Rewrite Universal Intrinsic code: gapi module (fluid part). by hanliutong · Pull Request #24324 · opencv/opencv

hanliutong · 2023-09-27T05:53:20Z

The goal of this series of PRs is to modify the SIMD code blocks guarded by CV_SIMD macro: rewrite them by using the new Universal Intrinsic API.

This is the modification to the gapi module, especially the fluid part.

All modifications to the gapi module have been completed, but this PR is marked as draft because many use cases failed in the test, and I am looking for the reason.

FAILED 76 tests on RVV QEMU, listed below:


[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/2, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x4e534a, DIV, true, 1, false)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/3, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x4e534a, DIV, true, 1, true)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/6, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x4e534a, MUL, true, 1, false)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/7, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x4e534a, MUL, true, 1, true) 
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/10, where GetParam() = (8UC3, 1280x720, 8UC1, 0x4e534a, DIV, true, 1, false)    
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/11, where GetParam() = (8UC3, 1280x720, 8UC1, 0x4e534a, DIV, true, 1, true)     
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/14, where GetParam() = (8UC3, 1280x720, 8UC1, 0x4e534a, MUL, true, 1, false)    
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/15, where GetParam() = (8UC3, 1280x720, 8UC1, 0x4e534a, MUL, true, 1, true)     
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/18, where GetParam() = (8UC3, 1280x720, 32FC1, 0x4e534a, DIV, true, 1, false)   
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/19, where GetParam() = (8UC3, 1280x720, 32FC1, 0x4e534a, DIV, true, 1, true)    
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/22, where GetParam() = (8UC3, 1280x720, 32FC1, 0x4e534a, MUL, true, 1, false)   
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/23, where GetParam() = (8UC3, 1280x720, 32FC1, 0x4e534a, MUL, true, 1, true)    
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/26, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x4e534a, DIV, true, 1, false)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/27, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x4e534a, DIV, true, 1, true) 
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/30, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x4e534a, MUL, true, 1, false)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/31, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x4e534a, MUL, true, 1, true) 
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/34, where GetParam() = (8UC3, 128x128, 8UC1, 0x4e534a, DIV, true, 1, false)     
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/35, where GetParam() = (8UC3, 128x128, 8UC1, 0x4e534a, DIV, true, 1, true)      
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/38, where GetParam() = (8UC3, 128x128, 8UC1, 0x4e534a, MUL, true, 1, false)     
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/39, where GetParam() = (8UC3, 128x128, 8UC1, 0x4e534a, MUL, true, 1, true)      
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/42, where GetParam() = (8UC3, 128x128, 32FC1, 0x4e534a, DIV, true, 1, false)    
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/43, where GetParam() = (8UC3, 128x128, 32FC1, 0x4e534a, DIV, true, 1, true)     
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/46, where GetParam() = (8UC3, 128x128, 32FC1, 0x4e534a, MUL, true, 1, false)    
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/47, where GetParam() = (8UC3, 128x128, 32FC1, 0x4e534a, MUL, true, 1, true)     
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/50, where GetParam() = (8UC1, 1280x720, SAME_TYPE, 0x4e534a, DIV, true, 1, false)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/74, where GetParam() = (8UC1, 128x128, SAME_TYPE, 0x4e534a, DIV, true, 1, false)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/82, where GetParam() = (8UC1, 128x128, 8UC1, 0x4e534a, DIV, true, 1, false)     
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/123, where GetParam() = (16SC1, 128x128, SAME_TYPE, 0x4e534a, DIV, true, 1, true)
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/37, where GetParam() = (32FC1, 1280x720, 8UC1, 0x4e8af6, CMP_GE, false, AbsExact())       
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/38, where GetParam() = (32FC1, 1280x720, 8UC1, 0x4e8af6, CMP_NE, false, AbsExact())       
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/39, where GetParam() = (32FC1, 1280x720, 8UC1, 0x4e8af6, CMP_GT, false, AbsExact())       
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/40, where GetParam() = (32FC1, 1280x720, 8UC1, 0x4e8af6, CMP_LT, false, AbsExact())       
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/41, where GetParam() = (32FC1, 1280x720, 8UC1, 0x4e8af6, CMP_LE, false, AbsExact())       
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/43, where GetParam() = (32FC1, 128x128, 8UC1, 0x4e8af6, CMP_GE, false, AbsExact())        
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/44, where GetParam() = (32FC1, 128x128, 8UC1, 0x4e8af6, CMP_NE, false, AbsExact())        
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/45, where GetParam() = (32FC1, 128x128, 8UC1, 0x4e8af6, CMP_GT, false, AbsExact())        
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/46, where GetParam() = (32FC1, 128x128, 8UC1, 0x4e8af6, CMP_LT, false, AbsExact())        
[  FAILED  ] CompareTestFluid/CmpTest.AccuracyTest/47, where GetParam() = (32FC1, 128x128, 8UC1, 0x4e8af6, CMP_LE, false, AbsExact())        
[  FAILED  ] CompareTestFluidScalar/CmpTest.AccuracyTest/38, where GetParam() = (32FC1, 1280x720, 8UC1, 0x4e9126, CMP_NE, true, AbsSimilarPoints(1, 0.01))
[  FAILED  ] CompareTestFluidScalar/CmpTest.AccuracyTest/39, where GetParam() = (32FC1, 1280x720, 8UC1, 0x4e9126, CMP_GT, true, AbsSimilarPoints(1, 0.01))
[  FAILED  ] CompareTestFluidScalar/CmpTest.AccuracyTest/43, where GetParam() = (32FC1, 128x128, 8UC1, 0x4e9126, CMP_GE, true, AbsSimilarPoints(1, 0.01))
[  FAILED  ] CompareTestFluidScalar/CmpTest.AccuracyTest/44, where GetParam() = (32FC1, 128x128, 8UC1, 0x4e9126, CMP_NE, true, AbsSimilarPoints(1, 0.01))
[  FAILED  ] CompareTestFluidScalar/CmpTest.AccuracyTest/45, where GetParam() = (32FC1, 128x128, 8UC1, 0x4e9126, CMP_GT, true, AbsSimilarPoints(1, 0.01))
[  FAILED  ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/12, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x4e733e)
[  FAILED  ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/13, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x4e733e)[  FAILED  ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/14, where GetParam() = (16UC3, 1280x720, SAME_TYPE, 0x4e733e)
[  FAILED  ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/15, where GetParam() = (16UC3, 128x128, SAME_TYPE, 0x4e733e)
[  FAILED  ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/16, where GetParam() = (16SC3, 1280x720, SAME_TYPE, 0x4e733e)
[  FAILED  ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/17, where GetParam() = (16SC3, 128x128, SAME_TYPE, 0x4e733e)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/3, where GetParam() = (8UC1, 1280x720, SAME_TYPE, 0x59fa46, AbsExact(), DIV)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/7, where GetParam() = (8UC1, 1280x720, SAME_TYPE, 0x59fa46, AbsExact(), DIVR)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/11, where GetParam() = (8UC1, 128x128, SAME_TYPE, 0x59fa46, AbsExact(), DIV)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/15, where GetParam() = (8UC1, 128x128, SAME_TYPE, 0x59fa46, AbsExact(), DIVR)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/23, where GetParam() = (16SC1, 1280x720, SAME_TYPE, 0x59fa46, AbsExact(), DIVR)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/48, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), GT)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/50, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), GE)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/53, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), NE)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/55, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), LTR)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/57, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), LER)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/59, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), NER)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/60, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), GT)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/62, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), GE)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/65, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), NE)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/67, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), LTR)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/69, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), LER)
[  FAILED  ] MathOperatorCompareTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/71, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x5a0188, AbsSimilarPoints(1, 0.01), NER)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/39, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x59f3dc, AbsExact(), GT)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/40, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x59f3dc, AbsExact(), LT)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/41, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x59f3dc, AbsExact(), GE)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/42, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x59f3dc, AbsExact(), LE)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/44, where GetParam() = (32FC1, 1280x720, SAME_TYPE, 0x59f3dc, AbsExact(), NE)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/48, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x59f3dc, AbsExact(), GT)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/49, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x59f3dc, AbsExact(), LT)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/50, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x59f3dc, AbsExact(), GE)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/51, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x59f3dc, AbsExact(), LE)
[  FAILED  ] MathOperatorTestFluid/MathOperatorMatMatTest.OperatorAccuracyTest/53, where GetParam() = (32FC1, 128x128, SAME_TYPE, 0x59f3dc, AbsExact(), NE)

FAILED 36 tests on RVV QEMU without this patch, listed below:

[ FAILED ] 36 tests, listed below:
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/2, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x3fb1d0, DIV, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/3, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x3fb1d0, DIV, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/6, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x3fb1d0, MUL, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/7, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x3fb1d0, MUL, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/10, where GetParam() = (8UC3, 1280x720, 8UC1, 0x3fb1d0, DIV, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/11, where GetParam() = (8UC3, 1280x720, 8UC1, 0x3fb1d0, DIV, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/14, where GetParam() = (8UC3, 1280x720, 8UC1, 0x3fb1d0, MUL, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/15, where GetParam() = (8UC3, 1280x720, 8UC1, 0x3fb1d0, MUL, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/18, where GetParam() = (8UC3, 1280x720, 32FC1, 0x3fb1d0, DIV, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/19, where GetParam() = (8UC3, 1280x720, 32FC1, 0x3fb1d0, DIV, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/22, where GetParam() = (8UC3, 1280x720, 32FC1, 0x3fb1d0, MUL, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/23, where GetParam() = (8UC3, 1280x720, 32FC1, 0x3fb1d0, MUL, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/26, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x3fb1d0, DIV, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/27, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x3fb1d0, DIV, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/30, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x3fb1d0, MUL, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/31, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x3fb1d0, MUL, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/34, where GetParam() = (8UC3, 128x128, 8UC1, 0x3fb1d0, DIV, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/35, where GetParam() = (8UC3, 128x128, 8UC1, 0x3fb1d0, DIV, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/38, where GetParam() = (8UC3, 128x128, 8UC1, 0x3fb1d0, MUL, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/39, where GetParam() = (8UC3, 128x128, 8UC1, 0x3fb1d0, MUL, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/42, where GetParam() = (8UC3, 128x128, 32FC1, 0x3fb1d0, DIV, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/43, where GetParam() = (8UC3, 128x128, 32FC1, 0x3fb1d0, DIV, true, 1, true)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/46, where GetParam() = (8UC3, 128x128, 32FC1, 0x3fb1d0, MUL, true, 1, false)
[ FAILED ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/47, where GetParam() = (8UC3, 128x128, 32FC1, 0x3fb1d0, MUL, true, 1, true)
[ FAILED ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/12, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x3fb504)
[ FAILED ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/13, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x3fb504)
[ FAILED ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/14, where GetParam() = (16UC3, 1280x720, SAME_TYPE, 0x3fb504)
[ FAILED ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/15, where GetParam() = (16UC3, 128x128, SAME_TYPE, 0x3fb504)
[ FAILED ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/16, where GetParam() = (16SC3, 1280x720, SAME_TYPE, 0x3fb504)
[ FAILED ] AbsDiffCTestFluid/AbsDiffCTest.AccuracyTest/17, where GetParam() = (16SC3, 128x128, SAME_TYPE, 0x3fb504)
[ FAILED ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/3, where GetParam() = (8UC1, 1280x720, SAME_TYPE, 0x48d782, AbsExact(), DIV)
[ FAILED ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/7, where GetParam() = (8UC1, 1280x720, SAME_TYPE, 0x48d782, AbsExact(), DIVR)
[ FAILED ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/11, where GetParam() = (8UC1, 128x128, SAME_TYPE, 0x48d782, AbsExact(), DIV)
[ FAILED ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/15, where GetParam() = (8UC1, 128x128, SAME_TYPE, 0x48d782, AbsExact(), DIVR)
[ FAILED ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/23, where GetParam() = (16SC1, 1280x720, SAME_TYPE, 0x48d782, AbsExact(), DIVR)
[ FAILED ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/31, where GetParam() = (16SC1, 128x128, SAME_TYPE, 0x48d782, AbsExact(), DIVR)

Pull Request Readiness Checklist

See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request

I agree to contribute to the project under Apache 2 License.
To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV
The PR is proposed to the proper branch
There is a reference to the original bug report and related work
There is accuracy test, performance test and test data in opencv_extra repository, if applicable
Patch to opencv_extra has the same branch name.
The feature is well documented and sample code can be built with the project CMake

hanliutong · 2023-10-03T14:08:06Z

After my various tests, the unit test case failures seem to be due to specific compiler and simulator versions and not related to code modifications🤔

In the following test environment, this patch did not introduce new failure test cases.

Test Environment:

clang version 16.0.6 7cbf1a2591520c2491aa35339f227775f4d3adf6
qemu-riscv64 version 8.0.3 (SCR-g066fc682) (from CI docker file)

And then, the following test failed both with and without this patch:

[==========] 15472 tests from 500 test cases ran. (4707378 ms total)
[  PASSED  ] 15462 tests.
[  FAILED  ] 10 tests, listed below:
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/2, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x3f8e34, DIV, true, 1, false)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/3, where GetParam() = (8UC3, 1280x720, SAME_TYPE, 0x3f8e34, DIV, true, 1, true)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/10, where GetParam() = (8UC3, 1280x720, 8UC1, 0x3f8e34, DIV, true, 1, false)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/27, where GetParam() = (8UC3, 128x128, SAME_TYPE, 0x3f8e34, DIV, true, 1, true)
[  FAILED  ] MathOpTestFluid/MathOpTest.MatricesAccuracyTest/35, where GetParam() = (8UC3, 128x128, 8UC1, 0x3f8e34, DIV, true, 1, true)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/3, where GetParam() = (8UC1, 1280x720, SAME_TYPE, 0x484fac, AbsExact(), DIV)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/7, where GetParam() = (8UC1, 1280x720, SAME_TYPE, 0x484fac, AbsExact(), DIVR)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/11, where GetParam() = (8UC1, 128x128, SAME_TYPE, 0x484fac, AbsExact(), DIV)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/15, where GetParam() = (8UC1, 128x128, SAME_TYPE, 0x484fac, AbsExact(), DIVR)
[  FAILED  ] MathOperatorArithmeticTestFluid/MathOperatorMatScalarTest.OperatorAccuracyTest/31, where GetParam() = (16SC1, 128x128, SAME_TYPE, 0x484fac, AbsExact(), DIVR)

10 FAILED TESTS

asmorkalov · 2023-10-04T08:53:29Z

Most probably, it's the same as in #19118 and #20413

asmorkalov · 2023-10-04T08:56:57Z

@TolyaTalamanov @dmatveev please review the PR.

Rewrite Universal Intrinsic code: float related part #24325 The goal of this series of PRs is to modify the SIMD code blocks guarded by CV_SIMD macro: rewrite them by using the new Universal Intrinsic API. The series of PRs is listed below: #23885 First patch, an example #23980 Core module #24058 ImgProc module, part 1 #24132 ImgProc module, part 2 #24166 ImgProc module, part 3 #24301 Features2d and calib3d module #24324 Gapi module This patch (hopefully) is the last one in the series. This patch mainly involves 3 parts 1. Add some modifications related to float (CV_SIMD_64F) 2. Use `#if (CV_SIMD || CV_SIMD_SCALABLE)` instead of `#if CV_SIMD || CV_SIMD_SCALABLE`, then we can get the `CV_SIMD` module that is not enabled for `CV_SIMD_SCALABLE` by looking for `if CV_SIMD` 3. Summary of `CV_SIMD` blocks that remains unmodified: Updated comments - Some blocks will cause test fail when enable for RVV, marked as `TODO: enable for CV_SIMD_SCALABLE, ....` - Some blocks can not be rewrited directly. (Not commented in the source code, just listed here) - ./modules/core/src/mathfuncs_core.simd.hpp (Vector type wrapped in class/struct) - ./modules/imgproc/src/color_lab.cpp (Array of vector type) - ./modules/imgproc/src/color_rgb.simd.hpp (Array of vector type) - ./modules/imgproc/src/sumpixels.simd.hpp (fixed length algorithm, strongly ralated with `CV_SIMD_WIDTH`) These algorithms will need to be redesigned to accommodate scalable backends. ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [ ] I agree to contribute to the project under Apache 2 License. - [ ] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [ ] The PR is proposed to the proper branch - [ ] There is a reference to the original bug report and related work - [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [ ] The feature is well documented and sample code can be built with the project CMake

Rewrite Universal Intrinsic code: float related part opencv#24325 The goal of this series of PRs is to modify the SIMD code blocks guarded by CV_SIMD macro: rewrite them by using the new Universal Intrinsic API. The series of PRs is listed below: opencv#23885 First patch, an example opencv#23980 Core module opencv#24058 ImgProc module, part 1 opencv#24132 ImgProc module, part 2 opencv#24166 ImgProc module, part 3 opencv#24301 Features2d and calib3d module opencv#24324 Gapi module This patch (hopefully) is the last one in the series. This patch mainly involves 3 parts 1. Add some modifications related to float (CV_SIMD_64F) 2. Use `#if (CV_SIMD || CV_SIMD_SCALABLE)` instead of `#if CV_SIMD || CV_SIMD_SCALABLE`, then we can get the `CV_SIMD` module that is not enabled for `CV_SIMD_SCALABLE` by looking for `if CV_SIMD` 3. Summary of `CV_SIMD` blocks that remains unmodified: Updated comments - Some blocks will cause test fail when enable for RVV, marked as `TODO: enable for CV_SIMD_SCALABLE, ....` - Some blocks can not be rewrited directly. (Not commented in the source code, just listed here) - ./modules/core/src/mathfuncs_core.simd.hpp (Vector type wrapped in class/struct) - ./modules/imgproc/src/color_lab.cpp (Array of vector type) - ./modules/imgproc/src/color_rgb.simd.hpp (Array of vector type) - ./modules/imgproc/src/sumpixels.simd.hpp (fixed length algorithm, strongly ralated with `CV_SIMD_WIDTH`) These algorithms will need to be redesigned to accommodate scalable backends. ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [ ] I agree to contribute to the project under Apache 2 License. - [ ] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [ ] The PR is proposed to the proper branch - [ ] There is a reference to the original bug report and related work - [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [ ] The feature is well documented and sample code can be built with the project CMake

asmorkalov · 2023-10-12T08:24:49Z

@TolyaTalamanov @dmatveev Friendly reminder.

asmorkalov

:+

dmatveev · 2023-10-13T19:59:09Z

Oh wow, this is big! If tests passed in the CI, it should be ok. :) Thanks for the contribution!

Rewrite Universal Intrinsic code: float related part opencv#24325 The goal of this series of PRs is to modify the SIMD code blocks guarded by CV_SIMD macro: rewrite them by using the new Universal Intrinsic API. The series of PRs is listed below: opencv#23885 First patch, an example opencv#23980 Core module opencv#24058 ImgProc module, part 1 opencv#24132 ImgProc module, part 2 opencv#24166 ImgProc module, part 3 opencv#24301 Features2d and calib3d module opencv#24324 Gapi module This patch (hopefully) is the last one in the series. This patch mainly involves 3 parts 1. Add some modifications related to float (CV_SIMD_64F) 2. Use `#if (CV_SIMD || CV_SIMD_SCALABLE)` instead of `#if CV_SIMD || CV_SIMD_SCALABLE`, then we can get the `CV_SIMD` module that is not enabled for `CV_SIMD_SCALABLE` by looking for `if CV_SIMD` 3. Summary of `CV_SIMD` blocks that remains unmodified: Updated comments - Some blocks will cause test fail when enable for RVV, marked as `TODO: enable for CV_SIMD_SCALABLE, ....` - Some blocks can not be rewrited directly. (Not commented in the source code, just listed here) - ./modules/core/src/mathfuncs_core.simd.hpp (Vector type wrapped in class/struct) - ./modules/imgproc/src/color_lab.cpp (Array of vector type) - ./modules/imgproc/src/color_rgb.simd.hpp (Array of vector type) - ./modules/imgproc/src/sumpixels.simd.hpp (fixed length algorithm, strongly ralated with `CV_SIMD_WIDTH`) These algorithms will need to be redesigned to accommodate scalable backends. ### Pull Request Readiness Checklist See details at https://github.com/opencv/opencv/wiki/How_to_contribute#making-a-good-pull-request - [ ] I agree to contribute to the project under Apache 2 License. - [ ] To the best of my knowledge, the proposed patch is not based on a code under GPL or another license that is incompatible with OpenCV - [ ] The PR is proposed to the proper branch - [ ] There is a reference to the original bug report and related work - [ ] There is accuracy test, performance test and test data in opencv_extra repository, if applicable Patch to opencv_extra has the same branch name. - [ ] The feature is well documented and sample code can be built with the project CMake

Rewrite fluid related part.

419060d

asmorkalov added category: g-api / gapi platform: riscv labels Sep 27, 2023

hanliutong mentioned this pull request Sep 27, 2023

Rewrite Universal Intrinsic code: float related part #24325

Merged

6 tasks

mshabunin self-assigned this Sep 27, 2023

opencv-alalek added the optimization label Sep 29, 2023

opencv-alalek added this to the 4.9.0 milestone Sep 29, 2023

hanliutong marked this pull request as ready for review October 3, 2023 12:04

asmorkalov requested review from TolyaTalamanov, dmatveev and mshabunin October 4, 2023 08:56

mshabunin approved these changes Oct 5, 2023

View reviewed changes

hanliutong mentioned this pull request Oct 7, 2023

Clean up the obsolete API of Universal Intrinsic #24371

Merged

6 tasks

asmorkalov approved these changes Oct 13, 2023

View reviewed changes

asmorkalov merged commit cd7cbe3 into opencv:4.x Oct 13, 2023

asmorkalov mentioned this pull request Oct 17, 2023

(5.x) Merge 4.x #24416

Merged

asmorkalov mentioned this pull request Oct 31, 2023

About the performance of opencv of the sizeless instruction(Riscv vector, SVE .etc) #21780

Closed

4 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Rewrite Universal Intrinsic code: gapi module (fluid part).#24324

Rewrite Universal Intrinsic code: gapi module (fluid part).#24324
asmorkalov merged 1 commit intoopencv:4.xfrom
hanliutong:rewrite-fluid

hanliutong commented Sep 27, 2023

Uh oh!

hanliutong commented Oct 3, 2023

Uh oh!

asmorkalov commented Oct 4, 2023 •

edited

Loading

Uh oh!

asmorkalov commented Oct 4, 2023

Uh oh!

asmorkalov commented Oct 12, 2023

Uh oh!

asmorkalov left a comment

Uh oh!

dmatveev commented Oct 13, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Uh oh!

Conversation

hanliutong commented Sep 27, 2023

Pull Request Readiness Checklist

Uh oh!

hanliutong commented Oct 3, 2023

Uh oh!

asmorkalov commented Oct 4, 2023 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

asmorkalov commented Oct 4, 2023

Uh oh!

asmorkalov commented Oct 12, 2023

Uh oh!

asmorkalov left a comment

Choose a reason for hiding this comment

Uh oh!

dmatveev commented Oct 13, 2023

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

asmorkalov commented Oct 4, 2023 •

edited

Loading