Improve performance on Arm64
* Improve performance on Apple silicon
This patch will
- Enable dot product intrinsics for macOS arm64 builds
- Enable for macOS arm64 builds
- Improve HAL primitives
- reduction (sum, min, max, sad)
- signmask
- mul_expand
- check_any / check_all
Results on a M1 Macbook Pro
* Updates to #20011 based on feedback
- Removes Apple Silicon specific workarounds
- Makes #ifdef sections smaller for v_mul_expand cases
- Moves dot product optimization to compiler optimization check
- Adds 4x4 matrix transpose optimization
* Remove dotprod and fix v_transpose
Based on the latest, we've removed dotprod entirely and will revisit in a future PR.
Added explicit cats with v_transpose4x4()
This should resolve all opens with this PR
* Remove commented out lines
Remove two extraneous comments
* Add the support for riscv64 vector 0.7.1.
* fixed GCC warnings
* cleaned whitespaces
* Remove the worning by the use of internal API of compiler.
* Update the license header.
* removed trailing whitespaces
Co-authored-by: Vadim Pisarevsky <vadim.pisarevsky@me.com>
Co-authored-by: yulj <linjie.ylj@alibaba-inc.com>
Co-authored-by: Vadim Pisarevsky <vadim.pisarevsky@gmail.com>
* Adding functions rbegin() and rend() functions to matrix class.
This is important to be more standard compliant with C++ and an ever increasing number of people using standard algorithms for better code readability- and maintainability.
The functions are copy pated from their counterparts (even though they should probably call the counterparts but this gave me some troube).
They return iterators using std::reverse_iterators
Follow up of an open feature request:
https://github.com/opencv/opencv/issues/4641
* Fix rbegin() and rend() and provide tests for them
* Removing unnecessary whitespaces
* Adding rbegin and rend to Mat_ class with the right parameters so we don't need to repeat the template argument.
An instantiating cv::Mat_<int> for example can call it's rbegin() function and doesn't need rbegin<int>() with this convience addition.
Follows what is done for forward iterators
* static cast the vector size (return size_t) to an int (that is required for opencv mat constructor)
Co-authored-by: Stefan <stefan.gerl@tum.de>
* Updated cpp reference implementations for a few intrinsics to address wide universal intrinsics as well
* Updated cpp reference implementations for a few more universal intrinsics
* Update polynom_solver.cpp
This pull request is in the response to Issue #19526. I have fixed the problem with the cube root calculation of 2*R. The Issue was in the usage of pow function with negative values of R, but if it is calculated for only positive values of R then changing x0 according to the parity of R, the Issue is resolved. Kindly consider it, Thanks!
* add cv::cubeRoot(double)
Co-authored-by: Alexander Alekhin <alexander.a.alekhin@gmail.com>
* Add Python Bindings for getCacheDirectory function
* Added getCacheDirectory interop test with image codecs.
Co-authored-by: Sergey Slashchinin <sergei.slashchinin@xperience.ai>
- follows iso c++ guideline C.44
- enables default compiler-created constructors to
also be noexcept
original commit: 77e26a7db3
- handled KernelArg, Image2D
* [hal][neon] Optimize the v_dotprod_fast intrinsics for aarch64.
On Armv8 in AArch64 execution mode, we can skip the sequence
v<op>_<ty>(vget_high_<ty>(x), vget_high_<ty>(y))
in favour of
v<op>_high_<ty>(x, y)
This has better changes for recent compilers to use less data movement
operations and better register allocation. See for example:
https://godbolt.org/z/bPq7vd
* [hal][neon] Fix build failure on armv7.
* [hal][neon] Address review comments in PR.
PR: https://github.com/opencv/opencv/pull/19486
* [hal][neon] Define macro to check for the AArch64 execution state of Armv8.
* [hal][neon] Fix macro definition for AArch64.
The fix is needed to prevent warnings when building for Armv7.
they might be thrown from third-party code (notably Ogre in the ovis
module).
While Linux is kind enough to print them, they cause instant termination
on Windows.
Arguably, they do not origin from OpenCV itself, but still this helps
understanding what went wrong when calling an OpenCV function.
Ordinary quaternion
* version 1.0
* add assumeUnit;
add UnitTest;
check boundary value;
fix the func using method: func(obj);
fix 4x4;
add rodrigues vector transformation;
fix mat to quat;
* fix blank and tab
* fix blank and tab
modify test;cpp to hpp
* mainly improve comment;
add rvec2Quat;fix toRodrigues;
fix throw to CV_Error
* fix bug of quatd * int;
combine hpp and cpp;
fix << overload error in win system;
modify include in test file;
* move implementation to quaternion.ini.hpp;
change some constructor to createFrom* function;
change Rodrigues vector to rotation vector;
change the matexpr to mat of 3x3 return type;
improve comments;
* try fix log function error in win
* add enums for assumeUnit;
improve docs;
add using std::cos funcs
* remove using std::* from header;
add std::* in affine.hpp,warpers_inl.hpp;
* quat: coding style
* quat: AssumeType => QuatAssumeType
[GSoC] OpenCV.js: WASM SIMD optimization 2.0
* gsoc_2020_simd Add perf test for filter2d
* add perf test for kernel scharr and kernel gaussianBlur
* add perf test for blur, medianBlur, erode, dilate
* fix the errors for the opencv PR robot
fix the trailing whitespace.
* add perf tests for kernel remap, warpAffine, warpPersepective, pyrDown
* fix a bug in modules/js/perf/perf_imgproc/perf_remap.js
* add function smoothBorder in helpfun.js and remove replicated function in perf test of warpAffine and warpPrespective
* fix the trailing white space issues
* add OpenCV.js loader
* Implement the Loader with help of WebAssembly Feature Detection, remove trailing whitespaces
* modify the explantion for loader in js_setup.markdown and fix bug in loader.js
- Added cross compile cmake file for target riscv64-clang
- Extended cmake for RISC-V and added instruction checks
- Created intrin_rvv.hpp with C++ version universal intrinsics
* Add documentation about usage of cv2eigen functions in eigen.hpp
* Fixed Doxygen syntax.
Co-authored-by: Alexander Smorkalov <smorkalov.a.m@gmail.com>
* fixed#17044
1. fixed Python part of the tutorial about using OpenCV XML-YAML-JSON I/O functionality from C++ and Python.
2. added startWriteStruct() and endWriteStruct() methods to FileStorage
3. modifed FileStorage::write() methods to make them work well inside sequences, not only mappings.
* try to fix the doc builder
* added Python regression test for FileStorage I/O API ([TODO] iterating through long sequences can be very slow)
* fixed yaml testing
* add eigen tensor conversion functions
* add eigen tensor conversion tests
* add support for column major order
* update eigen tensor tests
* fix coding style and add conditional compilation
* fix conditional compilation checks
* remove whitespace
* rearrange functions for easier reading
* reformat function documentation and add tensormap unit test
* cleanup documentation of unit test
* remove condition duplication
* check Eigen major version, not minor version
* restrict to Eigen v3.3.0+
* add documentation note and add type checking to cv2eigen_tensormap()
Vectorize minMaxIdx functions
* Updated documentation and intrinsic tests for v_reduce
* Add other files back in from the forced push
* Prevent an constant overflow with v_reduce for int8 type
* Another alternative to fix constant overflow warning.
* Fix another compiler warning.
* Update comments and change comparison form to be consistent with other vectorized loops.
* Change return type of v_reduce_min & max for v_uint8 and v_uint16 to be same as lane type.
* Cast v_reduce functions to int to avoid overflow. Reduce number of parameters in MINMAXIDX_REDUCE macro.
* Restore cast type for v_reduce_min & max to LaneType
* resize: HResizeLinear reduce duplicate work
There appears to be a 2x unroll of the HResizeLinear against k,
however the k value is only incremented by 1 during the unroll. This
results in k - 1 duplicate passes when k > 1.
Likewise, the final pass may not respect the work done by the vector
loop. Start it with the offset returned by the vector op if
implemented. Note, no vector ops are implemented today.
The performance is most noticable on a linear downscale. A set of
performance tests are added to characterize this. The performance
improvement is 10-50% depending on the scaling.
* imgproc: vectorize HResizeLinear
Performance is mostly gated by the gather operations
for x inputs.
Likewise, provide a 2x unroll against k, this reduces the
number of alpha gathers by 1/2 for larger k.
While not a 4x improvement, it still performs substantially
better under P9 for a 1.4x improvement. P8 baseline is
1.05-1.10x due to reduced VSX instruction set.
For float types, this results in a more modest
1.2x improvement.
* Update U8 processing for non-bitexact linear resize
* core: hal: vsx: improve v_load_expand_q
With a little help, we can do this quickly without gprs on
all VSX enabled targets.
* resize: Fix cn == 3 step per feedback
Per feedback, ensure we don't overrun. This was caught via the
failure observed in Test_TensorFlow.inception_accuracy.
* core: disable invalid constructors in C API by default
- C API objects will lose their default initializers through constructors
* samples: stop using of C API
Tests for argument conversion of Python bindings generator
* Tests for parsing elemental types from Python bindings
- Add positive and negative tests for int, float, double, size_t,
const char*, bool.
- Tests with wrong conversion behavior are skipped.
* Move implicit conversion of bool to integer/floating types to wrong
conversion behavior.
Improving VSX performance of integral function
* Adding support for vector get function on VSX datatypes so the
integral function gains a bit of performance.
* Removing get as a datatype member function and implementing a new HAL
instruction v_extract_n to get the n-th element of a vector register.
* Adding SSE/NEON/AVX intrinsics.
* Implement new HAL instruction v_broadcast_element on VSX/AVX/NEON/SSE.
* core(simd): add tests for v_extract_n/v_broadcast_element
- updated docs
- commented out code to repair compilation
- added WASM and MSA default implementations
* core(simd): fix compilation
- x86: avoid _mm256_extract_epi64/32/16/8 with MSVS 2015
- x86: _mm_extract_epi64 is 64-bit only
* cleanup
- move TLS & instrumentation code out of core/utility.hpp
- (*) TLSData lost .gather() method (to dispose thread data on thread termination)
- use TLSDataAccumulator for reliable collecting of thread data
- prefer using of .detachData() + .cleanupDetachedData() instead of .gather() method
(*) API is broken: replace TLSData => TLSDataAccumulator if gather required
(objects disposal on threads termination is not available in accumulator mode)
Fixing bug with comparison of v_int64x2 or v_uint64x2
* Casting v_uint64x2 to v_float64x2 and comparing does NOT work in all cases. Rewrite using epi64 instructions - faster too.
* Fix bad merge.
* Fix equal comparsion for non-SSE4.1. Add test cases for v_int64x2 comparisons.
* Try to fix merge conflict.
* Only test v_int64x2 comparisons if CV_SIMD_64F
* Fix compiler warning.
* New v_reverse HAL intrinsic for reversing the ordering of a vector
* Fix conflict.
* Try to resolve conflict again.
* Try one more time.
* Add _MM_SHUFFLE. Remove non-vectorize code in SSE2. Fix copy and paste issue with NEON.
* Change v_uint16x8 SSE2 version to use shuffles
* core: rework and optimize SIMD implementation of dotProd
- add new universal intrinsics v_dotprod[int32], v_dotprod_expand[u&int8, u&int16, int32], v_cvt_f64(int64)
- add a boolean param for all v_dotprod&_expand intrinsics that change the behavior of addition order between
pairs in some platforms in order to reach the maximum optimization when the sum among all lanes is what only matters
- fix clang build on ppc64le
- support wide universal intrinsics for dotProd_32s
- remove raw SIMD and activate universal intrinsics for dotProd_8
- implement SIMD optimization for dotProd_s16&u16
- extend performance test data types of dotprod
- fix GCC VSX workaround of vec_mule and vec_mulo (in little-endian it must be swapped)
- optimize v_mul_expand(int32) on VSX
* core: remove boolean param from v_dotprod&_expand and implement v_dotprod_fast&v_dotprod_expand_fast
this changes made depend on "terfendail" review
- renamed Cascade Lake AVX512_CEL => AVX512_CLX (align with Intel SDE tool)
- fixed CLX instruction sets (no IFMA/VBMI)
- added flag to bypass CPU baseline check: OPENCV_SKIP_CPU_BASELINE_CHECK
[GSoC 2019] Improve the performance of JavaScript version of OpenCV (OpenCV.js)
* [GSoC 2019]
Improve the performance of JavaScript version of OpenCV (OpenCV.js):
1. Create the base of OpenCV.js performance test:
This perf test is based on benchmark.js(https://benchmarkjs.com). And first add `cvtColor`, `Resize`, `Threshold` into it.
2. Optimize the OpenCV.js performance by WASM threads:
This optimization is based on Web Worker API and SharedArrayBuffer, so it can be only used in browser.
3. Optimize the OpenCV.js performance by WASM SIMD:
Add WASM SIMD backend for OpenCV Universal Intrinsics. It's experimental as WASM SIMD is still in development.
* [GSoC2019]
1. use short license header
2. fix documentation node issue
3. remove the unused `hasSIMD128()` api
* [GSoC2019]
1. fix emscripten define
2. use fallback function for f16
* [GSoC2019]
Fix rebase issue
* Added MSA implementations for mips platforms. Intrinsics for MSA and build scripts for MIPS platforms are added.
Signed-off-by: Fei Wu <fwu@wavecomp.com>
* Removed some unused code in mips.toolchain.cmake.
Signed-off-by: Fei Wu <fwu@wavecomp.com>
* Added comments for mips toolchain configuration and disabled compiling warnings for libpng.
Signed-off-by: Fei Wu <fwu@wavecomp.com>
* Fixed the build error of unsupported opcode 'pause' when mips isa_rev is less than 2.
Signed-off-by: Fei Wu <fwu@wavecomp.com>
* 1. Removed FP16 related item in MSA option defines in OpenCVCompilerOptimizations.cmake.
2. Use CV_CPU_COMPILE_MSA instead of __mips_msa for MSA feature check in cv_cpu_dispatch.h.
3. Removed hasSIMD128() in intrin_msa.hpp.
4. Define CPU_MSA as 150.
Signed-off-by: Fei Wu <fwu@wavecomp.com>
* 1. Removed unnecessary CV_SIMD128_64F guarding in intrin_msa.hpp.
2. Removed unnecessary CV_MSA related code block in dotProd_8u().
Signed-off-by: Fei Wu <fwu@wavecomp.com>
* 1. Defined CPU_MSA_FLAGS_ON as "-mmsa".
2. Removed CV_SIMD128_64F guardings in intrin_msa.hpp.
Signed-off-by: Fei Wu <fwu@wavecomp.com>
* Removed unused msa_mlal_u16() and msa_mlal_s16 from msa_macros.h.
Signed-off-by: Fei Wu <fwu@wavecomp.com>
ISA 2.07 (aka POWER8) effectively extended the expanding multiply
operation to word types. The altivec intrinsics prior to gcc 8 did
not get the update.
Workaround this deficiency similar to other fixes.
This was exposed by commit 33fb253a66
which leverages the int -> dword expanding multiply.
This fixes Issue #15506
Use 4x FMA chains to sum on SIMD 128 FP64 targets. On
x86 this showed about 1.4x improvement.
For PPC, do a full multiply (32x32->64b), convert to DP
then accumulate. This may be slightly less precise for
some inputs. But is 1.5x faster than the above which
is about 1.5x than the FMA above for ~2.5x speedup.
Implement cvRound using inline asm. No compiler support
exists today to properly optimize this. This results in
about a 4x speedup over the default rounding. Likewise,
simplify the growing number of rounding function overloads.
For P9 enabled targets, utilize the classification
testing instruction to test for Inf/Nan values. Operation
speedup is about 1.2x for FP32, and 1.5x for FP64 operands.
For P8 targets, fallback to the GCC nan inline. It provides
a 1.1/1.4x improvement for FP32/FP64 arguments.
Add a new macro definition OPENCV_USE_FASTMATH_GCC_BUILTINS to enable
usage of GCC inline math functions, if available and requested by the
user.
Likewise, enable it for POWER. This is nearly always a substantial
improvement over using integer manipulation as most operations can
be done in several instructions with no branching. The result is a
1.5-1.8x speedup in the ceil/floor operations.
1. As tested with AT 12.0-1 (GCC 8.3.1) compiler on P9 LE.
Due to the explicitly declared copy constructor Vec<T, n>::Vec(Vec <T,n>&)
GCC 9 warns if there is no assignment operator, as having one typically
requires the other (rule-of-three, constructor/desctructor/assginment).
As the values are just a plain array the default assignment operator does
the right thing. Tell the compiler explicitly to default it.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
