Commit Graph

52 Commits

Author SHA1 Message Date
Dmitry Budnikov
eedd7bfd50 fix size 2018-12-19 17:45:00 +03:00
Ruslan Garnov
5340073770 Removed UNUSED macro 2018-12-13 14:39:33 +03:00
Ruslan Garnov
d38676085a Added "kernels" pass stage to compiler, removed unused opaque from cv::gimpl::Op 2018-12-12 17:22:52 +03:00
Alexander Alekhin
c577c2cc6a gapi: eliminate KW issues 2018-12-03 15:20:47 +03:00
Alexander Alekhin
197285d12a Merge pull request #13331 from dbudniko:dbudniko/gapi_gpu_to_ocl_renaming_patch 2018-11-29 15:39:14 +00:00
Evgeny Latkin
ab430b8c87 Merge pull request #13329 from elatkin:el/gapi_perf_medblur
GAPI (fluid): Median blur optimization (#13329)

* GAPI (fluid): Median blur optimization: reference 3x3

* GAPI (fluid): Median blur optimization: CPU dispatcher

* GAPI (fluid): Median blur optimization: manual CV_SIMD
2018-11-29 18:02:29 +03:00
Dmitry Budnikov
d1029faa90 patch removes gpu mention from ocl backend 2018-11-29 17:05:41 +03:00
Dmitry Budnikov
6374b99a1a Merge pull request #13240 from dbudniko:dbudniko/gapi_gpu_to_ocl_renaming
G-API rename GPU backend to OCL backend (#13240)

* renaming draft

* inline namespace instead non-safe define

* more back compatibility

* Updates after review from Dmitry
2018-11-29 16:29:11 +03:00
Evgeny Latkin
c928c21fe7 Merge pull request #13319 from elatkin:el/gapi_perf_erdilate_2
GAPI (fluid): Erode/Dilate optimization, part 2 (#13319)

* GAPI (fluid): Erode/Dilate optimization: hard-code 3x3 case

* GAPI (fluid): Erode/Dilate optimization: CPU dispatcher

* GAPI (fluid): Erode/Dilate optimization: speed-up 10-15x times with CV_SIMD

* GAPI (fluid): Erode/Dilate optimization: 20-30% speed-up
2018-11-28 19:50:39 +03:00
Evgeny Latkin
992d5b8bcd Merge pull request #13315 from elatkin:el/gapi_perf_erdilate
GAPI (fluid): Erode/Dilate optimization (#13315)

* GAPI (fluid): Erode/Dilate optimization: hard-code 3x3 case

* GAPI (fluid): Erode/Dilate optimization: CPU dispatcher

* GAPI (fluid): Erode/Dilate optimization: speed-up 10-15x times with CV_SIMD
2018-11-28 18:20:31 +03:00
Evgeny Latkin
6808d33b2f Merge pull request #13290 from elatkin:el/gapi_perf_filter2d
GAPI (fluid): Filter 2D optimization (#13290)

* GAPI (fluid): Filter 2D optimization: speedup 13x if float, 2x if integral

* GAPI (fluid): Filter 2D speedup 8x if output is short/ushort

* GAPI (fluid): Filter 2D speedup 7x if output is uchar

* GAPI (fluid): Filter 2D optimization: fixed compiler warnings

* GAPI (fluid): fix compiler warnings on Mac

* GAPI (fluid): fix compiler warnings on Mac

* GAPI (fluid): fix compiler errors on VS2015

* GAPI (fluid): fix compiler errors on VS2015

* GAPI (fluid): fix compiler errors on VS2015
2018-11-27 19:12:14 +03:00
Dmitry Budnikov
51cc78b2a2 Merge pull request #13251 from dbudniko:dbudniko/gapi_more_fixes_for_tests
More fixes for G-API tests (#13251)

* scalar comparator and more fixes for tests

* add weighted aligned

* white space

* more white space

* Add weighted test accuracy check enabled
2018-11-26 17:44:46 +03:00
Evgeny Latkin
f07856eab9 Merge pull request #13221 from elatkin:el/gapi_perf_sepfilter
GAPI (fluid): optimization of Separable filter (#13221)

* GAPI (fluid): Separable filter: performance test

* GAPI (fluid): enable all performance tests

* GAPI: separable filters: alternative code for Sobel

* GAPI (fluid): hide unused old code for Sobel filter

* GAPI (fluid): especial code for Sobel if U8 into S16

* GAPI (fluid): back to old code for Sobel

* GAPI (fluid): run_sepfilter3x3_impl() with CPU dispatcher

* GAPI (fluid): run_sepfilter3x3_impl(): fix compiler warnings

* GAPI (fluid): new engine for separable filters (but Sobel)

* GAPI (fluid): new performance engine for Sobel

* GAPI (fluid): Sepfilters performance: fixed compilation error
2018-11-26 15:05:35 +03:00
Alexander Alekhin
c6daa4aa16 Merge pull request #13260 from alalek:cmake_ade_12856 2018-11-23 17:09:47 +00:00
Dmitry Budnikov
a518e7063d Merge pull request #13120 from dbudniko:dbudniko/gapi_opencl_kernel_example
* custom OpenCL G-API kernel draft

* clean up and warnings fix

* more warnings

* white space

* new blank line at the EOF removed

* HAVE_OPENCL guard

* remove unnecessary ocl API call

* remove sum test workaround

* check if opencl activated

* fix std::str warning

* CPU fall back for symm7x7

* gpu test kernel draft

* adjust have opencl guard

* more guards

* one more attempt to adjust guards

* empty stub files and kernel source files creation in the test directory

* try to force auto generation

* one more attempt to force build

* remove symm7x7 custom from gapi module

* looks like that this version works properly on Win desktop

* clean up

* more clean up

* address some suggestions from Dmitry's review

* const kernel coefficients

* CV_Error in kernel + try to fix cpu fallback

* CV_Error_ instead CV_Error

* everything in one gapi_gpu_test.cpp

* fix warning

* remove kernel generation, add kernel string

* avoid generated code and ocl internal namespace

* fix misprint

* c_str
2018-11-23 17:51:15 +03:00
Alexander Alekhin
de8696aa43 cmake: allow to disable ADE build too (BUILD_opencv_gapi=OFF is not enough)
CMake option: WITH_ADE=OFF
2018-11-23 12:52:41 +03:00
Ruslan Garnov
a3df05d93b Merge pull request #13215 from rgarnov:rg/overhead
* Added caching of agents execution sequence

* Merged linesRead() and nextWindow() methods on FluidAgent in one method

* Added caching of input lines for fluid::View

* Added caching of output lines for fluid::Buffer

* Fixed GAPI_Assert to work in standalone mode
2018-11-20 17:25:04 +03:00
Dmitry Matveev
1b13df5368 G-API: Recent inclusion has broken STANDALONE build
This MR fixes this.
2018-11-20 14:00:08 +03:00
Evgeny Latkin
083332f85f Merge pull request #13206 from elatkin:el/gapi_perf_rgb2lab
GAPI (fluid): RGB to Lab optimization (#13206)

* GAPI (fluid): BGR2LUV, RGB2Lab: performance test

* GAPI (fluid): BGR2LUV, RGB2Lab: using cv::hal::cvtBGRtoLab

* GAPI (fluid): BGR2LUV, RGB2Lab: hide reference code with #ifdef
2018-11-19 18:52:48 +03:00
Evgeny Latkin
6757c2c5a6 Merge pull request #13174 from elatkin:el/gapi_perf_rgb2yuv
GAPI (fluid): RGB to YUV optimization (#13174)

* GAPI (fluid): RGB to YUV: activate performance tests

* GAPI (fluid): speedup 4-8x RGB-to-YUV, 2.5x YUV to RGB with int16 arithmetic

* GAPI (fluid): RGB <--> YUV: fixed compiler warning

* GAPI (fluid): RGB <--> YUV: additional speedup 2-3x times (10-15x over original) via manual CV_SIMD

* GAPI (fluid): RGB <--> YUV: dynamic CV_SIMD dispatching
2018-11-19 17:30:14 +03:00
Alexander Alekhin
a574788e89 move legacy C-API constants into separate files 2018-11-17 23:47:51 +00:00
Dmitry Matveev
d7540c9a3c Merge pull request #13176 from dmatveev:gapi_doxygen
G-API: Doxygen class reference

* G-API Doxygen documentation: covered cv::GComputation

* G-API Doxygen documentation: added sections on compile arguments

* G-API Doxygen documentation: restructuring & more text

* Added new sections (organized API reference into it);
* Documented GCompiled, compile args, backends, etc.

* G-API Doxygen documentation: documented GKernelPackage and added group for meta
2018-11-16 23:38:10 +03:00
Evgeny Latkin
f81370232a Merge pull request #13162 from elatkin:el/gapi_perf_rgb2gray
GAPI (fluid): RGB/BGR to gray: optimization (#13162)

* GAPI (fluid): RGB/BGR to Gray: add performance tests

* GAPI (fluid): RGB/BGR to Gray: speedup 8-12x with manual CV_SIMD

* GAPI (fluid): RGB/BGR to Gray: fix compiler warning

* GAPI (fluid): RGB/BGR to Gray: dynamic dispatching to AVX2

* GAPI (fluid): RGB/BGR to Gray: check R/G/B coefficients

* GAPI (fluid): RGB/BGR to Gray: fixed compilation error (caused by change in master)
2018-11-15 18:14:27 +03:00
Dmitry Matveev
85fad1504a Merge pull request #13030 from dmatveev:tutorial
* G-API: First steps with tutorial

* G-API Tutorial: First iteration

* G-API port of anisotropic image segmentation tutorial;
* Currently works via OpenCV only;
* Some new kernels have been required.

* G-API Tutorial: added chapters on execution code, inspection, and profiling

* G-API Tutorial: make Fluid kernel headers public

For some reason, these headers were not moved to the public
headers subtree during the initial development. Somehow it even
worked for the existing workloads.

* G-API Tutorial: Fix a couple of issues found during the work

* Introduced Phase & Sqrt kernels, OCV & Fluid versions
* Extended GKernelPackage to allow kernel removal & policies on include()

All the above stuff needs to be tested, tests will be added later

* G-API Tutorial: added chapter on running Fluid backend

* G-API Tutorial: fix a number of issues in the text

* G-API Tutorial - some final updates

- Fixed post-merge issues after Sobel kernel renaming;
- Simplified G-API code a little bit;
- Put a conclusion note in text.

* G-API Tutorial - fix build issues in test/perf targets

Public headers were refactored but tests suites were not updated in time

* G-API Tutorial: Added tests & reference docs on new kernels

* Phase
* Sqrt

* G-API Tutorial: added link to the tutorial from the main module doc

* G-API Tutorial: Added tests on new GKernelPackage functionality

* G-API Tutorial: Extended InRange tests to cover 32F

* G-API Tutorial: Misc fixes

* Avoid building examples when gapi module is not there
* Added a volatile API disclaimer to G-API root documentation page

* G-API Tutorial: Fix perf tests build issue

This change came from master where Fluid kernels are still used
incorrectly.

* G-API Tutorial: Fixed channels support in Sqrt/Phase fluid kernels

Extended tests to cover this case

* G-API Tutorial: Fix text problems found on team review
2018-11-15 18:12:36 +03:00
Alexander Alekhin
ca9aa180c1 Merge pull request #13096 from alalek:gapi_tests_checks 2018-11-14 15:59:01 +00:00
Alexander Alekhin
6189b47648 gapi(test): more reliable checks
avoid `countNonZero()`, use `norm()`
2018-11-14 13:30:53 +03:00
Alexander Alekhin
dd6f5949c2 gapi(test): use relative error check for Norm/Sum tests 2018-11-14 13:30:53 +03:00
Latkin, Yevgeny I
a62539489d GAPI (fluid): Sobel 3x3 optimization: remove needless file 2018-11-14 10:17:40 +03:00
Alexander Alekhin
8b7f805642 Merge pull request #13150 from dmatveev:update_ade011d 2018-11-13 16:18:32 +00:00
Evgeny Latkin
cc5190eb91 Merge pull request #13133 from elatkin:el/gapi_perf_sobel_2
GAPI (fluid): Sobel 3x3 optimization: CV_SIMD dynamic dispatching (#13133)

* GAPI (fluid): Sobel 3x3: remove template for run_sobel_row()

* GAPI (fluid): Sobel 3x3: dynamic dispatching of CV_SIMD code

* GAPI (fluid): Sobel 3x3 optimization: fixed CV_SIMD dynamic dispatcher
2018-11-13 17:48:10 +03:00
Dmitry Matveev
4eff798270 Update ADE to version 0.1.1d 2018-11-13 16:44:29 +03:00
Evgeny Latkin
4e40e5bb88 Merge pull request #13070 from elatkin:el/gapi_perf_sobel
GAPI (fluid): optimization of Sobel 3x3 (#13070)

* GAPI: performance test for Sobel

* GAPI: performance test for Sobel w/FP32 input

* GAPI: Sobel speedup: 2.5x (U8) up to 10x (float)

* GAPI: Sobel 3x3 to support U8 into S16

* GAPI (fluid): Sobel 3x3 speedup: 10% (uchar), 1.5x (float)

* GAPI (fluid): Sobel 3x3 speedup: +10x (uchar), but -20% (float)

* GAPI (fluid): Sobel 3x3 speedup: +10% (float)

* GAPI (fluid): Sobel 3x3 speedup: +15% (float), +10% (uchar)

* GAPI (fluid): Sobel 3x3: address GCC warnings

* GAPI (fluid): Sobel 3x3: separate *.cpp file w/SIMD code

* GAPI (fluid): Sobel 3x3: fixed AVX2 code, AVX2 speedup 20-50% (uchar), 10-20% (float)

* GAPI (fluid): Sobel 3x3: fix CV_SIMD code for AVX2

* GAPI (fluid): Sobel 3x3: refactor
2018-11-13 15:04:37 +03:00
Alexander Alekhin
75a6eaf147 Merge pull request #13024 from rgarnov:gapi_fix_lut_test 2018-11-09 19:16:39 +00:00
Ruslan Garnov
858ba6c798 Fixed uniniitialized mats in tests, added operator<< for compare_f 2018-11-09 12:00:55 +03:00
Alexander Alekhin
386733aea1 gapi: fix ARM build 2018-11-09 06:37:48 +00:00
Dmitry Budnikov
5087ff0814 Merge pull request #13008 from dbudniko:dbudniko/gpu_opencl_backend
G-API GPU-OpenCL backend (#13008)

* gpu/ocl backend core

* accuracy tests added and adjusted + license headers

* GPU perf. tests added; almost all adjusted to pass

* all tests adjusted and passed - ready for pull request

* missing license headers

* fix warning (workaround RGB2Gray)

* fix c++ magic

* precompiled header

* white spaces

* try to fix warning and blur test

* try to fix Blur perf tests

* more alignments with the latest cpu backend

* more gapi tests refactoring + 1 more UB issue fix + more informative tolerance exceed reports

* white space fix

* try workaround for SumTest

* GAPI_EXPORTS instead CV_EXPORTS
2018-11-08 22:14:53 +03:00
Alexander Alekhin
3c99dc96b1 Merge pull request #13001 from rgarnov:gapi_standalone_cmake 2018-11-07 15:12:29 +00:00
Ruslan Garnov
c972c94662 Added Standalone.cmake 2018-11-07 14:54:18 +03:00
Evgeny Latkin
000110e760 Merge pull request #13052 from elatkin:yl/gapi_sobel3x3_f32_v2
* GAPI: Sobel 3x3 with FP32 input

* GAPI: Sobel 3x3 with FP32 input, v2
2018-11-06 23:08:50 +03:00
Ruslan Garnov
443fed796a Merge pull request #12990 from rgarnov:gapi_fluid_reshape_support
G-API: Introduce new `reshape()` API (#12990)

* Moved initFluidUnits, initLineConsumption, calcLatency, calcSkew to separate functions

* Added Fluid::View::allocate method (moved allocation logic from constructor)

* Changed util::zip to util::indexed, utilized collectInputMeta in GFluidExecutable constructor

* Added makeReshape method to FluidExecutable

* Removed m_outputRoi from GFluidExecutable

* Added reshape feature

* Added switch of resize mapper if agent ratio was changed

* Added more TODOs and renamed a function

* G-API reshape(): add missing `override` specifiers

Fix warnings on all platforms
2018-10-30 21:12:36 +03:00
AsyaPronina
08536943ad Merge pull request #12949 from AsyaPronina:missed_multiply_operator_for_GMAT
Made scale parameter optional for mul kernel wrapper (#12949)

* Added missed operator*(GMat, GMat). Made scale parameter optional for mul kernel.

* Fixed perf test for mul(GMat, GMat) kernel

* Removed operator*(GMat, GMat) as not needed
2018-10-30 21:10:47 +03:00
Dmitry Budnikov
220b278575 Merge pull request #12944 from dbudniko:dbudniko/tests_thresholds
* added threshold to g-api imgproc tests

* trailing spaces remove

* try to fix warnings

* fix license header

* fix random numbers issue

* fix Sobel test

* license headers + precompiled headers included
2018-10-29 17:34:20 +03:00
AsyaPronina
e6dd9a78ae Merge pull request #12947 from AsyaPronina:sobel_renaming
* Renamed Sobel operator GAPI kernel to match with OpenCV naming rules

* Fixed perf tests

* Small refactoring to check CI issue

* Refactored alignment for kernel wrappers in imgproc.hpp
2018-10-27 19:54:32 +03:00
Dmitry Matveev
dbed39a931 Merge pull request #12857 from dmatveev:hld
* G-API Documentation: first submission

This PR introduces a number of new OpenCV documentation chapters for
Graph API module.

In particular, the following topics are covered:
- Introduction & background information;
- High-level design overview;
- Kernel API;
- Pipeline example.

All changes are done in Markdown files, no headers, etc modified.

Doxygen references for main API classes will be added later.

Also, a tutorial will be introduced soon (in the common Tutorials place)

* G-API Documentation - fix warnings & trailing whitespaces

* G-API Documentation: address review issues

* G-API Documentation: export code snippets to compileable files

* gapi: move documentation samples
2018-10-24 07:47:56 +03:00
Dmitry Matveev
5e9750d1f5 Merge pull request #12870 from dmatveev:gapi_fluid_basic_hetero_support
* G-API Fluid basic heterogeneity support: initial upload

* G-API Fluid heterogeneity: address some coding style issues

* G-API Fluid heterogeneity: fix compiler warnings

* G-API Fluid heterogeneity: fix warnings on Windows & ARMv7

* G-API Fluid heterogeneity: finally fix Windows warnings

* G-API Fluid heterogeneity: fix dangling reference problem
2018-10-19 18:32:48 +03:00
Alexander Alekhin
1188c6fe14 Merge pull request #12838 from dmatveev:gapi_fluid_resize_lpi 2018-10-17 09:22:28 +00:00
Alexander Alekhin
13aa23f706 Merge pull request #12836 from dmatveev:gapi_upd151018 2018-10-15 16:20:52 +00:00
Dmitry Matveev
922d5796b9 G-API: Introduce LPI (multiple Lines-Per-Iteration) support for Resize
Several Resize optimizations count on fetching multiple input lines at
once to do interpolation more efficiently.

At the moment, Fluid backend supports only LPI=1 for Resize kernels.

This patch introduces scheduling support for Resizes with LPI>1 and
covers these cases with new tests.

The support is initially written by Ruslan Garnov.
2018-10-15 17:11:55 +03:00
Dmitry Matveev
f0ddc2a53c G-API: Fixed static analysis issue in own::Mat
Tool reported a false alarm on possible out-of-bounds access which
was work-arounded to make code more clear
2018-10-15 16:31:31 +03:00
Alexander Alekhin
d5d059690f gapi(test): disable divide tests with zeros 2018-10-14 10:47:50 +00:00