Simd v6.0.134

Algorithms

New features

• SSE4.1 optimizations of ResizerFloatBilinear class.

Improving

• Improve AVX2 optimizations of ResizerFloatBilinear class (AMD CPU).
• Improve AVX2 optimizations of ResizerShortBilinear class (AMD CPU).

Bug fixing

• MSVS compiler bug in file SimdAvx512bwYuvToBgraV2.
• Linux, GCC-13 - crash in function SimdSynetInnerProduct32fForward.
• MSVS compiler bug (Cmake, Windows for ARM64) with functions Extract4Sums.
• GCC-9/10 - compiler error in AVX-512BW optimization of function YToGray.
• GCC-9/10 - compiler error in AVX-512BW optimization of function GrayToY.

Replacing

• Replace AVX optimizations to AVX2 for function CosineDistance32f.
• Replace AVX optimizations to AVX2 for function Fill32f.
• Replace AVX optimizations to AVX2 for ResizerFloatBilinear class.
• Replace AVX optimizations to AVX2 for function SquaredDifferenceSum32f.
• Replace AVX optimizations to AVX2 for function SquaredDifferenceKahanSum32f.
• Replace AVX optimizations to AVX2 for function HogLiteFilterFeatures.
• Replace AVX optimizations to AVX2 for function HogLiteResizeFeatures.
• Replace AVX optimizations to AVX2 for function HogLiteCompressFeatures.
• Replace AVX optimizations to AVX2 for function HogLiteFilterSeparable.
• Replace AVX optimizations to AVX2 for function NeuralPooling2x2Max2x2.
• Replace AVX optimizations to AVX2 for function NeuralProductSum.
• Replace AVX optimizations to AVX2 for function NeuralAddVectorMultipliedByValue.
• Replace AVX optimizations to AVX2 for function NeuralAddVector.
• Replace AVX optimizations to AVX2 for function NeuralAddValue.
• Replace AVX optimizations to AVX2 for function NeuralRoughSigmoid.
• Replace AVX optimizations to AVX2 for function NeuralRoughSigmoid2.
• Replace AVX optimizations to AVX2 for function NeuralRoughTanh.
• Replace AVX optimizations to AVX2 for function NeuralDerivativeRelu.
• Replace AVX optimizations to AVX2 for function NeuralDerivativeTanh.
• Replace AVX optimizations to AVX2 for function NeuralDerivativeSigmoid.
• Replace AVX optimizations to AVX2 for function NeuralUpdateWeights.
• Replace AVX optimizations to AVX2 for function NeuralAdaptiveGradientUpdate.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution2x2Forward.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution3x3Forward.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution4x4Forward.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution5x5Forward.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution2x2Backward.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution3x3Backward.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution4x4Backward.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution5x5Backward.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution2x2Sum.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution3x3Sum.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution4x4Sum.
• Replace AVX optimizations to AVX2 for function NeuralAddConvolution5x5Sum.
• Replace AVX optimizations to AVX2 for function NeuralConvolutionForward.
• Replace AVX optimizations to AVX2 for function SynetAddBias.
• Replace AVX optimizations to AVX2 for function SynetFusedLayerForward0.
• Replace AVX optimizations to AVX2 for function SynetFusedLayerForward1.
• Replace AVX optimizations to AVX2 for function SynetFusedLayerForward2.
• Replace AVX optimizations to AVX2 for function SynetFusedLayerForward3.
• Replace AVX optimizations to AVX2 for function SynetFusedLayerForward4.
• Replace AVX optimizations to AVX2 for function SynetFusedLayerForward8.
• Replace AVX optimizations to AVX2 for function SynetFusedLayerForward9.
• Replace AVX optimizations to AVX2 for function SynetPoolingAverage.
• Replace AVX optimizations to AVX2 for function SynetShuffleLayerForward.
• Replace AVX optimizations to AVX2 for function SynetHardSigmoid32f.
• Replace AVX optimizations to AVX2 for function SynetHswish32f.
• Replace AVX optimizations to AVX2 for function SynetPreluLayerForward.
• Replace AVX optimizations to AVX2 for function SynetRelu32f.
• Replace AVX optimizations to AVX2 for function SynetRestrictRange32f.
• Replace AVX optimizations to AVX2 for function WinogradKernel1x3Block1x4SetFilter.
• Replace AVX optimizations to AVX2 for function WinogradKernel1x3Block1x4SetInput.
• Replace AVX optimizations to AVX2 for function WinogradKernel1x3Block1x4SetOutput.
• Replace AVX optimizations to AVX2 for function WinogradKernel1x5Block1x4SetFilter.
• Replace AVX optimizations to AVX2 for function WinogradKernel1x5Block1x4SetInput.
• Replace AVX optimizations to AVX2 for function WinogradKernel1x5Block1x4SetOutput.
• Replace AVX optimizations to AVX2 for function WinogradKernel2x2Block2x2SetFilter.
• Replace AVX optimizations to AVX2 for function WinogradKernel2x2Block2x2SetInput.
• Replace AVX optimizations to AVX2 for function WinogradKernel2x2Block2x2SetOutput.
• Replace AVX optimizations to AVX2 for function WinogradKernel2x2Block4x4SetFilter.
• Replace AVX optimizations to AVX2 for function WinogradKernel2x2Block4x4SetInput.
• Replace AVX optimizations to AVX2 for function WinogradKernel2x2Block4x4SetOutput.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block2x2SetFilter.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block2x2SetInput.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block2x2SetOutput.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block3x3SetFilter.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block3x3SetInput.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block3x3SetOutput.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block4x4SetFilter.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block4x4SetInput.
• Replace AVX optimizations to AVX2 for function WinogradKernel3x3Block4x4SetOutput.
• Replace AVX optimizations to AVX2 for function GemmPackA.
• Replace AVX optimizations to AVX2 for function GemmPackB.
• Replace AVX optimizations to AVX2 for function GemmScaleC.
• Replace AVX optimizations to AVX2 for function SynetScaleLayerForward.
• Replace AVX optimizations to AVX2 for function SynetInnerProductLayerForward.
• Replace AVX optimizations to AVX2 for function SynetInnerProduct32fInit.
• Replace AVX optimizations to AVX2 for function SynetEltwiseLayerForward.
• Replace AVX optimizations to AVX2 for function SynetDeconvolution32fInit.
• Replace AVX optimizations to AVX2 for function SynetMergedConvolution32fInit.
• Replace AVX optimizations to AVX2 for SynetInnerProduct32fGemm class.
• Replace AVX optimizations to AVX2 for SynetInnerProduct32fProd class.
• Replace AVX optimizations to AVX2 for SynetDeconvolution32fGemmNN class.
• Replace AVX optimizations to AVX2 for SynetDeconvolution32fNhwcDirect2x2 class.
• Replace AVX optimizations to AVX2 for SynetMergedConvolution32fCdc class.
• Replace AVX optimizations to AVX2 for SynetMergedConvolution32fCd class.
• Replace AVX optimizations to AVX2 for SynetMergedConvolution32fDc class.
• Replace AVX optimizations to AVX2 for SynetConvolution32fDepthwiseDotProduct class.
• Replace AVX optimizations to AVX2 for SynetConvolution32fDirectNchw class.
• Replace AVX optimizations to AVX2 for SynetConvolution32fDirectNhwc class.
• Replace AVX optimizations to AVX2 for SynetConvolution32fNhwcDirect class.
• Replace AVX optimizations to AVX2 for SynetConvolution32fGemmNT class.
• Replace AVX optimizations to AVX2 for SynetConvolution32fGemmNТ class.
• Replace AVX optimizations to AVX2 for SynetConvolution32fWinograd class.
• Replace AVX optimizations to AVX2 for function SynetConvolution32fInit.
• Replace AVX optimizations to AVX2 for function SynetMergedConvolution32fInit.
• Replace AVX optimizations to AVX2 for function SynetMergedConvolution32fInit.

Removing

• Base implementation, SSE4.1, AVX, AVX-512BW, NEON, VSX optimizations of function SvmSumLinear.
• Stopping of separate support of AVX extension (only together with AVX2).
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON, VMX optimizations of function EdgeBackgroundGrowRangeSlow.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON, VMX optimizations of function EdgeBackgroundGrowRangeFast.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON, VMX optimizations of function EdgeBackgroundIncrementCount.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON, VMX optimizations of function EdgeBackgroundAdjustRange.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON, VMX optimizations of function EdgeBackgroundAdjustRangeMasked.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON, VMX optimizations of function EdgeBackgroundShiftRange.
• Base implementatio...

Simd v5.4.133

Algorithms

New features

• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function SynetNormalizeLayerForwardV4.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of class SynetConvolution32fNhwcGroupedBlock1x2.
• Function ImageSaveToFile can choose output file format (if it is undefined) by file extension.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function GrayToY.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function YToGray.
• Add yuvType parameter in Frame structure.
• Support of SimdSynetUnaryOperation32fCeil in function SynetUnaryOperation32f.
• Support of SimdSynetUnaryOperation32fFloor in function SynetUnaryOperation32f.
• Function of Simd::Yuva444pToBgra.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuva422pToBgraV2.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuva420pToBgraV2.
• The mark of function SimdYuva420pToBgra as deprecated.

Python wrapper

New features

• Wrapper for enumeration Simd.WarpAffineFlags.
• Wrapper for function SimdWarpAffineInit.
• Wrapper for function SimdWarpAffineRun.
• Function Simd.WarpAffine.
• Wrapper for function SimdAbsGradientSaturatedSum.
• Function Simd.AbsGradientSaturatedSum.
• Wrapper for function SimdBgraToBgr.
• Method Simd.Image.Copy.
• Method Simd.Image.Convert.
• Method Simd.Image.Converted.
• Wrapper for function SimdBgraToGray.
• Wrapper for function SimdBgraToRgb.
• Wrapper for function SimdBgraToRgba.
• Wrapper for function SimdCopy.
• Wrapper for function SimdBgrToBgra.
• Wrapper for function SimdBgrToGray.
• Wrapper for function SimdBgrToRgb.
• Wrapper for function SimdRgbToBgra.
• Wrapper for function SimdRgbToGray.
• Wrapper for function SimdRgbaToGray.
• Wrapper for function SimdBgraToYuv420pV2.
• Wrapper for enumeration Simd.FrameFormat.
• Class Simd.ImageFrame.
• Method Simd.ImageFrame.Copy.
• Method Simd.ImageFrame.Convert.
• Method Simd.ImageFrame.Converted.
• Wrapper for function SimdDeinterleaveUv.
• Wrapper for function SimdInterleaveUv.
• Wrapper for function SimdBgrToYuv420pV2.
• Wrapper for function SimdGrayToBgra.
• Wrapper for function SimdGrayToBgr.
• Method Simd.Image.Fill.
• Wrapper for function SimdYuv420pToBgraV2.
• Wrapper for function SimdYuv420pToBgrV2.
• Wrapper for function SimdYuv420pToRgbV2.
• Wrapper for function SimdYToGray.
• Wrapper for function SimdGrayToY.

Test framework

New features

• Tests for verifying functionality of function SynetNormalizeLayerForwardV4.
• Special test for verifying functionality of function Yuv420pToRgbV2.
• Tests for verifying functionality of function GrayToY.
• Tests for verifying functionality of function YToGray.
• Tests for verifying functionality of function Yuva422pToBgraV2.
• Tests for verifying functionality of function Yuva420pToBgraV2.

Infrastructure

Bug fixing

• Error in CMake for ARM platform.

Simd v5.4.132

Algorithms

New features

• Support of RGBA-32 input image format in base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function SynetSetInput.

Bug fixing

• Wrong order of SIMD_DEPRECATED macro.
• Error in AVX-512BW optimizations of function SynetSoftmaxLayerForward.

Python wrapper

New features

• Wrapper for function SimdVersion.
• Wrapper for function SimdRelease.
• Wrapper for function SimdCpuDesc.
• Wrapper for function SimdCpuInfo.
• Wrapper for enumeration SimdCpuDescType.
• Wrapper for enumeration SimdCpuInfoType.
• Wrapper for function SimdPerformanceStatistic.
• Wrapper for function SimdAllocate.
• Wrapper for function SimdFree.
• Wrapper for function SimdAlign.
• Wrapper for function SimdAlignment.
• Wrapper for function SimdGetThreadNumber.
• Wrapper for function SimdSetThreadNumber.
• Wrapper for function SimdEmpty.
• Wrapper for function SimdGetFastMode.
• Wrapper for function SimdSetFastMode.
• Wrapper for enumeration SimdPixelFormatType.
• Class Simd.Image.
• Wrapper for function SimdCrc32.
• Wrapper for function SimdCrc32c.
• Wrapper for enumeration SimdImageFileType.
• Wrapper for function SimdImageSaveToFile.
• Wrapper for function SimdImageLoadFromFile.
• Wrapper for enumeration Simd::View::Position.
• Method Simd.Image.Region.
• Method Simd.Image.RegionAt.
• Wrapper for enumeration Simd.ResizeMethod.
• Wrapper for enumeration Simd.ResizeChannel.
• Wrapper for function SimdResizerInit.
• Wrapper for function SimdResizerRun
• Function Simd.Resize.
• Function Simd.Resized.
• Wrapper for function SimdSynetSetInput.
• Function Simd.SynetSetInput.
• Wrapper for enumeration Simd.TensorFormat.
• Wrapper for enumeration Simd.TensorData.
• Wrapper for enumeration Simd.YuvType.
• Wrapper for function SimdFillPixel.
• Function Simd.FillPixel.

Infrastructure

New features

• SimdPy MSVS project.

Documentation

New features

• Doxygen generation of documentation for Python wrapper.

Simd v5.3.131

Algorithms

New features

• NEON optimizations of function DescrIntCosineDistance.
• NEON optimizations of function DescrIntCosineDistancesMxNa.
• NEON optimizations of function DescrIntCosineDistancesMxNp.

Improving

• NEON optimizations of function DescrIntDecode32f.
• NEON optimizations of function CorrelationSum.
• Base implementation and SSE4.1 optimizations of ImageJpegLoader class.

Bug fixing

• Error in defenition of SIMD_CPP_2011_ENABLE macro for Visual Studio.
• Error in defenition of SIMD_CPP_2014_ENABLE macro for Visual Studio.
• Error in defenition of SIMD_CPP_2017_ENABLE macro for Visual Studio.
• Compiler warning in method Detection::InitLevels.
• Compiler warning in method Detection::FillLevels.

Infrastructure

New features

• Cmake SIMD_UNINSTALL option.
• Cmake SIMD_INSTALL option.

The use examples

New features

• An example how to use Simd::ImageMatcher to find image duplicates.

Simd v5.3.130

Algorithms

New features

• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function BgrToYuv420pV2.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function BgrToYuv422pV2.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function BgrToYuv444pV2.
• Parameter yuvType in function Simd::BgrToYuv420p.
• Parameter yuvType in function Simd::BgrToYuv422p.
• Parameter yuvType in function Simd::BgrToYuv444p.
• The mark of function SimdBgrToYuv420p as deprecated.
• The mark of function SimdBgrToYuv422p as deprecated.
• The mark of function SimdBgrToYuv444p as deprecated.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuv420pToRgbV2.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuv422pToRgbV2.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuv444pToRgbV2.
• Parameter yuvType in function Simd::Yuv420pToRgb.
• Parameter yuvType in function Simd::Yuv422pToRgb.
• Parameter yuvType in function Simd::Yuv444pToRgb
• The mark of function SimdYuv420pToRgb as deprecated.
• The mark of function SimdYuv422pToRgb as deprecated.
• The mark of function SimdYuv444pToRgb as deprecated.
• NEON optimizations of function AlphaBlendingBgraToYuv420p.
• NEON optimizations of function DescrIntEncode32f.
• NEON optimizations of function DescrIntEncode16f.
• NEON optimizations of function DescrIntDecode32f.
• NEON optimizations of function DescrIntDecode16f.

Bug fixing

• Error in AVX-512BW optimizations of function SynetSoftmaxLayerForward.
• Error in AVX2 optimizations of class ResizerByteArea2x2 (internal buffer overflow).
• Error in function Simd::BgraToYuv420p.
• Error in function Simd::BgraToYuv422p.
• Error in function Simd::BgraToYuv444p.
• Error in NEON optimizations of class MergedConvolution32fCd.

Test framework

New features

• Tests for verifying functionality of function BgrToYuv420pV2.
• Tests for verifying functionality of function BgrToYuv422pV2.
• Tests for verifying functionality of function BgrToYuv444pV2.
• Tests for verifying functionality of function Yuv420pToRgbV2.
• Tests for verifying functionality of function Yuv422pToRgbV2.
• Tests for verifying functionality of function Yuv444pToRgbV2.

Simd v5.3.129

Algorithms

New features

• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuv420pToBgrV2.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuv422pToBgrV2.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuv444pToBgrV2.
• Parameter yuvType in function Simd::Yuv420pToBgr.
• Parameter yuvType in function Simd::Yuv422pToBgr.
• Parameter yuvType in function Simd::Yuv444pToBgr.
• The mark of function SimdYuv420pToBgr as deprecated.
• The mark of function SimdYuv422pToBgr as deprecated.
• The mark of function SimdYuv444pToBgr as deprecated.
• Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Yuv422pToBgraV2.
• Parameter yuvType in function Simd::Yuv420pToBgra.
• Parameter yuvType in function Simd::Yuv422pToBgra.
• Parameter yuvType in function Simd::Yuv444pToBgra.
• The mark of function SimdYuv420pToBgra as deprecated.
• The mark of function SimdYuv422pToBgra as deprecated.
• The mark of function SimdYuv444pToBgra as deprecated.
• Parameter yuvType in function Simd::BgraToYuv420p.
• Parameter yuvType in function Simd::BgraToYuv422p.
• Parameter yuvType in function Simd::BgraToYuv444p.
• Parameter yuvType in function Simd::BgraToYuva420p.
• The mark of function SimdBgraToYuv420p as deprecated.
• The mark of function SimdBgraToYuv422p as deprecated.
• The mark of function SimdBgraToYuv444p as deprecated.
• The mark of function SimdBgraToYuva420p as deprecated.
• The mark of function SimdResizeBilinear as deprecated.
• The mark of function Simd::ResizeBilinear as deprecated.
• The mark of function Simd::ResizeAreaGray as deprecated.
• The mark of function Simd::ResizeArea as deprecated.
• The mark of function Simd::InterferenceIncrement as deprecated.
• The mark of function Simd::InterferenceIncrementMasked as deprecated.
• The mark of function Simd::InterferenceDecrement as deprecated.
• The mark of function Simd::InterferenceDecrementMasked as deprecated.
• The mark of function SimdSynetFusedLayerForward0 as deprecated.
• The mark of function SimdSynetFusedLayerForward1 as deprecated.
• The mark of function SimdSynetFusedLayerForward2 as deprecated.
• The mark of function SimdSynetFusedLayerForward3 as deprecated.
• The mark of function SimdSynetFusedLayerForward4 as deprecated.
• The mark of function SimdSynetFusedLayerForward8 as deprecated.
• The mark of function SimdSynetFusedLayerForward9 as deprecated.

Bug fixing

• Error in NEON optimizations of Resizer engine.
• The memory reading outside border of input array in Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntDecode32f.
• The memory reading outside border of input array in Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntDecode16f.
• The memory reading outside border of input array in Base implementation, SSE4.1, AVX2 optimizations of function DescrIntCosineDistance.
• The memory reading outside border of input array in Base implementation, SSE4.1, AVX2 optimizations of function DescrIntCosineDistancesMxNa.
• The memory reading outside border of input array in Base implementation, SSE4.1, AVX2 optimizations of function DescrIntCosineDistancesMxNp.
• Error in AVX-512BW optimizations of function DescrIntEncode32f.
• Error in AVX-512BW optimizations of function DescrIntEncode16f.
• Compiler error in function Simd::ResizeArea.
• Error in SSE4.1, AVX, AVX2, AVX-512BW, NEON optimizations of class MergedConvolution32fCd.

Test framework

New features

• Tests for verifying functionality of function Yuv420pToBgrV2.
• Tests for verifying functionality of function Yuv422pToBgrV2.
• Tests for verifying functionality of function Yuv444pToBgrV2.
• Tests for verifying functionality of function Yuv422pToBgraV2.
• Special test for verifying functionality of function Simd::ResizeAreaGray.
• Special test for verifying functionality of function Simd::ResizeArea.

Simd v5.3.128

Algorithms

New features

• Support of SimdCpuInfoRam in function SimdCpuInfo.
• Support of SimdCpuInfoRam in function Simd::PrintInfo.
• Base implementation of function SimdCpuDesc.
• Base implementation of SynetGridSample2dRef class.
• Base implementation, SSE4.1, AVX2 optimizations of SynetGridSample2d32fBlZ class.

Bug fixing

• Error in AVX-512VNNI optimizations of class SynetMergedConvolution8iCdc.
• Error in AVX-512VNNI optimizations of class SynetMergedConvolution8iCd.
• Error in AVX-512VNNI optimizations of class SynetMergedConvolution8iDc.
• Error (assert) in Base implementation of class ResizerNearest.
• Error in SSE4.1, AVX, AVX2, AVX-512BW optimizations of class SynetMergedConvolution32fCd.

Test framework

New features

• Tests for verifying functionality of SynetGridSample2d engine.

Improving

• WIN32 performance report.

Infrastructure

New features

• Github actions script for CMake (build and test for GCC-13 (instead of GCC-12), Linux).

Documentation

Bug fixing

• Wrong description of function SimdDescrIntInit.

Simd v5.3.127

Algorithms

New features

• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntEncode16f.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntDecode16f.
• Support of 4-bit and 5-bit depth in Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntEncode32f.
• Support of 4-bit and 5-bit depth in Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntEncode16f.
• Support of 4-bit and 5-bit depth in Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntDecode32f.
• Support of 4-bit and 5-bit depth in Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntDecode16f.
• Support of 4-bit and 5-bit depth in Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntCosineDistance.
• Support of 4-bit and 5-bit depth in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512VNNI optimizations of function DescrIntCosineDistancesMxNp.
• Support of 4-bit and 5-bit depth in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512VNNI optimizations of function DescrIntCosineDistancesMxNa.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function SynetNormalizeLayerForwardV3.

Improving

• SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntCosineDistancesMxNp for 4, 5, 6, 7-bits depth.
• SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntCosineDistancesMxNa for 4, 5, 6, 7-bits depth.

Bug fixing

• Compiler error in file SimdYuvToBgr.h.

Renaming

• Function DescrIntEncode to DescrIntEncode32f.
• Function DescrIntDecode to DescrIntDecode32f.

Test framework

New features

• Tests for verifying functionality of function DescrIntEncode16f.
• Tests for verifying functionality of function DescrIntDecode16f.
• Tests for verifying functionality of function SynetNormalizeLayerForwardV3

Improving

• WIN32 exception handling.

Infrastructure

Improving

• Host Properties step in Github actions script for MSBuild.

Simd v5.3.126

Algorithms

New features

• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntEncode.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntDecode.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntCosineDistance.
• Base implementation optimizations of function DescrIntVectorNorm.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntCosineDistancesMxNp.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function DescrIntCosineDistancesMxNa.
• SimdSynetUnaryOperation32fRcp member in enumeration SimdSynetUnaryOperationType.
• Support of SimdSynetUnaryOperation32fRcp in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function SynetUnaryOperation32f.
• SimdSynetUnaryOperation32fNot member in enumeration SimdSynetUnaryOperationType.
• Support of SimdSynetUnaryOperation32fNot in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function SynetUnaryOperation32f.
• Support of SimdSynetUnaryOperation32fNot in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of function SynetUnaryOperation32f.
• Helper function Simd::InvertAffineTransform.

Improving

• SSE4.1, AVX, AVX-512BW optimizations of method SynetDeconvolution32fGemmNN::RowToImg.
• AVX-512BW optimizations of function SynetUnaryOperation32f (case of SimdSynetUnaryOperation32fLog, SimdSynetUnaryOperation32fExp).
• AVX-512BW optimizations of function SynetSoftplus32f.

Bug fixing

• Error in AVX2 optimizations of function SynetSoftmaxLayerForward.
• GCC compiler error in file SimdDrawing.hpp (Windows, MinGW).
• GCC compiler error in function Test::FileExists (Windows, MinGW).
• Crash in function SimdSynetDeconvolution32fForward (Linux, GCC-12, GCC-13).
• Crash in function Base::AlgCacheL3 (Windows, MinGW).
• Using of _WIN32 macro instead of WIN32.

Test framework

New features

• Tests for verifying functionality of function DescrIntEncode.
• Tests for verifying functionality of function DescrIntDecode.
• Tests for verifying functionality of function DescrIntCosineDistance.
• Tests for verifying functionality of function DescrIntCosineDistancesMxNp.
• Tests for verifying functionality of function DescrIntCosineDistancesMxNa.
• Test command line argument '-cc' to run CheckCpp tests.

Bug fixing

• Test log messages after test error stoppage (multithreaded run).
• Error in CheckCpp tests.

Infrastructure

New features

• Github actions script for CMake (build and test for MinGW, Windows).
• Github actions script for CMake (build and test for GCC-12, Linux).

Simd v5.2.125

Algorithms

New features

• AVX-512BW, NEON optimizations of function SynetGelu32f.
• SimdConvolutionActivationGelu member in enumeration SimdConvolutionActivationType.
• Support of SynetUnaryOperation32fErf in NEON optimizations of function SynetUnaryOperation32f.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetConvolution32fGemmNN.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetConvolution32fWinograd.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetConvolution32fGemmNT.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetConvolution32fDirectNchw.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetConvolution32fDirectNhwc.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetConvolution32fDepthwiseDotProduct.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetConvolution32fNhwcDirect.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512BF16, AMX-BF16 optimizations of class SynetConvolution32fBf16Gemm.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512BF16, AMX-BF16 optimizations of class SynetConvolution32fBf16Nhwc.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetDeconvolution32fGemmNN.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, NEON optimizations of class SynetDeconvolution32fNhwcDirect2x2.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512VNNI, AMX-INT8, NEON optimizations of class SynetConvolution8iGemmNN.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512VNNI, AMX-INT8, NEON optimizations of class SynetConvolution8iNhwcDirect.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512VNNI, AMX-INT8, NEON optimizations of class SynetConvolution8iNhwcDepthwise.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX, AVX2, AVX-512BW, NEON optimizations of class SynetMergedConvolution32fCd.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX, AVX2, AVX-512BW, NEON optimizations of class SynetMergedConvolution32fCdc.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX, AVX2, AVX-512BW, NEON optimizations of class SynetMergedConvolution32fDc.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512BF16, AMX-BF16 optimizations of class SynetMergedConvolution32fBf16Cd.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512BF16, AMX-BF16 optimizations of class SynetMergedConvolution32fBf16Cdc.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512BF16, AMX-BF16 optimizations of class SynetMergedConvolution32fBf16Dc.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512VNNI, AMX-INT8 optimizations of class SynetMergedConvolution8iCd.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512VNNI, AMX-INT8 optimizations of class SynetMergedConvolution8iCdc.
• Support of SimdConvolutionActivationGelu in Base implementation, SSE4.1, AVX2, AVX-512BW, AVX-512VNNI, AMX-INT8 optimizations of class SynetMergedConvolution8iDc.
• Base implementation, SSE4.1, AVX2, AVX-512BW optimizations of function SynetNormalizeLayerForwardV2.

Improving

• SSE4.1, AVX2, AVX-512BW, NEON optimizations of function Erf.
• Performance of function Simd::Parallel.
• Using of resize method SimdResizeMethodArea in ImageMatcher::Create (more precise than SimdResizeMethodBilinear).

Bug fixing

• Compiler error in SimdImageMatcher.hpp (using of internal functions).
• Wrong API of function ImageMatcher.Skip().

Test framework

New features

• Tests for verifying functionality of function SynetNormalizeLayerForwardV2.

Infrastructure

Bug fixing

• Github actions script for MSBuild (Restore NuGet packages timeout exit).