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>
* core: improve AVX512 infrastructure by adding more CPU features groups
* cmake: use groups for AVX512 optimization flags
* core: remove gap in CPU flags enumeration
* cmake: restore default CPU_DISPATCH
OE-11 Logging revamp (#13909)
* Initial commit for log tag support.
Part of #11003, incomplete. Should pass build.
Moved LogLevel enum to logger.defines.hpp
LogTag struct used to convey both name and log level threshold as
one argument to the new logging macro. See logtag.hpp file, and
CV_LOG_WITH_TAG macro.
Global log level is now associated with a global log tag, when a
logging statement doesn't specify any log tag. See getLogLevel and
getGlobalLogTag functions.
A macro CV_LOGTAG_FALLBACK is allowed to be re-defined by other modules
or compilation units, internally, so that logging statements inside
that unit that specify NULL as tag will fall back to the re-defined tag.
Line-of-code information (file name, line number, function name),
together with tag name, are passed into the new log message sink.
See writeLogMessageEx function.
Fixed old incorrect CV_LOG_VERBOSE usage in ocl4dnn_conv_spatial.cpp.
* Implemented tag-based log filtering
Added LogTagManager. This is an initial version, using standard C++
approach as much as possible, to allow easier code review. Will
optimize later.
A workaround for all static dynamic initialization issues is
implemented. Refer to code comments.
* Added LogTagConfigParser.
Note: new code does not fully handle old log config parsing behavior.
* Fix log tag config vs registering ordering issue.
* Started testing LogTagConfigParser, incomplete.
The intention of this commit is to illustrate the capabilities of
the current design of LogTagConfigParser.
The test contained in this commit is not complete. Also, design changes
may require throwing away this commit and rewriting test code from
scratch.
Does not test whitespace segmentation (multiple tags on the config);
will do in next commit.
* Added CV_LOGTAG_EXPAND_NAME macro
This macro allows to be re-defined locally in other compilation units
to apply a prefix to whatever argument is passed as the "tag" argument
into CV_LOG_WITH_TAG. The default definition in logger.hpp does not
modify the argument. It is recommended to include the address-of
operator (ampersand) when re-defined locally.
* Added a few tests for LogTagManager, some fail.
See test_logtagmanager.cpp
Failed tests are: non-global ("something"), setting level by name-part
(first part or any part) has no effect at all.
* LogTagManagerTests substring non-confusion tests
* Fix major bugs in LogTagManager
The code change is intended to approximate the spec documented in
https://gist.github.com/kinchungwong/ec25bc1eba99142e0be4509b0f67d0c6
Refer to test suite in test_logtagmanager.cpp
Filter test result in "opencv_test_core" ...
with gtest_filter "LogTagManager*"
To see the test code that finds the bugs, refer to original commits
(before rebase; might be gone)
.. f3451208 (2019-03-03T19:45:17Z)
.... LogTagManagerTests substring non-confusion tests
.. 1b848f5f (2019-03-03T01:55:18Z)
.... Added a few tests for LogTagManager, some fail.
* Added LogTagManagerNamePartNonConfusionTest.
See test_logtagmanager.cpp in modules/core/test.
* Added LogTagAuto for auto registration in ctor
* Rewritten LogTagManager to resolve issues.
* Resolves code review issues around 2019-04-10
LogTagConfigParser::parseLogLevel - as part of resolving code review
issues, this function is rewritten to simplify control flow and to
improve conformance with legacy usage (for string values "OFF",
"DISABLED", and "WARNINGS").
- added functionality to collect memory usage of OpenCL sybsystem
- memory usage of fastMalloc() (disabled by default):
* It is not accurate sometimes - external memory profiler is required.
- specify common `CV_TEST_TAG_` macros
- added applyTestTag() function
- write memory usage / enabled tags into Google Tests output file (.xml)
- allow cmake to check sanity of vsx aligned ld/st
- force universal intrinsics v_load_aligned/v_store_aligned
to failback to unaligned ld/st if cmake runtime vsx aligned test fail
Lab/XYZ modes have been postponed (color_lab.cpp):
- need to split code for tables initialization and for pixels processing first
- no significant performance improvements for switching between SSE42 / AVX2 code generation
Resize reworked using wide universal intrinsics (#13781)
* Added wide universal intrinsics optimized implementation for 3 channel bit-exact linear resize
* Reworked linear resize using new wide LUT intrinsics
* Fix for VSX intrinsics
Due to size limit of shared memory, histogram is built on
the global memory for CV_16UC1 case.
The amount of memory needed for building histogram is:
65536 * 4byte = 256KB
and shared memory limit is 48KB typically.
Added test cases for CV_16UC1 and various clip limits.
Added perf tests for CV_16UC1 on both CPU and CUDA code.
There was also a bug in CV_8UC1 case when redistributing
"residual" clipped pixels. Adding the test case where clip
limit is 5.0 exposes this bug.
* Add Operator override for multi-channel Mat with literal constant.
* simple test
* Operator overloading channel constraint for primitive types
* fix some test for #13586
* added performance test for compareHist
* compareHist reworked to use wide universal intrinsics
* Disabled vectorization for CV_COMP_CORREL and CV_COMP_BHATTACHARYYA if f64 is unsupported
* Added performance tests for hal::norm functions
* Added sum of absolute differences intrinsic
* norm implementation updated to use wide universal intrinsics
* improve and fix v_reduce_sad on VSX
- add infrastructure support for Power9/VSX3
- fix missing VSX flags on GCC4.9 and CLANG4(#13210, #13222)
- fix disable VSX optimzation on GCC by using flag ENABLE_VSX
- flag ENABLE_VSX is deprecated now, use CPU_BASELINE, CPU_DISPATCH instead
- add VSX3 to arithmetic dispatchable flags
* Support for Matx read/write by FileStorage
* Only empty filestorage read now produces default Matx. Split Matx IO test into smaller units. Test checks for exception thrown if reading a Mat into a Matx of different size.
* Updated boxFilter implementations to use wide universal intrinsics
* boxFilter implementation moved to separate file
* Replaced ROUNDUP macro with roundUp() function
* integrated the new C++ persistence; removed old persistence; most of OpenCV compiles fine! the tests have not been run yet
* fixed multiple bugs in the new C++ persistence
* fixed raw size of the parsed empty sequences
* [temporarily] excluded obsolete applications traincascade and createsamples from build
* fixed several compiler warnings and multiple test failures
* undo changes in cocoa window rendering (that was fixed in another PR)
* fixed more compile warnings and the remaining test failures (hopefully)
* trying to fix the last little warning
- initialize arithmetic dispatcher
- add new universal intrinsic v_absdiffs
- add new universal intrinsic v_pack_b
- add accumulate version of universal intrinsic v_round
- fix sse/avx2:uint8 multiplication overflow
- reimplement arithmetic, logic and comparison operations into wide universal intrinsics
with full support for all types
- reimplement IPP arithmetic, logic and comparison operations in a sperate file arithm_ipp.hpp
- avoid scalar multiplication if scaling factor eq 1 and use integer multiplication
- move C arithmetic operations to precomp.hpp and delete [arithm_simd|arithm_core].hpp
- add compatibility with new opencv4 divide policy
* js: update build script
- support emscipten 1.38.12 (wasm is ON by default)
- verbose build messages
* js: use builtin Math functions
* js: disable tracing code completelly
* hide use of std::mutex from /clr compilation under Visual Studio
C++11 <mutex> is not available when compiling with /clr under Visual Studio, thus opencv cannot be easily included.
It is fixed by making a CEEMutex wrapper class, around an opaque implementation using std::mutex internally
* fixed compilation outside of Visual Studio
fixed compilation outside of Visual Studio by avoiding some macros
* fixed indentation, prepare getting rid of CEEMutex/CEELockGuard
fixed indentation
After discussion, CEEMutex and CEELockGuard can be totally removed, letting the developer in a /clr context to provide his own implementation
* remove CEEMutex/CEELockGuard
Fixes for instrumentation of IPP and OCL (#12637)
* fixed warning about re-declaring variable when both IPP and instrumentation are enabled
* fixed segfault when no funName provided
* compilation fixed when both OCL and instrumentation are enabled
- avoid updating of input image during equalizeHist() call
- avoid for() with double variable (use 'int' instead)
- more CV_Check*() macros
- use Mat_<T>, Matx
- static for local variables
* Remove isIntel check from deep learning layers
* Remove fp16->fp32 fallbacks where it's not necessary
* Fix Kernel::run to prevent localsize > globalsize
* added basic support for CV_16F (the new datatype etc.). CV_USRTYPE1 is now equal to CV_16F, which may break some [rarely used] functionality. We'll see
* fixed just introduced bug in norm; reverted errorneous changes in Torch importer (need to find a better solution)
* addressed some issues found during the PR review
* restored the patch to fix some perf test failures
* may be an typo fix
* remove identical branch,may be paste error
* add parentheses around macro parameter
* simplify if condition
* check malloc fail
* change the condition of branch removed by commit 3041502861
* rewrote Mat::convertTo() and convertScaleAbs() to wide universal intrinsics; added always-available and SIMD-optimized FP16<=>FP32 conversion
* fixed compile warnings
* fix some more compile errors
* slightly relaxed accuracy threshold for int->float conversion (since we now do it using single-precision arithmetics, not double-precision)
* fixed compile errors on iOS, Android and in the baseline C++ version (intrin_cpp.hpp)
* trying to fix ARM-neon builds
* trying to fix ARM-neon builds
* trying to fix ARM-neon builds
* trying to fix ARM-neon builds
* trying to fix the custom AVX2 builder test failures (false alarms)
* fixed compile error with CPU_BASELINE=AVX2 on x86; raised tolerance thresholds in a couple of tests
* fixed compile error with CPU_BASELINE=AVX2 on x86; raised tolerance thresholds in a couple of tests
* fixed compile error with CPU_BASELINE=AVX2 on x86; raised tolerance thresholds in a couple of tests
* seemingly disabled false alarm warning in surf.cpp; increased tolerance thresholds in the tests for SolvePnP and in DNN/ENet
Intrinsics must be effective, so don't declare FP16 type/operations if there is no native support.
- CV_FP16: supports load/store into/from float32
- CV_SIMD_FP16: declares FP16 types and native FP16 operations
for some big negative values less than -INT_MAX+32767 the sign of the numbers is lost due to overflow that leads to
incorrect saturation to MAX value, instead of zero.
The issue is not reproduced with CV_ENABLED_INTRINSICS=OFF
* 1. changed static const __m128/256 to const __m128/256 to avoid wierd instructions and calls inserted by compiler.
2. added universal intrinsics that wrap MOVNTPS and other such (non-temporary or "no cache" store) instructions. v_store_interleave() and v_store() got respective flags/overloaded variants
3. rewrote split & merge to use the "no cache" store instructions. It resulted in dramatic performance improvement when processing big arrays
* hopefully, fixed some test failures where 4-channel v_store_interleave() is used
* added missing implementation of the new universal intrinsics (v_store_aligned_nocache() etc.)
* fixed silly typo in the new intrinsics in intrin_vsx.hpp
* still trying to fix VSX compiler errors
* still trying to fix VSX compiler errors
* still trying to fix VSX compiler errors
* still trying to fix VSX compiler errors
* fixed/updated v_load_deinterleave and v_store_interleave intrinsics; modified split() and merge() functions to use those intrinsics
* fixed a few compile errors and bug in v_load_deinterleave(ptr, v_uint32x4& a, v_uint32x4& b)
* fixed few more compile errors
* core:OE-27 prepare universal intrinsics to expand (#11022)
* core:OE-27 prepare universal intrinsics to expand (#11022)
* core: Add universal intrinsics for AVX2
* updated implementation of wide univ. intrinsics; converted several OpenCV HAL functions: sqrt, invsqrt, magnitude, phase, exp to the wide universal intrinsics.
* converted log to universal intrinsics; cleaned up the code a bit; added v_lut_deinterleave intrinsics.
* core: Add universal intrinsics for AVX2
* fixed multiple compile errors
* fixed many more compile errors and hopefully some test failures
* fixed some more compile errors
* temporarily disabled IPP to debug exp & log; hopefully fixed Doxygen complains
* fixed some more compile errors
* fixed v_store(short*, v_float16&) signatures
* trying to fix the test failures on Linux
* fixed some issues found by alalek
* restored IPP optimization after the patch with AVX wide intrinsics has been properly tested
* restored IPP optimization after the patch with AVX wide intrinsics has been properly tested
- 'if' logic is moved into templates.
- removed unnecessary cv::Mat objects creation.
- fixed inv() test (invA * A == eye)
- added more Matx tests to cover all defined template specializations
* a part of https://github.com/opencv/opencv/pull/11364 by Tetragramm. Rewritten and extended findNonZero & PSNR to support more types, not just 8u.
* fixed compile & doxygen warnings
* fixed small bug in findNonZero test
fixes handling of empty matrices in some functions (#11634)
* a part of PR #11416 by Yuki Takehara
* moved the empty mat check in Mat::copyTo()
* fixed some test failures
* make sure that the matrix with more than INT_MAX elements is marked as non-continuous, and thus all the pixel-wise functions process it correctly (i.e. row-by-row, not as a single row, where integer overflow may occur when computing the total number of elements)
* Issue 11242 intrinsics v_extract, v_rotate improvement, branch 3.4, without C++11 (remove type restrictions for SSE2, use PALIGNR on SSSE3, compile to no-op when imm is 0 or nlanes).
* fix whitespace
* Fix#11242 (NEON intrinsics v_rotate...) branch 3.4
Separate macro expansion OPENCV_HAL_IMPL_NEON_SHIFT_OP for bitwise shifts for integers, from macro expansion OPENCV_HAL_IMPL_NEON_ROTATE for lane rotations. Bitwise shifts do not apply to floats, but lane-rotations can apply to both.
* fix whitespace
* Fix#11242 compile error (VSX intrinsics v_rotate(a)) branch 3.4 no-c++11
* Fix CV_Asserts with negation of strings
{!"string"} causes some compilers to throw a warning.
The value of the string is not that important -- it's only for printing
the assertion message.
Replace these calls with:
CV_Error(Error::StsError, "string")
to suppress the warning.
* remove unnecessary 'break' after CV_Error()
* use universal intrinsic instead of raw intrinsic
* add 2 channels de-interleave on x86 platform
* add v_int32x4 version of v_muladd
* add accumulate version of v_dotprod based on the commit from seiko2plus on bf1852d
* remove some verify check in performance test
* avoid the out of boundary access and keep the performance
* remove unnecessary defines from vsx_utils
* fix v_load_expand, load lower 64bit
* use vec_ld, vec_st with alignment load/store on all types except 64bit
* map v_extract to v_rotate_right
* update license header
* enable VSX by default on clang since #11167
To avoid compilation of this code:
- buf = 0;
This code can be received after refactoring of 1D cv::Mat to cv::AutoBuffer.
- "cv_mat = 0" calls setTo().
- cv::AutoBuffer calls "allocate(0)" - this is wrong.
* Make <array> #ifdef true for MSVC
I think MSVC had `std::array` for quite a while (possibly going back as far as VS 2012, but it's definitely there in 2015 and 2017. So I think `_MSC_VER` `1900` is a safe bet. Probably `1800` and maybe even `1700` could work as well but I can't test that locally.
* fix test
* Update BufferReader documentation with some example code
* Add warning to BufferPool doc regarding deallocation of StackAllocator
* Added a sample code that satisfies LIFO rule for StackAllocator
OpenCV pthreads-based implementation changes:
- rework worker threads pool, allow to execute job by the main thread too
- rework synchronization scheme (wait for job completion, threads 'pong' answer is not required)
- allow "active wait" (spin) by worker threads and by the main thread
- use _mm_pause() during active wait (support for Hyper-Threading technology)
- use sched_yield() to avoid preemption of still working other workers
- don't use getTickCount()
- optional builtin thread pool profiler (disabled by compilation flag)
UMatData locks are not mapped on real locks (they are mapped to some "pre-initialized" pool).
Concurrent execution of these statements may lead to deadlock:
- a.copyTo(b) from thread 1
- c.copyTo(d) from thread 2
where:
- 'a' and 'd' are mapped to single lock "A".
- 'b' and 'c' are mapped to single lock "B".
Workaround is to process locks with strict order.
The opencv infrastructure mostly has the basics for supporting avx512 math functions,
but it wasn't hooked up (likely due to lack of users)
In order to compile the DNN functions for AVX512, a few things need to be hooked up
and this patch does that
Signed-off-by: Arjan van de Ven <arjan@linux.intel.com>
- don't store ProgramSource in compiled Programs (resolved problem with "source" buffers lifetime)
- completelly remove Program::read/write methods implementation:
- replaced with method to query RAW OpenCL binary without any "custom" data
- deprecate Program::getPrefix() methods
If there are no OpenCL/UMat methods calls from application.
OpenCL subsystem is initialized:
- haveOpenCL() is called from application
- useOpenCL() is called from application
- access to OpenCL allocator: UMat is created (empty UMat is ignored) or UMat <-> Mat conversions are called
Don't call OpenCL functions if OPENCV_OPENCL_RUNTIME=disabled
(independent from OpenCL linkage type)
* add accuracy test and performance check for matmul
* add performance tests for transform and dotProduct
* add test Core_TransformLargeTest for 8u version of transform
* remove raw SSE2/NEON implementation from matmul.cpp
* use universal intrinsic instead of raw intrinsic
* remove unused templated function
* add v_matmuladd which multiply 3x3 matrix and add 3x1 vector
* add v_rotate_left/right in universal intrinsic
* suppress intrinsic on some function and platform
* add pure SW implementation of new universal intrinsics
* add test for new universal intrinsics
* core: prevent memory access after the end of buffer
* fix perf tests
When elements are 64 bits, the vec_st_interleave()/vec_ld_deinterleave()
doesn't interleave 4 elements correctly.
For vec_st_interleave(), following is saved into mem:
a0 b0 a1 b1 c0 d0 c1 d1
-> we expected:
a0 b0 c0 d0 a1 b1 c1 d1
for vec_ld_deinterleave(), following is loaded into a b c d for memory
string { 1 2 3 4 5 6 7 8 }:
a: 1 3
b: 2 4
c: 5 7
d: 6 8
-> we expected:
a: 1 5
b: 2 6
c: 3 7
d: 4 8
This patch corrects this behavior.
Signed-off-by: Simon Guo <wei.guo.simon@gmail.com>
- changed behavior of vec_ctf, vec_ctu, vec_cts
in gcc and clang to make them compatible with XLC
- implemented most of missing conversion intrinsics in gcc and clang
- implemented conversions intrinsics of odd-numbered elements
- ignored gcc bug warning that caused by -Wunused-but-set-variable in rare cases
- replaced right shift with algebraic right shift for signed vectors
to shift in the sign bit.
- added new universal intrinsics v_matmuladd, v_rotate_left/right
- avoid using floating multiply-add in RNG
* Update OpenCVCompilerOptimizations.cmake
Neon not supported on MSVC ARM breaking build fix
* Update OpenCVCompilerOptimizations.cmake
Whitespace
* Update intrin.hpp
Many problems in MSVC ARM builds (at least on VS2017) being fixed in this PR now.
C:\Users\Gregory\DOCUME~1\MYLIBR~1\OPENCV~3\opencv\sources\modules\core\include\opencv2/core/hal/intrin.hpp(444): error C3861: '_tzcnt_u32': identifier not found
* Update hal_replacement.hpp
Passing variadic expansion in a macro to another macro does not work properly in MSVC and a famous known workaround is hereby applied. Discussion of it: https://stackoverflow.com/questions/5134523/msvc-doesnt-expand-va-args-correctly
Only needed the fix for ARM builds: TEGRA_ macros are used for cv_hal_ functions in the carotene library.
C:\Users\Gregory\Documents\My Libraries\opencv330\opencv\sources\modules\core\src\arithm.cpp(2378): warning C4003: not enough actual parameters for macro 'TEGRA_ADD'
C:\Users\Gregory\Documents\My Libraries\opencv330\opencv\sources\modules\core\src\arithm.cpp(2378): error C2143: syntax error: missing ')' before ','
C:\Users\Gregory\Documents\My Libraries\opencv330\opencv\sources\modules\core\src\arithm.cpp(2378): error C2059: syntax error: ')'
* Update hal_replacement.hpp
All hal_replacement's using carotene\hal\tegra_hal.hpp TEGRA_ functions as macros preprocessed by variadic macros should be changed, identical as was done in core.
C:\Users\Gregory\Documents\My Libraries\opencv330\opencv\sources\modules\imgproc\src\color.cpp(9604): warning C4003: not enough actual parameters for macro 'TEGRA_CVTBGRTOBGR'
C:\Users\Gregory\Documents\My Libraries\opencv330\opencv\sources\modules\imgproc\src\color.cpp(9604): error C2059: syntax error: '=='
* Update OpenCVCompilerOptimizations.cmake
* Update hal_replacement.hpp
* Update hal_replacement.hpp
The original template based mat ptr for indexing is not implemented,
add the similar implementation as uchar type, but cast to
user-defined type from the uchar pointer.
Adds fitEllipseDirect to imgproc: The Direct least square (Direct) method by Fitzgibbon1999.
New Tests are included for the methods.
fitEllipseAMS Tests
fitEllipseDirect Tests
Comparative examples are added to fitEllipse.cpp in Samples.
add libdnn acceleration to dnn module (#9114)
* import libdnn code
Signed-off-by: Li Peng <peng.li@intel.com>
* add convolution layer ocl acceleration
Signed-off-by: Li Peng <peng.li@intel.com>
* add pooling layer ocl acceleration
Signed-off-by: Li Peng <peng.li@intel.com>
* add softmax layer ocl acceleration
Signed-off-by: Li Peng <peng.li@intel.com>
* add lrn layer ocl acceleration
Signed-off-by: Li Peng <peng.li@intel.com>
* add innerproduct layer ocl acceleration
Signed-off-by: Li Peng <peng.li@intel.com>
* add HAVE_OPENCL macro
Signed-off-by: Li Peng <peng.li@intel.com>
* fix for convolution ocl
Signed-off-by: Li Peng <peng.li@intel.com>
* enable getUMat() for multi-dimension Mat
Signed-off-by: Li Peng <peng.li@intel.com>
* use getUMat for ocl acceleration
Signed-off-by: Li Peng <peng.li@intel.com>
* use CV_OCL_RUN macro
Signed-off-by: Li Peng <peng.li@intel.com>
* set OPENCL target when it is available
and disable fuseLayer for OCL target for the time being
Signed-off-by: Li Peng <peng.li@intel.com>
* fix innerproduct accuracy test
Signed-off-by: Li Peng <peng.li@intel.com>
* remove trailing space
Signed-off-by: Li Peng <peng.li@intel.com>
* Fixed tensorflow demo bug.
Root cause is that tensorflow has different algorithm with libdnn
to calculate convolution output dimension.
libdnn don't calculate output dimension anymore and just use one
passed in by config.
* split gemm ocl file
split it into gemm_buffer.cl and gemm_image.cl
Signed-off-by: Li Peng <peng.li@intel.com>
* Fix compile failure
Signed-off-by: Li Peng <peng.li@intel.com>
* check env flag for auto tuning
Signed-off-by: Li Peng <peng.li@intel.com>
* switch to new ocl kernels for softmax layer
Signed-off-by: Li Peng <peng.li@intel.com>
* update softmax layer
on some platform subgroup extension may not work well,
fallback to non subgroup ocl acceleration.
Signed-off-by: Li Peng <peng.li@intel.com>
* fallback to cpu path for fc layer with multi output
Signed-off-by: Li Peng <peng.li@intel.com>
* update output message
Signed-off-by: Li Peng <peng.li@intel.com>
* update fully connected layer
fallback to gemm API if libdnn return false
Signed-off-by: Li Peng <peng.li@intel.com>
* Add ReLU OCL implementation
* disable layer fusion for now
Signed-off-by: Li Peng <peng.li@intel.com>
* Add OCL implementation for concat layer
Signed-off-by: Wu Zhiwen <zhiwen.wu@intel.com>
* libdnn: update license and copyrights
Also refine libdnn coding style
Signed-off-by: Wu Zhiwen <zhiwen.wu@intel.com>
Signed-off-by: Li Peng <peng.li@intel.com>
* DNN: Don't link OpenCL library explicitly
* DNN: Make default preferableTarget to DNN_TARGET_CPU
User should set it to DNN_TARGET_OPENCL explicitly if want to
use OpenCL acceleration.
Also don't fusion when using DNN_TARGET_OPENCL
* DNN: refine coding style
* Add getOpenCLErrorString
* DNN: Use int32_t/uint32_t instread of alias
* Use namespace ocl4dnn to include libdnn things
* remove extra copyTo in softmax ocl path
Signed-off-by: Li Peng <peng.li@intel.com>
* update ReLU layer ocl path
Signed-off-by: Li Peng <peng.li@intel.com>
* Add prefer target property for layer class
It is used to indicate the target for layer forwarding,
either the default CPU target or OCL target.
Signed-off-by: Li Peng <peng.li@intel.com>
* Add cl_event based timer for cv::ocl
* Rename libdnn to ocl4dnn
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: wzw <zhiwen.wu@intel.com>
* use UMat for ocl4dnn internal buffer
Remove allocateMemory which use clCreateBuffer directly
Signed-off-by: Li Peng <peng.li@intel.com>
Signed-off-by: wzw <zhiwen.wu@intel.com>
* enable buffer gemm in ocl4dnn innerproduct
Signed-off-by: Li Peng <peng.li@intel.com>
* replace int_tp globally for ocl4dnn kernels.
Signed-off-by: wzw <zhiwen.wu@intel.com>
Signed-off-by: Li Peng <peng.li@intel.com>
* create UMat for layer params
Signed-off-by: Li Peng <peng.li@intel.com>
* update sign ocl kernel
Signed-off-by: Li Peng <peng.li@intel.com>
* update image based gemm of inner product layer
Signed-off-by: Li Peng <peng.li@intel.com>
* remove buffer gemm of inner product layer
call cv::gemm API instead
Signed-off-by: Li Peng <peng.li@intel.com>
* change ocl4dnn forward parameter to UMat
Signed-off-by: Li Peng <peng.li@intel.com>
* Refine auto-tuning mechanism.
- Use OPENCV_OCL4DNN_KERNEL_CONFIG_PATH to set cache directory
for fine-tuned kernel configuration.
e.g. export OPENCV_OCL4DNN_KERNEL_CONFIG_PATH=/home/tmp,
the cache directory will be /home/tmp/spatialkernels/ on Linux.
- Define environment OPENCV_OCL4DNN_ENABLE_AUTO_TUNING to enable
auto-tuning.
- OPENCV_OPENCL_ENABLE_PROFILING is only used to enable profiling
for OpenCL command queue. This fix basic kernel get wrong running
time, i.e. 0ms.
- If creating cache directory failed, disable auto-tuning.
* Detect and create cache dir on windows
Signed-off-by: Li Peng <peng.li@intel.com>
* Refine gemm like convolution kernel.
Signed-off-by: Li Peng <peng.li@intel.com>
* Fix redundant swizzleWeights calling when use cached kernel config.
* Fix "out of resource" bug when auto-tuning too many kernels.
* replace cl_mem with UMat in ocl4dnnConvSpatial class
* OCL4DNN: reduce the tuning kernel candidate.
This patch could reduce 75% of the tuning candidates with less
than 2% performance impact for the final result.
Signed-off-by: Zhigang Gong <zhigang.gong@intel.com>
* replace cl_mem with umat in ocl4dnn convolution
Signed-off-by: Li Peng <peng.li@intel.com>
* remove weight_image_ of ocl4dnn inner product
Actually it is unused in the computation
Signed-off-by: Li Peng <peng.li@intel.com>
* Various fixes for ocl4dnn
1. OCL_PERFORMANCE_CHECK(ocl::Device::getDefault().isIntel())
2. Ptr<OCL4DNNInnerProduct<float> > innerProductOp
3. Code comments cleanup
4. ignore check on OCL cpu device
Signed-off-by: Li Peng <peng.li@intel.com>
* add build option for log softmax
Signed-off-by: Li Peng <peng.li@intel.com>
* remove unused ocl kernels in ocl4dnn
Signed-off-by: Li Peng <peng.li@intel.com>
* replace ocl4dnnSet with opencv setTo
Signed-off-by: Li Peng <peng.li@intel.com>
* replace ALIGN with cv::alignSize
Signed-off-by: Li Peng <peng.li@intel.com>
* check kernel build options
Signed-off-by: Li Peng <peng.li@intel.com>
* Handle program compilation fail properly.
* Use std::numeric_limits<float>::infinity() for large float number
* check ocl4dnn kernel compilation result
Signed-off-by: Li Peng <peng.li@intel.com>
* remove unused ctx_id
Signed-off-by: Li Peng <peng.li@intel.com>
* change clEnqueueNDRangeKernel to kernel.run()
Signed-off-by: Li Peng <peng.li@intel.com>
* change cl_mem to UMat in image based gemm
Signed-off-by: Li Peng <peng.li@intel.com>
* check intel subgroup support for lrn and pooling layer
Signed-off-by: Li Peng <peng.li@intel.com>
* Fix convolution bug if group is greater than 1
Signed-off-by: Li Peng <peng.li@intel.com>
* Set default layer preferableTarget to be DNN_TARGET_CPU
Signed-off-by: Li Peng <peng.li@intel.com>
* Add ocl perf test for convolution
Signed-off-by: Li Peng <peng.li@intel.com>
* Add more ocl accuracy test
Signed-off-by: Li Peng <peng.li@intel.com>
* replace cl_image with ocl::Image2D
Signed-off-by: Li Peng <peng.li@intel.com>
* Fix build failure in elementwise layer
Signed-off-by: Li Peng <peng.li@intel.com>
* use getUMat() to get blob data
Signed-off-by: Li Peng <peng.li@intel.com>
* replace cl_mem handle with ocl::KernelArg
Signed-off-by: Li Peng <peng.li@intel.com>
* dnn(build): don't use C++11, OPENCL_LIBRARIES fix
* dnn(ocl4dnn): remove unused OpenCL kernels
* dnn(ocl4dnn): extract OpenCL code into .cl files
* dnn(ocl4dnn): refine auto-tuning
Defaultly disable auto-tuning, set OPENCV_OCL4DNN_ENABLE_AUTO_TUNING
environment variable to enable it.
Use a set of pre-tuned configs as default config if auto-tuning is disabled.
These configs are tuned for Intel GPU with 48/72 EUs, and for googlenet,
AlexNet, ResNet-50
If default config is not suitable, use the first available kernel config
from the candidates. Candidate priority from high to low is gemm like kernel,
IDLF kernel, basick kernel.
* dnn(ocl4dnn): pooling doesn't use OpenCL subgroups
* dnn(ocl4dnn): fix perf test
OpenCV has default 3sec time limit for each performance test.
Warmup OpenCL backend outside of perf measurement loop.
* use ocl::KernelArg as much as possible
Signed-off-by: Li Peng <peng.li@intel.com>
* dnn(ocl4dnn): fix bias bug for gemm like kernel
* dnn(ocl4dnn): wrap cl_mem into UMat
Signed-off-by: Li Peng <peng.li@intel.com>
* dnn(ocl4dnn): Refine signature of kernel config
- Use more readable string as signture of kernel config
- Don't count device name and vendor in signature string
- Default kernel configurations are tuned for Intel GPU with
24/48/72 EUs, and for googlenet, AlexNet, ResNet-50 net model.
* dnn(ocl4dnn): swap width/height in configuration
* dnn(ocl4dnn): enable configs for Intel OpenCL runtime only
* core: make configuration helper functions accessible from non-core modules
* dnn(ocl4dnn): update kernel auto-tuning behavior
Avoid unwanted creation of directories
* dnn(ocl4dnn): simplify kernel to workaround OpenCL compiler crash
* dnn(ocl4dnn): remove redundant code
* dnn(ocl4dnn): Add more clear message for simd size dismatch.
* dnn(ocl4dnn): add const to const argument
Signed-off-by: Li Peng <peng.li@intel.com>
* dnn(ocl4dnn): force compiler use a specific SIMD size for IDLF kernel
* dnn(ocl4dnn): drop unused tuneLocalSize()
* dnn(ocl4dnn): specify OpenCL queue for Timer and convolve() method
* dnn(ocl4dnn): sanitize file names used for cache
* dnn(perf): enable Network tests with OpenCL
* dnn(ocl4dnn/conv): drop computeGlobalSize()
* dnn(ocl4dnn/conv): drop unused fields
* dnn(ocl4dnn/conv): simplify ctor
* dnn(ocl4dnn/conv): refactor kernelConfig localSize=NULL
* dnn(ocl4dnn/conv): drop unsupported double / untested half types
* dnn(ocl4dnn/conv): drop unused variable
* dnn(ocl4dnn/conv): alignSize/divUp
* dnn(ocl4dnn/conv): use enum values
* dnn(ocl4dnn): drop unused innerproduct variable
Signed-off-by: Li Peng <peng.li@intel.com>
* dnn(ocl4dnn): add an generic function to check cl option support
* dnn(ocl4dnn): run softmax subgroup version kernel first
Signed-off-by: Li Peng <peng.li@intel.com>
Added forkfour Latex command to math js support.
Split cv::norm documentation between the cv::norm and its overload, to make things clearer
Corrected some typos and cleaned up grammar.
Result is clearer documentation for the norms.
Work pending...
This adds the possibility to use multi-channel masks for the functions
cv::mean, cv::meanStdDev and the method Mat::setTo. The tests have now a
probability to use multi-channel masks for operations that support them.
This also includes Mat::copyTo, which supported multi-channel masks
before, but there was no test confirming this.
This function is the counterpart of "Context::getProg".
With this function, users have chance to unload a program
from global run-time cached programs, and save resource.
- Optimizations set change. Now IPP integrations will provide code for SSE42, AVX2 and AVX512 (SKX) CPUs only. For HW below SSE42 IPP code is disabled.
- Performance regressions fixes for IPP code paths;
- cv::boxFilter integration improvement;
- cv::filter2D integration improvement;
Add constructors taking initializer_list for some of OpenCV data types (#9034)
* Add a constructor taking initializer_list for Matx
* Add a constructor taking initializer list for Mat and Mat_
* Add one more method to initialize Mat to the corresponding tutorial
* Add a note how to initialize Matx
* CV_CXX_11->CV_CXX11
BufferPoolController has a non virtual protected destructor (which is legitimate)
However, Visual Studio sees this as a bug, if you enable more warnings, like below
```
add_compile_options(/W3) # level 3 warnings
add_compile_options(/we4265) # warning about missing virtual destructors
```
This is a proposition in order to silence this warning.
See https://github.com/ivsgroup/boost_warnings_minimal_demo for a demo of the same problem
with boost/exception.hpp
Remove unnecessary Non-ASCII characters from source code (#9075)
* Remove unnecessary Non-ASCII characters from source code
Remove unnecessary Non-ASCII characters and replace them with ASCII
characters
* Remove dashes in the @param statement
Remove dashes and place single space in the @param statement to keep
coding style
* misc: more fixes for non-ASCII symbols
* misc: fix non-ASCII symbol in CMake file
* another round of dnn optimization:
* increased malloc alignment across OpenCV from 16 to 64 bytes to make it AVX2 and even AVX-512 friendly
* improved SIMD optimization of pooling layer, optimized average pooling
* cleaned up convolution layer implementation
* made activation layer "attacheable" to all other layers, including fully connected and addition layer.
* fixed bug in the fusion algorithm: "LayerData::consumers" should not be cleared, because it desctibes the topology.
* greatly optimized permutation layer, which improved SSD performance
* parallelized element-wise binary/ternary/... ops (sum, prod, max)
* also, added missing copyrights to many of the layer implementation files
* temporarily disabled (again) the check for intermediate blobs consistency; fixed warnings from various builders
