Commit Graph

138 Commits

Author SHA1 Message Date
Alexander Alekhin
6997d423c8 fix normalize in case of inplace operations
fixes #5876
2015-12-25 15:33:06 +03:00
Maksim Shabunin
b4bcdd10a1 HAL: improvements
- added new functions from core module: split, merge, add, sub, mul, div, ...
- added function replacement mechanism
- added example of HAL replacement library
2015-12-03 14:43:37 +03:00
hoangviet1985
e679d97100 remove redundant code 2015-11-22 14:32:18 -05:00
Maksim Shabunin
6e9d0d9a0c Visual Studio 2015 warning and test fixes 2015-10-20 12:48:37 +03:00
Pavel Vlasov
14b006e808 IPP_VERSION_X100 was changed to:
IPP_VERSION_MAJOR * 100 + IPP_VERSION_MINOR*10 + IPP_VERSION_UPDATE
to manage changes between updates more easily.

IPP_DISABLE_BLOCK was added to ease tracking of disabled IPP functions;
2015-09-25 17:50:15 +03:00
Pavel Vlasov
e57609836c Warning fix; 2015-06-24 14:43:42 +03:00
Pavel Vlasov
101607a7d0 Imgproc_Hist_MinMaxVal.accuracy fix;
Some code style corrections;
2015-06-24 13:50:17 +03:00
Dmitry Budnikov
a5a21019b2 ipp_countNonZero build fix;
Removed IPP port for tiny arithm.cpp functions

Additional warnings fix on various platforms.

Build without OPENCL and GCC warnings fixed

Fixed warnings, trailing spaces and removed unused secure_cpy.

IPP code refactored.

IPP code path  implemented as separate static functions to simplify future work with IPP code and make it more readable.
2015-06-18 12:47:07 +03:00
Ilya Lavrenov
6bce6ee34a checks 2015-01-12 10:59:31 +03:00
Ilya Lavrenov
f2cd65cf1e fixes 2015-01-12 10:59:30 +03:00
Ilya Lavrenov
70933ea999 convert from f64 2015-01-12 10:59:30 +03:00
Ilya Lavrenov
612b8ce2cb fixed compilation issues 2015-01-12 10:59:30 +03:00
Ilya Lavrenov
d87457a025 split/merge 2015-01-12 10:59:30 +03:00
Ilya Lavrenov
fc0869735d used popcnt 2015-01-12 10:59:30 +03:00
Ilya Lavrenov
63fc6ef316 convertTo from 64f 2015-01-12 10:59:29 +03:00
Ilya Lavrenov
8870ef4159 convertTo from 16u, 16s, 32s, 32f 2015-01-12 10:59:29 +03:00
Ilya Lavrenov
116fb275a8 convertTo from 8s 2015-01-12 10:59:29 +03:00
Ilya Lavrenov
19e77e4787 convertTo from 8u 2015-01-12 10:59:29 +03:00
Ilya Lavrenov
b758dbd384 convertTo AVX2 2015-01-12 10:59:29 +03:00
Ilya Lavrenov
3a78a22733 convertScaleAbs for s8, f64 2015-01-12 10:59:29 +03:00
Ilya Lavrenov
8d48632ebe avx2 2015-01-12 10:59:28 +03:00
vbystricky
a8aa6381d9 Optimize OpenCL version of conversScaleAbs function 2014-10-21 19:20:20 +04:00
Pavel Vlasov
45958eaabc Implementation detector and selector for IPP and OpenCL;
IPP can be switched on and off on runtime;

Optional implementation collector was added (switched off by default in CMake). Gathers data of implementation used in functions and report this info through performance TS;

TS modifications for implementations control;
2014-10-15 14:24:41 +04:00
Vadim Pisarevsky
a798386660 Merge pull request #3326 from ilya-lavrenov:neon_canny 2014-10-11 17:58:24 +00:00
Ilya Lavrenov
5f23d99918 the rest modes of cv::Mat::convertTo 2014-10-10 14:10:50 +00:00
Ilya Lavrenov
00f16e9178 neon 2014-10-03 08:43:02 +00:00
Ilya Lavrenov
bbc161e1cb fix for cv::Mat::convertTo with scale 2014-09-28 14:51:30 -07:00
Ilya Lavrenov
e46332a183 cv::Mat::convertTo with scale and shift 2014-09-28 03:49:56 -07:00
Ilya Lavrenov
345b1369be correct neon rounding 2014-09-25 07:54:52 +00:00
Ilya Lavrenov
4b3f2c1972 Neon optimization of Mat::convertTo 2014-09-23 15:06:17 +00:00
Ilya Lavrenov
515be70867 Neon optimization of cv::convertScaleAbs 2014-09-22 15:47:46 +00:00
Ilya Lavrenov
5d3a128cd3 NEON impl on cv::convertScaleAba CV_32f 2014-09-01 17:04:36 +00:00
Alexander Alekhin
52ac61d87c Merge pull request #3088 from vbystricky:ocl_enableNormEtc 2014-08-14 14:34:40 +00:00
Vadim Pisarevsky
5240a03217 Merge pull request #3100 from vpisarev:disabled_neon_split_merge_on_ios 2014-08-14 08:51:44 +00:00
Vadim Pisarevsky
5d8098a2c3 avoid compiler warnings 2014-08-14 10:06:10 +04:00
Vadim Pisarevsky
a37876ab95 fixed build of iOS framework another time; clang does not like the new NEON-optimized split/merge 2014-08-13 19:59:26 +04:00
vbystricky
1fe403f461 Enable OpenCL version of norm and convertScaleAbs or 32F data
Fix error in minmaxloc.cl
Change test for convertScaleAbs
Fix minMaxIdx for _src2 align
Change epsilon on the tests
2014-08-13 18:33:01 +04:00
Adil Ibragimov
8a4a1bb018 Several type of formal refactoring:
1. someMatrix.data -> someMatrix.prt()
2. someMatrix.data + someMatrix.step * lineIndex -> someMatrix.ptr( lineIndex )
3. (SomeType*) someMatrix.data -> someMatrix.ptr<SomeType>()
4. someMatrix.data -> !someMatrix.empty() ( or !someMatrix.data -> someMatrix.empty() ) in logical expressions
2014-08-13 15:21:35 +04:00
Alexander Alekhin
8fd61c9aa6 Merge pull request #3075 from akarsakov:ipp_imgproc_fix 2014-08-11 15:02:53 +00:00
Elena Gvozdeva
b73b28ab9c fix for ocl_lut 2014-08-11 16:21:55 +04:00
Alexander Karsakov
835b5e2179 Disabled ippiConvert_32f16s_C1R since it breaks OCL accuracy tests of remap with test_loop_times=30. 2014-08-11 15:50:06 +04:00
Vadim Pisarevsky
ca183707f2 Merge pull request #3039 from alalek:ocv_world 2014-08-06 14:52:52 +00:00
Vadim Pisarevsky
d611684d9e Merge pull request #3041 from adrians:optimize_split 2014-08-06 09:22:38 +00:00
Alexander Alekhin
55188fe991 world fix 2014-08-05 20:12:35 +04:00
Adrian Stratulat
b2acd1f75a Vectorize split & merge for NEON 2014-08-05 06:51:36 +00:00
Vadim Pisarevsky
5267ed46c7 1. disabled OpenCL acceleration for cv::multiply() (CV_32F), cv::divide (CV_32F), cv::convertScaleAbs (CV_32F) and cv::reduce (SUM, CV_32F), cv::reduce (MIN & MAX), cv::flip (3-channel case).
2. changed the number of test loops from 1 to 30 (except for cv::pow() test, which fails for yet unknown reason)
3. disabled IPP acceleration for 3-channel norms.
4. modified relativeNorm test function to handle very small values
2014-07-30 18:19:47 +04:00
Ilya Lavrenov
10c772fa7f minor optimization of cv::LUT 2014-07-02 18:50:21 +04:00
Ilya Lavrenov
972119bff3 sse2 optimization of cv::convertScaleAbs 2014-07-01 02:21:00 +04:00
Alexander Alekhin
ea417ac04a Merge pull request #2755 from vbystricky:oclopt_LUT 2014-06-06 18:52:40 +04:00
VBystricky
5d924b7a75 If lut table has one channel and src aligned to 4, work with src as with one channel matrix 2014-06-05 19:31:31 +04:00