit-Queue: Frank Barchard <fbarchard@chromium.org>
