tomoto 0.2.3 → 0.3.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +6 -0
- data/README.md +8 -10
- data/ext/tomoto/extconf.rb +6 -2
- data/ext/tomoto/{ext.cpp → tomoto.cpp} +1 -1
- data/lib/tomoto/version.rb +1 -1
- data/lib/tomoto.rb +5 -1
- data/vendor/EigenRand/EigenRand/Core.h +10 -10
- data/vendor/EigenRand/EigenRand/Dists/Basic.h +208 -9
- data/vendor/EigenRand/EigenRand/Dists/Discrete.h +52 -31
- data/vendor/EigenRand/EigenRand/Dists/GammaPoisson.h +9 -8
- data/vendor/EigenRand/EigenRand/Dists/NormalExp.h +28 -21
- data/vendor/EigenRand/EigenRand/EigenRand +11 -6
- data/vendor/EigenRand/EigenRand/Macro.h +13 -7
- data/vendor/EigenRand/EigenRand/MorePacketMath.h +348 -740
- data/vendor/EigenRand/EigenRand/MvDists/Multinomial.h +5 -3
- data/vendor/EigenRand/EigenRand/MvDists/MvNormal.h +9 -3
- data/vendor/EigenRand/EigenRand/PacketFilter.h +11 -253
- data/vendor/EigenRand/EigenRand/PacketRandomEngine.h +21 -47
- data/vendor/EigenRand/EigenRand/RandUtils.h +50 -344
- data/vendor/EigenRand/EigenRand/arch/AVX/MorePacketMath.h +619 -0
- data/vendor/EigenRand/EigenRand/arch/AVX/PacketFilter.h +149 -0
- data/vendor/EigenRand/EigenRand/arch/AVX/RandUtils.h +228 -0
- data/vendor/EigenRand/EigenRand/arch/NEON/MorePacketMath.h +473 -0
- data/vendor/EigenRand/EigenRand/arch/NEON/PacketFilter.h +142 -0
- data/vendor/EigenRand/EigenRand/arch/NEON/RandUtils.h +126 -0
- data/vendor/EigenRand/EigenRand/arch/SSE/MorePacketMath.h +501 -0
- data/vendor/EigenRand/EigenRand/arch/SSE/PacketFilter.h +133 -0
- data/vendor/EigenRand/EigenRand/arch/SSE/RandUtils.h +120 -0
- data/vendor/EigenRand/EigenRand/doc.h +24 -12
- data/vendor/EigenRand/README.md +57 -4
- data/vendor/eigen/COPYING.APACHE +203 -0
- data/vendor/eigen/COPYING.BSD +1 -1
- data/vendor/eigen/COPYING.MINPACK +51 -52
- data/vendor/eigen/Eigen/Cholesky +0 -1
- data/vendor/eigen/Eigen/Core +112 -265
- data/vendor/eigen/Eigen/Eigenvalues +2 -3
- data/vendor/eigen/Eigen/Geometry +5 -8
- data/vendor/eigen/Eigen/Householder +0 -1
- data/vendor/eigen/Eigen/Jacobi +0 -1
- data/vendor/eigen/Eigen/KLUSupport +41 -0
- data/vendor/eigen/Eigen/LU +2 -5
- data/vendor/eigen/Eigen/OrderingMethods +0 -3
- data/vendor/eigen/Eigen/PaStiXSupport +1 -0
- data/vendor/eigen/Eigen/PardisoSupport +0 -0
- data/vendor/eigen/Eigen/QR +2 -3
- data/vendor/eigen/Eigen/QtAlignedMalloc +0 -1
- data/vendor/eigen/Eigen/SVD +0 -1
- data/vendor/eigen/Eigen/Sparse +0 -2
- data/vendor/eigen/Eigen/SparseCholesky +0 -8
- data/vendor/eigen/Eigen/SparseLU +4 -0
- data/vendor/eigen/Eigen/SparseQR +0 -1
- data/vendor/eigen/Eigen/src/Cholesky/LDLT.h +42 -27
- data/vendor/eigen/Eigen/src/Cholesky/LLT.h +39 -23
- data/vendor/eigen/Eigen/src/CholmodSupport/CholmodSupport.h +90 -47
- data/vendor/eigen/Eigen/src/Core/ArithmeticSequence.h +413 -0
- data/vendor/eigen/Eigen/src/Core/Array.h +99 -11
- data/vendor/eigen/Eigen/src/Core/ArrayBase.h +3 -3
- data/vendor/eigen/Eigen/src/Core/ArrayWrapper.h +21 -21
- data/vendor/eigen/Eigen/src/Core/Assign.h +1 -1
- data/vendor/eigen/Eigen/src/Core/AssignEvaluator.h +125 -50
- data/vendor/eigen/Eigen/src/Core/Assign_MKL.h +10 -10
- data/vendor/eigen/Eigen/src/Core/BandMatrix.h +16 -16
- data/vendor/eigen/Eigen/src/Core/Block.h +56 -60
- data/vendor/eigen/Eigen/src/Core/BooleanRedux.h +29 -31
- data/vendor/eigen/Eigen/src/Core/CommaInitializer.h +7 -3
- data/vendor/eigen/Eigen/src/Core/CoreEvaluators.h +325 -272
- data/vendor/eigen/Eigen/src/Core/CoreIterators.h +5 -0
- data/vendor/eigen/Eigen/src/Core/CwiseBinaryOp.h +21 -22
- data/vendor/eigen/Eigen/src/Core/CwiseNullaryOp.h +153 -18
- data/vendor/eigen/Eigen/src/Core/CwiseUnaryOp.h +6 -6
- data/vendor/eigen/Eigen/src/Core/CwiseUnaryView.h +14 -10
- data/vendor/eigen/Eigen/src/Core/DenseBase.h +132 -42
- data/vendor/eigen/Eigen/src/Core/DenseCoeffsBase.h +25 -21
- data/vendor/eigen/Eigen/src/Core/DenseStorage.h +153 -71
- data/vendor/eigen/Eigen/src/Core/Diagonal.h +21 -23
- data/vendor/eigen/Eigen/src/Core/DiagonalMatrix.h +50 -2
- data/vendor/eigen/Eigen/src/Core/DiagonalProduct.h +1 -1
- data/vendor/eigen/Eigen/src/Core/Dot.h +10 -10
- data/vendor/eigen/Eigen/src/Core/EigenBase.h +10 -9
- data/vendor/eigen/Eigen/src/Core/ForceAlignedAccess.h +8 -4
- data/vendor/eigen/Eigen/src/Core/Fuzzy.h +3 -3
- data/vendor/eigen/Eigen/src/Core/GeneralProduct.h +20 -10
- data/vendor/eigen/Eigen/src/Core/GenericPacketMath.h +599 -152
- data/vendor/eigen/Eigen/src/Core/GlobalFunctions.h +40 -33
- data/vendor/eigen/Eigen/src/Core/IO.h +40 -7
- data/vendor/eigen/Eigen/src/Core/IndexedView.h +237 -0
- data/vendor/eigen/Eigen/src/Core/Inverse.h +9 -10
- data/vendor/eigen/Eigen/src/Core/Map.h +7 -7
- data/vendor/eigen/Eigen/src/Core/MapBase.h +10 -3
- data/vendor/eigen/Eigen/src/Core/MathFunctions.h +767 -125
- data/vendor/eigen/Eigen/src/Core/MathFunctionsImpl.h +118 -19
- data/vendor/eigen/Eigen/src/Core/Matrix.h +131 -25
- data/vendor/eigen/Eigen/src/Core/MatrixBase.h +21 -3
- data/vendor/eigen/Eigen/src/Core/NestByValue.h +25 -50
- data/vendor/eigen/Eigen/src/Core/NoAlias.h +4 -3
- data/vendor/eigen/Eigen/src/Core/NumTraits.h +107 -20
- data/vendor/eigen/Eigen/src/Core/PartialReduxEvaluator.h +232 -0
- data/vendor/eigen/Eigen/src/Core/PermutationMatrix.h +3 -31
- data/vendor/eigen/Eigen/src/Core/PlainObjectBase.h +152 -59
- data/vendor/eigen/Eigen/src/Core/Product.h +30 -25
- data/vendor/eigen/Eigen/src/Core/ProductEvaluators.h +192 -125
- data/vendor/eigen/Eigen/src/Core/Random.h +37 -1
- data/vendor/eigen/Eigen/src/Core/Redux.h +180 -170
- data/vendor/eigen/Eigen/src/Core/Ref.h +121 -23
- data/vendor/eigen/Eigen/src/Core/Replicate.h +8 -8
- data/vendor/eigen/Eigen/src/Core/Reshaped.h +454 -0
- data/vendor/eigen/Eigen/src/Core/ReturnByValue.h +7 -5
- data/vendor/eigen/Eigen/src/Core/Reverse.h +18 -12
- data/vendor/eigen/Eigen/src/Core/Select.h +8 -6
- data/vendor/eigen/Eigen/src/Core/SelfAdjointView.h +33 -20
- data/vendor/eigen/Eigen/src/Core/Solve.h +14 -14
- data/vendor/eigen/Eigen/src/Core/SolveTriangular.h +16 -16
- data/vendor/eigen/Eigen/src/Core/SolverBase.h +41 -3
- data/vendor/eigen/Eigen/src/Core/StableNorm.h +100 -70
- data/vendor/eigen/Eigen/src/Core/StlIterators.h +463 -0
- data/vendor/eigen/Eigen/src/Core/Stride.h +9 -4
- data/vendor/eigen/Eigen/src/Core/Swap.h +5 -4
- data/vendor/eigen/Eigen/src/Core/Transpose.h +88 -27
- data/vendor/eigen/Eigen/src/Core/Transpositions.h +26 -47
- data/vendor/eigen/Eigen/src/Core/TriangularMatrix.h +93 -75
- data/vendor/eigen/Eigen/src/Core/VectorBlock.h +5 -5
- data/vendor/eigen/Eigen/src/Core/VectorwiseOp.h +159 -70
- data/vendor/eigen/Eigen/src/Core/Visitor.h +137 -29
- data/vendor/eigen/Eigen/src/Core/arch/AVX/Complex.h +50 -129
- data/vendor/eigen/Eigen/src/Core/arch/AVX/MathFunctions.h +126 -337
- data/vendor/eigen/Eigen/src/Core/arch/AVX/PacketMath.h +1092 -155
- data/vendor/eigen/Eigen/src/Core/arch/AVX/TypeCasting.h +65 -1
- data/vendor/eigen/Eigen/src/Core/arch/AVX512/Complex.h +422 -0
- data/vendor/eigen/Eigen/src/Core/arch/AVX512/MathFunctions.h +207 -236
- data/vendor/eigen/Eigen/src/Core/arch/AVX512/PacketMath.h +1482 -495
- data/vendor/eigen/Eigen/src/Core/arch/AVX512/TypeCasting.h +89 -0
- data/vendor/eigen/Eigen/src/Core/arch/AltiVec/Complex.h +152 -165
- data/vendor/eigen/Eigen/src/Core/arch/AltiVec/MathFunctions.h +19 -251
- data/vendor/eigen/Eigen/src/Core/arch/AltiVec/MatrixProduct.h +2937 -0
- data/vendor/eigen/Eigen/src/Core/arch/AltiVec/MatrixProductCommon.h +221 -0
- data/vendor/eigen/Eigen/src/Core/arch/AltiVec/MatrixProductMMA.h +629 -0
- data/vendor/eigen/Eigen/src/Core/arch/AltiVec/PacketMath.h +2042 -392
- data/vendor/eigen/Eigen/src/Core/arch/CUDA/Complex.h +235 -80
- data/vendor/eigen/Eigen/src/Core/arch/Default/BFloat16.h +700 -0
- data/vendor/eigen/Eigen/src/Core/arch/Default/ConjHelper.h +102 -14
- data/vendor/eigen/Eigen/src/Core/arch/Default/GenericPacketMathFunctions.h +1649 -0
- data/vendor/eigen/Eigen/src/Core/arch/Default/GenericPacketMathFunctionsFwd.h +110 -0
- data/vendor/eigen/Eigen/src/Core/arch/Default/Half.h +942 -0
- data/vendor/eigen/Eigen/src/Core/arch/Default/Settings.h +1 -1
- data/vendor/eigen/Eigen/src/Core/arch/Default/TypeCasting.h +120 -0
- data/vendor/eigen/Eigen/src/Core/arch/{CUDA → GPU}/MathFunctions.h +16 -4
- data/vendor/eigen/Eigen/src/Core/arch/GPU/PacketMath.h +1685 -0
- data/vendor/eigen/Eigen/src/Core/arch/GPU/TypeCasting.h +80 -0
- data/vendor/eigen/Eigen/src/Core/arch/HIP/hcc/math_constants.h +23 -0
- data/vendor/eigen/Eigen/src/Core/arch/MSA/Complex.h +648 -0
- data/vendor/eigen/Eigen/src/Core/arch/MSA/MathFunctions.h +387 -0
- data/vendor/eigen/Eigen/src/Core/arch/MSA/PacketMath.h +1233 -0
- data/vendor/eigen/Eigen/src/Core/arch/NEON/Complex.h +313 -219
- data/vendor/eigen/Eigen/src/Core/arch/NEON/GeneralBlockPanelKernel.h +183 -0
- data/vendor/eigen/Eigen/src/Core/arch/NEON/MathFunctions.h +54 -70
- data/vendor/eigen/Eigen/src/Core/arch/NEON/PacketMath.h +4376 -549
- data/vendor/eigen/Eigen/src/Core/arch/NEON/TypeCasting.h +1419 -0
- data/vendor/eigen/Eigen/src/Core/arch/SSE/Complex.h +59 -179
- data/vendor/eigen/Eigen/src/Core/arch/SSE/MathFunctions.h +65 -428
- data/vendor/eigen/Eigen/src/Core/arch/SSE/PacketMath.h +893 -283
- data/vendor/eigen/Eigen/src/Core/arch/SSE/TypeCasting.h +65 -0
- data/vendor/eigen/Eigen/src/Core/arch/SVE/MathFunctions.h +44 -0
- data/vendor/eigen/Eigen/src/Core/arch/SVE/PacketMath.h +752 -0
- data/vendor/eigen/Eigen/src/Core/arch/SVE/TypeCasting.h +49 -0
- data/vendor/eigen/Eigen/src/Core/arch/SYCL/InteropHeaders.h +232 -0
- data/vendor/eigen/Eigen/src/Core/arch/SYCL/MathFunctions.h +301 -0
- data/vendor/eigen/Eigen/src/Core/arch/SYCL/PacketMath.h +670 -0
- data/vendor/eigen/Eigen/src/Core/arch/SYCL/SyclMemoryModel.h +694 -0
- data/vendor/eigen/Eigen/src/Core/arch/SYCL/TypeCasting.h +85 -0
- data/vendor/eigen/Eigen/src/Core/arch/ZVector/Complex.h +212 -183
- data/vendor/eigen/Eigen/src/Core/arch/ZVector/MathFunctions.h +101 -5
- data/vendor/eigen/Eigen/src/Core/arch/ZVector/PacketMath.h +510 -395
- data/vendor/eigen/Eigen/src/Core/functors/AssignmentFunctors.h +11 -2
- data/vendor/eigen/Eigen/src/Core/functors/BinaryFunctors.h +112 -46
- data/vendor/eigen/Eigen/src/Core/functors/NullaryFunctors.h +31 -30
- data/vendor/eigen/Eigen/src/Core/functors/StlFunctors.h +32 -2
- data/vendor/eigen/Eigen/src/Core/functors/UnaryFunctors.h +355 -16
- data/vendor/eigen/Eigen/src/Core/products/GeneralBlockPanelKernel.h +1075 -586
- data/vendor/eigen/Eigen/src/Core/products/GeneralMatrixMatrix.h +49 -24
- data/vendor/eigen/Eigen/src/Core/products/GeneralMatrixMatrixTriangular.h +41 -35
- data/vendor/eigen/Eigen/src/Core/products/GeneralMatrixMatrixTriangular_BLAS.h +6 -6
- data/vendor/eigen/Eigen/src/Core/products/GeneralMatrixMatrix_BLAS.h +4 -2
- data/vendor/eigen/Eigen/src/Core/products/GeneralMatrixVector.h +382 -483
- data/vendor/eigen/Eigen/src/Core/products/Parallelizer.h +22 -5
- data/vendor/eigen/Eigen/src/Core/products/SelfadjointMatrixMatrix.h +53 -30
- data/vendor/eigen/Eigen/src/Core/products/SelfadjointMatrixMatrix_BLAS.h +16 -8
- data/vendor/eigen/Eigen/src/Core/products/SelfadjointMatrixVector.h +8 -6
- data/vendor/eigen/Eigen/src/Core/products/SelfadjointProduct.h +4 -4
- data/vendor/eigen/Eigen/src/Core/products/SelfadjointRank2Update.h +5 -4
- data/vendor/eigen/Eigen/src/Core/products/TriangularMatrixMatrix.h +33 -27
- data/vendor/eigen/Eigen/src/Core/products/TriangularMatrixMatrix_BLAS.h +14 -12
- data/vendor/eigen/Eigen/src/Core/products/TriangularSolverMatrix.h +36 -34
- data/vendor/eigen/Eigen/src/Core/products/TriangularSolverMatrix_BLAS.h +8 -4
- data/vendor/eigen/Eigen/src/Core/products/TriangularSolverVector.h +13 -10
- data/vendor/eigen/Eigen/src/Core/util/BlasUtil.h +304 -119
- data/vendor/eigen/Eigen/src/Core/util/ConfigureVectorization.h +512 -0
- data/vendor/eigen/Eigen/src/Core/util/Constants.h +25 -9
- data/vendor/eigen/Eigen/src/Core/util/DisableStupidWarnings.h +26 -3
- data/vendor/eigen/Eigen/src/Core/util/ForwardDeclarations.h +29 -9
- data/vendor/eigen/Eigen/src/Core/util/IndexedViewHelper.h +186 -0
- data/vendor/eigen/Eigen/src/Core/util/IntegralConstant.h +272 -0
- data/vendor/eigen/Eigen/src/Core/util/MKL_support.h +8 -1
- data/vendor/eigen/Eigen/src/Core/util/Macros.h +709 -246
- data/vendor/eigen/Eigen/src/Core/util/Memory.h +222 -52
- data/vendor/eigen/Eigen/src/Core/util/Meta.h +355 -77
- data/vendor/eigen/Eigen/src/Core/util/ReenableStupidWarnings.h +5 -1
- data/vendor/eigen/Eigen/src/Core/util/ReshapedHelper.h +51 -0
- data/vendor/eigen/Eigen/src/Core/util/StaticAssert.h +8 -5
- data/vendor/eigen/Eigen/src/Core/util/SymbolicIndex.h +293 -0
- data/vendor/eigen/Eigen/src/Core/util/XprHelper.h +65 -30
- data/vendor/eigen/Eigen/src/Eigenvalues/ComplexEigenSolver.h +1 -1
- data/vendor/eigen/Eigen/src/Eigenvalues/ComplexSchur.h +7 -4
- data/vendor/eigen/Eigen/src/Eigenvalues/EigenSolver.h +2 -2
- data/vendor/eigen/Eigen/src/Eigenvalues/GeneralizedSelfAdjointEigenSolver.h +1 -1
- data/vendor/eigen/Eigen/src/Eigenvalues/HessenbergDecomposition.h +2 -2
- data/vendor/eigen/Eigen/src/Eigenvalues/MatrixBaseEigenvalues.h +2 -2
- data/vendor/eigen/Eigen/src/Eigenvalues/RealQZ.h +9 -6
- data/vendor/eigen/Eigen/src/Eigenvalues/RealSchur.h +21 -9
- data/vendor/eigen/Eigen/src/Eigenvalues/SelfAdjointEigenSolver.h +77 -43
- data/vendor/eigen/Eigen/src/Eigenvalues/Tridiagonalization.h +20 -15
- data/vendor/eigen/Eigen/src/Geometry/AlignedBox.h +99 -5
- data/vendor/eigen/Eigen/src/Geometry/AngleAxis.h +4 -4
- data/vendor/eigen/Eigen/src/Geometry/EulerAngles.h +3 -3
- data/vendor/eigen/Eigen/src/Geometry/Homogeneous.h +15 -11
- data/vendor/eigen/Eigen/src/Geometry/Hyperplane.h +1 -1
- data/vendor/eigen/Eigen/src/Geometry/OrthoMethods.h +3 -2
- data/vendor/eigen/Eigen/src/Geometry/ParametrizedLine.h +39 -2
- data/vendor/eigen/Eigen/src/Geometry/Quaternion.h +70 -14
- data/vendor/eigen/Eigen/src/Geometry/Rotation2D.h +3 -3
- data/vendor/eigen/Eigen/src/Geometry/Scaling.h +23 -5
- data/vendor/eigen/Eigen/src/Geometry/Transform.h +88 -67
- data/vendor/eigen/Eigen/src/Geometry/Translation.h +6 -12
- data/vendor/eigen/Eigen/src/Geometry/Umeyama.h +1 -1
- data/vendor/eigen/Eigen/src/Geometry/arch/Geometry_SIMD.h +168 -0
- data/vendor/eigen/Eigen/src/Householder/BlockHouseholder.h +9 -2
- data/vendor/eigen/Eigen/src/Householder/Householder.h +8 -4
- data/vendor/eigen/Eigen/src/Householder/HouseholderSequence.h +123 -48
- data/vendor/eigen/Eigen/src/IterativeLinearSolvers/BasicPreconditioners.h +15 -15
- data/vendor/eigen/Eigen/src/IterativeLinearSolvers/BiCGSTAB.h +7 -23
- data/vendor/eigen/Eigen/src/IterativeLinearSolvers/ConjugateGradient.h +5 -22
- data/vendor/eigen/Eigen/src/IterativeLinearSolvers/IncompleteCholesky.h +41 -47
- data/vendor/eigen/Eigen/src/IterativeLinearSolvers/IncompleteLUT.h +51 -60
- data/vendor/eigen/Eigen/src/IterativeLinearSolvers/IterativeSolverBase.h +70 -20
- data/vendor/eigen/Eigen/src/IterativeLinearSolvers/LeastSquareConjugateGradient.h +2 -20
- data/vendor/eigen/Eigen/src/IterativeLinearSolvers/SolveWithGuess.h +11 -9
- data/vendor/eigen/Eigen/src/Jacobi/Jacobi.h +31 -10
- data/vendor/eigen/Eigen/src/KLUSupport/KLUSupport.h +358 -0
- data/vendor/eigen/Eigen/src/LU/Determinant.h +35 -19
- data/vendor/eigen/Eigen/src/LU/FullPivLU.h +29 -43
- data/vendor/eigen/Eigen/src/LU/InverseImpl.h +25 -8
- data/vendor/eigen/Eigen/src/LU/PartialPivLU.h +71 -58
- data/vendor/eigen/Eigen/src/LU/arch/InverseSize4.h +351 -0
- data/vendor/eigen/Eigen/src/OrderingMethods/Amd.h +7 -17
- data/vendor/eigen/Eigen/src/OrderingMethods/Eigen_Colamd.h +297 -277
- data/vendor/eigen/Eigen/src/OrderingMethods/Ordering.h +6 -10
- data/vendor/eigen/Eigen/src/PaStiXSupport/PaStiXSupport.h +1 -1
- data/vendor/eigen/Eigen/src/PardisoSupport/PardisoSupport.h +9 -7
- data/vendor/eigen/Eigen/src/QR/ColPivHouseholderQR.h +41 -20
- data/vendor/eigen/Eigen/src/QR/CompleteOrthogonalDecomposition.h +100 -27
- data/vendor/eigen/Eigen/src/QR/FullPivHouseholderQR.h +59 -22
- data/vendor/eigen/Eigen/src/QR/HouseholderQR.h +48 -23
- data/vendor/eigen/Eigen/src/SPQRSupport/SuiteSparseQRSupport.h +25 -3
- data/vendor/eigen/Eigen/src/SVD/BDCSVD.h +183 -63
- data/vendor/eigen/Eigen/src/SVD/JacobiSVD.h +22 -14
- data/vendor/eigen/Eigen/src/SVD/SVDBase.h +83 -22
- data/vendor/eigen/Eigen/src/SVD/UpperBidiagonalization.h +3 -3
- data/vendor/eigen/Eigen/src/SparseCholesky/SimplicialCholesky.h +17 -9
- data/vendor/eigen/Eigen/src/SparseCholesky/SimplicialCholesky_impl.h +12 -37
- data/vendor/eigen/Eigen/src/SparseCore/AmbiVector.h +3 -2
- data/vendor/eigen/Eigen/src/SparseCore/CompressedStorage.h +16 -0
- data/vendor/eigen/Eigen/src/SparseCore/ConservativeSparseSparseProduct.h +6 -6
- data/vendor/eigen/Eigen/src/SparseCore/SparseAssign.h +81 -27
- data/vendor/eigen/Eigen/src/SparseCore/SparseBlock.h +25 -57
- data/vendor/eigen/Eigen/src/SparseCore/SparseCompressedBase.h +40 -11
- data/vendor/eigen/Eigen/src/SparseCore/SparseCwiseBinaryOp.h +11 -15
- data/vendor/eigen/Eigen/src/SparseCore/SparseCwiseUnaryOp.h +4 -2
- data/vendor/eigen/Eigen/src/SparseCore/SparseDenseProduct.h +30 -8
- data/vendor/eigen/Eigen/src/SparseCore/SparseMatrix.h +126 -11
- data/vendor/eigen/Eigen/src/SparseCore/SparseMatrixBase.h +5 -12
- data/vendor/eigen/Eigen/src/SparseCore/SparseProduct.h +13 -1
- data/vendor/eigen/Eigen/src/SparseCore/SparseRef.h +7 -7
- data/vendor/eigen/Eigen/src/SparseCore/SparseSelfAdjointView.h +5 -2
- data/vendor/eigen/Eigen/src/SparseCore/SparseUtil.h +8 -0
- data/vendor/eigen/Eigen/src/SparseCore/SparseVector.h +1 -1
- data/vendor/eigen/Eigen/src/SparseCore/SparseView.h +1 -0
- data/vendor/eigen/Eigen/src/SparseLU/SparseLU.h +162 -12
- data/vendor/eigen/Eigen/src/SparseLU/SparseLU_Memory.h +1 -1
- data/vendor/eigen/Eigen/src/SparseLU/SparseLU_SupernodalMatrix.h +76 -2
- data/vendor/eigen/Eigen/src/SparseLU/SparseLU_column_dfs.h +2 -2
- data/vendor/eigen/Eigen/src/SparseLU/SparseLU_gemm_kernel.h +1 -1
- data/vendor/eigen/Eigen/src/SparseLU/SparseLU_panel_bmod.h +1 -1
- data/vendor/eigen/Eigen/src/SparseQR/SparseQR.h +19 -6
- data/vendor/eigen/Eigen/src/StlSupport/StdDeque.h +2 -12
- data/vendor/eigen/Eigen/src/StlSupport/StdList.h +2 -2
- data/vendor/eigen/Eigen/src/StlSupport/StdVector.h +2 -2
- data/vendor/eigen/Eigen/src/SuperLUSupport/SuperLUSupport.h +6 -8
- data/vendor/eigen/Eigen/src/UmfPackSupport/UmfPackSupport.h +175 -39
- data/vendor/eigen/Eigen/src/misc/lapacke.h +5 -4
- data/vendor/eigen/Eigen/src/plugins/ArrayCwiseBinaryOps.h +28 -2
- data/vendor/eigen/Eigen/src/plugins/ArrayCwiseUnaryOps.h +155 -11
- data/vendor/eigen/Eigen/src/plugins/BlockMethods.h +626 -242
- data/vendor/eigen/Eigen/src/plugins/CommonCwiseUnaryOps.h +14 -0
- data/vendor/eigen/Eigen/src/plugins/IndexedViewMethods.h +262 -0
- data/vendor/eigen/Eigen/src/plugins/MatrixCwiseBinaryOps.h +4 -4
- data/vendor/eigen/Eigen/src/plugins/MatrixCwiseUnaryOps.h +10 -0
- data/vendor/eigen/Eigen/src/plugins/ReshapedMethods.h +149 -0
- data/vendor/eigen/README.md +2 -0
- data/vendor/eigen/bench/btl/README +1 -1
- data/vendor/eigen/bench/tensors/README +6 -7
- data/vendor/eigen/ci/README.md +56 -0
- data/vendor/eigen/demos/mix_eigen_and_c/README +1 -1
- data/vendor/eigen/unsupported/Eigen/CXX11/src/Tensor/README.md +213 -158
- data/vendor/eigen/unsupported/README.txt +1 -1
- data/vendor/tomotopy/README.kr.rst +21 -0
- data/vendor/tomotopy/README.rst +20 -0
- data/vendor/tomotopy/src/Labeling/FoRelevance.cpp +2 -2
- data/vendor/tomotopy/src/Labeling/Phraser.hpp +1 -1
- data/vendor/tomotopy/src/TopicModel/CTModel.hpp +2 -1
- data/vendor/tomotopy/src/TopicModel/DMRModel.hpp +2 -1
- data/vendor/tomotopy/src/TopicModel/DTModel.hpp +1 -1
- data/vendor/tomotopy/src/TopicModel/GDMRModel.hpp +2 -2
- data/vendor/tomotopy/src/TopicModel/HDP.h +1 -0
- data/vendor/tomotopy/src/TopicModel/HDPModel.hpp +53 -2
- data/vendor/tomotopy/src/TopicModel/HLDAModel.hpp +1 -1
- data/vendor/tomotopy/src/TopicModel/HPAModel.hpp +1 -0
- data/vendor/tomotopy/src/TopicModel/LDACVB0Model.hpp +2 -2
- data/vendor/tomotopy/src/TopicModel/LDAModel.hpp +16 -5
- data/vendor/tomotopy/src/TopicModel/LLDAModel.hpp +1 -0
- data/vendor/tomotopy/src/TopicModel/MGLDAModel.hpp +1 -0
- data/vendor/tomotopy/src/TopicModel/PLDAModel.hpp +1 -0
- data/vendor/tomotopy/src/TopicModel/PT.h +3 -1
- data/vendor/tomotopy/src/TopicModel/PTModel.hpp +31 -1
- data/vendor/tomotopy/src/TopicModel/SLDAModel.hpp +2 -2
- data/vendor/tomotopy/src/TopicModel/TopicModel.hpp +7 -5
- data/vendor/tomotopy/src/Utils/EigenAddonOps.hpp +36 -1
- data/vendor/tomotopy/src/Utils/exception.h +6 -0
- data/vendor/tomotopy/src/Utils/sample.hpp +14 -12
- data/vendor/tomotopy/src/Utils/sse_gamma.h +0 -3
- metadata +60 -14
- data/vendor/eigen/Eigen/CMakeLists.txt +0 -19
- data/vendor/eigen/Eigen/src/Core/arch/CUDA/Half.h +0 -674
- data/vendor/eigen/Eigen/src/Core/arch/CUDA/PacketMath.h +0 -333
- data/vendor/eigen/Eigen/src/Core/arch/CUDA/PacketMathHalf.h +0 -1124
- data/vendor/eigen/Eigen/src/Core/arch/CUDA/TypeCasting.h +0 -212
- data/vendor/eigen/Eigen/src/Geometry/arch/Geometry_SSE.h +0 -161
- data/vendor/eigen/Eigen/src/LU/arch/Inverse_SSE.h +0 -338
@@ -0,0 +1,49 @@
|
|
1
|
+
// This file is part of Eigen, a lightweight C++ template library
|
2
|
+
// for linear algebra.
|
3
|
+
//
|
4
|
+
// Copyright (C) 2020, Arm Limited and Contributors
|
5
|
+
//
|
6
|
+
// This Source Code Form is subject to the terms of the Mozilla
|
7
|
+
// Public License v. 2.0. If a copy of the MPL was not distributed
|
8
|
+
// with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
|
9
|
+
|
10
|
+
#ifndef EIGEN_TYPE_CASTING_SVE_H
|
11
|
+
#define EIGEN_TYPE_CASTING_SVE_H
|
12
|
+
|
13
|
+
namespace Eigen {
|
14
|
+
namespace internal {
|
15
|
+
|
16
|
+
template <>
|
17
|
+
struct type_casting_traits<float, numext::int32_t> {
|
18
|
+
enum { VectorizedCast = 1, SrcCoeffRatio = 1, TgtCoeffRatio = 1 };
|
19
|
+
};
|
20
|
+
|
21
|
+
template <>
|
22
|
+
struct type_casting_traits<numext::int32_t, float> {
|
23
|
+
enum { VectorizedCast = 1, SrcCoeffRatio = 1, TgtCoeffRatio = 1 };
|
24
|
+
};
|
25
|
+
|
26
|
+
template <>
|
27
|
+
EIGEN_STRONG_INLINE PacketXf pcast<PacketXi, PacketXf>(const PacketXi& a) {
|
28
|
+
return svcvt_f32_s32_z(svptrue_b32(), a);
|
29
|
+
}
|
30
|
+
|
31
|
+
template <>
|
32
|
+
EIGEN_STRONG_INLINE PacketXi pcast<PacketXf, PacketXi>(const PacketXf& a) {
|
33
|
+
return svcvt_s32_f32_z(svptrue_b32(), a);
|
34
|
+
}
|
35
|
+
|
36
|
+
template <>
|
37
|
+
EIGEN_STRONG_INLINE PacketXf preinterpret<PacketXf, PacketXi>(const PacketXi& a) {
|
38
|
+
return svreinterpret_f32_s32(a);
|
39
|
+
}
|
40
|
+
|
41
|
+
template <>
|
42
|
+
EIGEN_STRONG_INLINE PacketXi preinterpret<PacketXi, PacketXf>(const PacketXf& a) {
|
43
|
+
return svreinterpret_s32_f32(a);
|
44
|
+
}
|
45
|
+
|
46
|
+
} // namespace internal
|
47
|
+
} // namespace Eigen
|
48
|
+
|
49
|
+
#endif // EIGEN_TYPE_CASTING_SVE_H
|
@@ -0,0 +1,232 @@
|
|
1
|
+
// This file is part of Eigen, a lightweight C++ template library
|
2
|
+
// for linear algebra.
|
3
|
+
//
|
4
|
+
// Mehdi Goli Codeplay Software Ltd.
|
5
|
+
// Ralph Potter Codeplay Software Ltd.
|
6
|
+
// Luke Iwanski Codeplay Software Ltd.
|
7
|
+
// Contact: <eigen@codeplay.com>
|
8
|
+
//
|
9
|
+
// This Source Code Form is subject to the terms of the Mozilla
|
10
|
+
// Public License v. 2.0. If a copy of the MPL was not distributed
|
11
|
+
// with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
|
12
|
+
|
13
|
+
/*****************************************************************
|
14
|
+
* InteropHeaders.h
|
15
|
+
*
|
16
|
+
* \brief:
|
17
|
+
* InteropHeaders
|
18
|
+
*
|
19
|
+
*****************************************************************/
|
20
|
+
|
21
|
+
#ifndef EIGEN_INTEROP_HEADERS_SYCL_H
|
22
|
+
#define EIGEN_INTEROP_HEADERS_SYCL_H
|
23
|
+
|
24
|
+
namespace Eigen {
|
25
|
+
|
26
|
+
#if !defined(EIGEN_DONT_VECTORIZE_SYCL)
|
27
|
+
|
28
|
+
namespace internal {
|
29
|
+
|
30
|
+
template <int has_blend, int lengths>
|
31
|
+
struct sycl_packet_traits : default_packet_traits {
|
32
|
+
enum {
|
33
|
+
Vectorizable = 1,
|
34
|
+
AlignedOnScalar = 1,
|
35
|
+
size = lengths,
|
36
|
+
HasHalfPacket = 0,
|
37
|
+
HasDiv = 1,
|
38
|
+
HasLog = 1,
|
39
|
+
HasExp = 1,
|
40
|
+
HasSqrt = 1,
|
41
|
+
HasRsqrt = 1,
|
42
|
+
HasSin = 1,
|
43
|
+
HasCos = 1,
|
44
|
+
HasTan = 1,
|
45
|
+
HasASin = 1,
|
46
|
+
HasACos = 1,
|
47
|
+
HasATan = 1,
|
48
|
+
HasSinh = 1,
|
49
|
+
HasCosh = 1,
|
50
|
+
HasTanh = 1,
|
51
|
+
HasLGamma = 0,
|
52
|
+
HasDiGamma = 0,
|
53
|
+
HasZeta = 0,
|
54
|
+
HasPolygamma = 0,
|
55
|
+
HasErf = 0,
|
56
|
+
HasErfc = 0,
|
57
|
+
HasNdtri = 0,
|
58
|
+
HasIGamma = 0,
|
59
|
+
HasIGammac = 0,
|
60
|
+
HasBetaInc = 0,
|
61
|
+
HasBlend = has_blend,
|
62
|
+
// This flag is used to indicate whether packet comparison is supported.
|
63
|
+
// pcmp_eq, pcmp_lt and pcmp_le should be defined for it to be true.
|
64
|
+
HasCmp = 1,
|
65
|
+
HasMax = 1,
|
66
|
+
HasMin = 1,
|
67
|
+
HasMul = 1,
|
68
|
+
HasAdd = 1,
|
69
|
+
HasFloor = 1,
|
70
|
+
HasRound = 1,
|
71
|
+
HasRint = 1,
|
72
|
+
HasLog1p = 1,
|
73
|
+
HasExpm1 = 1,
|
74
|
+
HasCeil = 1,
|
75
|
+
};
|
76
|
+
};
|
77
|
+
|
78
|
+
#ifdef SYCL_DEVICE_ONLY
|
79
|
+
#define SYCL_PACKET_TRAITS(packet_type, has_blend, unpacket_type, lengths) \
|
80
|
+
template <> \
|
81
|
+
struct packet_traits<unpacket_type> \
|
82
|
+
: sycl_packet_traits<has_blend, lengths> { \
|
83
|
+
typedef packet_type type; \
|
84
|
+
typedef packet_type half; \
|
85
|
+
};
|
86
|
+
|
87
|
+
SYCL_PACKET_TRAITS(cl::sycl::cl_float4, 1, float, 4)
|
88
|
+
SYCL_PACKET_TRAITS(cl::sycl::cl_float4, 1, const float, 4)
|
89
|
+
SYCL_PACKET_TRAITS(cl::sycl::cl_double2, 0, double, 2)
|
90
|
+
SYCL_PACKET_TRAITS(cl::sycl::cl_double2, 0, const double, 2)
|
91
|
+
#undef SYCL_PACKET_TRAITS
|
92
|
+
|
93
|
+
// Make sure this is only available when targeting a GPU: we don't want to
|
94
|
+
// introduce conflicts between these packet_traits definitions and the ones
|
95
|
+
// we'll use on the host side (SSE, AVX, ...)
|
96
|
+
#define SYCL_ARITHMETIC(packet_type) \
|
97
|
+
template <> \
|
98
|
+
struct is_arithmetic<packet_type> { \
|
99
|
+
enum { value = true }; \
|
100
|
+
};
|
101
|
+
SYCL_ARITHMETIC(cl::sycl::cl_float4)
|
102
|
+
SYCL_ARITHMETIC(cl::sycl::cl_double2)
|
103
|
+
#undef SYCL_ARITHMETIC
|
104
|
+
|
105
|
+
#define SYCL_UNPACKET_TRAITS(packet_type, unpacket_type, lengths) \
|
106
|
+
template <> \
|
107
|
+
struct unpacket_traits<packet_type> { \
|
108
|
+
typedef unpacket_type type; \
|
109
|
+
enum { size = lengths, vectorizable = true, alignment = Aligned16 }; \
|
110
|
+
typedef packet_type half; \
|
111
|
+
};
|
112
|
+
SYCL_UNPACKET_TRAITS(cl::sycl::cl_float4, float, 4)
|
113
|
+
SYCL_UNPACKET_TRAITS(cl::sycl::cl_double2, double, 2)
|
114
|
+
|
115
|
+
#undef SYCL_UNPACKET_TRAITS
|
116
|
+
#endif
|
117
|
+
|
118
|
+
} // end namespace internal
|
119
|
+
|
120
|
+
#endif
|
121
|
+
|
122
|
+
namespace TensorSycl {
|
123
|
+
namespace internal {
|
124
|
+
|
125
|
+
template <typename PacketReturnType, int PacketSize>
|
126
|
+
struct PacketWrapper;
|
127
|
+
// This function should never get called on the device
|
128
|
+
#ifndef SYCL_DEVICE_ONLY
|
129
|
+
template <typename PacketReturnType, int PacketSize>
|
130
|
+
struct PacketWrapper {
|
131
|
+
typedef typename ::Eigen::internal::unpacket_traits<PacketReturnType>::type
|
132
|
+
Scalar;
|
133
|
+
template <typename Index>
|
134
|
+
EIGEN_DEVICE_FUNC static Scalar scalarize(Index, PacketReturnType &) {
|
135
|
+
eigen_assert(false && "THERE IS NO PACKETIZE VERSION FOR THE CHOSEN TYPE");
|
136
|
+
abort();
|
137
|
+
}
|
138
|
+
EIGEN_DEVICE_FUNC static PacketReturnType convert_to_packet_type(Scalar in,
|
139
|
+
Scalar) {
|
140
|
+
return ::Eigen::internal::template plset<PacketReturnType>(in);
|
141
|
+
}
|
142
|
+
EIGEN_DEVICE_FUNC static void set_packet(PacketReturnType, Scalar *) {
|
143
|
+
eigen_assert(false && "THERE IS NO PACKETIZE VERSION FOR THE CHOSEN TYPE");
|
144
|
+
abort();
|
145
|
+
}
|
146
|
+
};
|
147
|
+
|
148
|
+
#elif defined(SYCL_DEVICE_ONLY)
|
149
|
+
template <typename PacketReturnType>
|
150
|
+
struct PacketWrapper<PacketReturnType, 4> {
|
151
|
+
typedef typename ::Eigen::internal::unpacket_traits<PacketReturnType>::type
|
152
|
+
Scalar;
|
153
|
+
template <typename Index>
|
154
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static Scalar scalarize(Index index, PacketReturnType &in) {
|
155
|
+
switch (index) {
|
156
|
+
case 0:
|
157
|
+
return in.x();
|
158
|
+
case 1:
|
159
|
+
return in.y();
|
160
|
+
case 2:
|
161
|
+
return in.z();
|
162
|
+
case 3:
|
163
|
+
return in.w();
|
164
|
+
default:
|
165
|
+
//INDEX MUST BE BETWEEN 0 and 3.There is no abort function in SYCL kernel. so we cannot use abort here.
|
166
|
+
// The code will never reach here
|
167
|
+
__builtin_unreachable();
|
168
|
+
}
|
169
|
+
__builtin_unreachable();
|
170
|
+
}
|
171
|
+
|
172
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static PacketReturnType convert_to_packet_type(
|
173
|
+
Scalar in, Scalar other) {
|
174
|
+
return PacketReturnType(in, other, other, other);
|
175
|
+
}
|
176
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static void set_packet(PacketReturnType &lhs, Scalar *rhs) {
|
177
|
+
lhs = PacketReturnType(rhs[0], rhs[1], rhs[2], rhs[3]);
|
178
|
+
}
|
179
|
+
};
|
180
|
+
|
181
|
+
template <typename PacketReturnType>
|
182
|
+
struct PacketWrapper<PacketReturnType, 1> {
|
183
|
+
typedef typename ::Eigen::internal::unpacket_traits<PacketReturnType>::type
|
184
|
+
Scalar;
|
185
|
+
template <typename Index>
|
186
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static Scalar scalarize(Index, PacketReturnType &in) {
|
187
|
+
return in;
|
188
|
+
}
|
189
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static PacketReturnType convert_to_packet_type(Scalar in,
|
190
|
+
Scalar) {
|
191
|
+
return PacketReturnType(in);
|
192
|
+
}
|
193
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static void set_packet(PacketReturnType &lhs, Scalar *rhs) {
|
194
|
+
lhs = rhs[0];
|
195
|
+
}
|
196
|
+
};
|
197
|
+
|
198
|
+
template <typename PacketReturnType>
|
199
|
+
struct PacketWrapper<PacketReturnType, 2> {
|
200
|
+
typedef typename ::Eigen::internal::unpacket_traits<PacketReturnType>::type
|
201
|
+
Scalar;
|
202
|
+
template <typename Index>
|
203
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static Scalar scalarize(Index index, PacketReturnType &in) {
|
204
|
+
switch (index) {
|
205
|
+
case 0:
|
206
|
+
return in.x();
|
207
|
+
case 1:
|
208
|
+
return in.y();
|
209
|
+
default:
|
210
|
+
//INDEX MUST BE BETWEEN 0 and 1.There is no abort function in SYCL kernel. so we cannot use abort here.
|
211
|
+
// The code will never reach here
|
212
|
+
__builtin_unreachable();
|
213
|
+
}
|
214
|
+
__builtin_unreachable();
|
215
|
+
}
|
216
|
+
|
217
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static PacketReturnType convert_to_packet_type(
|
218
|
+
Scalar in, Scalar other) {
|
219
|
+
return PacketReturnType(in, other);
|
220
|
+
}
|
221
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE static void set_packet(PacketReturnType &lhs, Scalar *rhs) {
|
222
|
+
lhs = PacketReturnType(rhs[0], rhs[1]);
|
223
|
+
}
|
224
|
+
};
|
225
|
+
|
226
|
+
#endif
|
227
|
+
|
228
|
+
} // end namespace internal
|
229
|
+
} // end namespace TensorSycl
|
230
|
+
} // end namespace Eigen
|
231
|
+
|
232
|
+
#endif // EIGEN_INTEROP_HEADERS_SYCL_H
|
@@ -0,0 +1,301 @@
|
|
1
|
+
// This file is part of Eigen, a lightweight C++ template library
|
2
|
+
// for linear algebra.
|
3
|
+
//
|
4
|
+
// Mehdi Goli Codeplay Software Ltd.
|
5
|
+
// Ralph Potter Codeplay Software Ltd.
|
6
|
+
// Luke Iwanski Codeplay Software Ltd.
|
7
|
+
// Contact: <eigen@codeplay.com>
|
8
|
+
//
|
9
|
+
// This Source Code Form is subject to the terms of the Mozilla
|
10
|
+
// Public License v. 2.0. If a copy of the MPL was not distributed
|
11
|
+
// with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
|
12
|
+
|
13
|
+
/*****************************************************************
|
14
|
+
* MathFunctions.h
|
15
|
+
*
|
16
|
+
* \brief:
|
17
|
+
* MathFunctions
|
18
|
+
*
|
19
|
+
*****************************************************************/
|
20
|
+
|
21
|
+
#ifndef EIGEN_MATH_FUNCTIONS_SYCL_H
|
22
|
+
#define EIGEN_MATH_FUNCTIONS_SYCL_H
|
23
|
+
namespace Eigen {
|
24
|
+
|
25
|
+
namespace internal {
|
26
|
+
|
27
|
+
// Make sure this is only available when targeting a GPU: we don't want to
|
28
|
+
// introduce conflicts between these packet_traits definitions and the ones
|
29
|
+
// we'll use on the host side (SSE, AVX, ...)
|
30
|
+
#if defined(SYCL_DEVICE_ONLY)
|
31
|
+
#define SYCL_PLOG(packet_type) \
|
32
|
+
template <> \
|
33
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type plog<packet_type>( \
|
34
|
+
const packet_type& a) { \
|
35
|
+
return cl::sycl::log(a); \
|
36
|
+
}
|
37
|
+
|
38
|
+
SYCL_PLOG(cl::sycl::cl_float4)
|
39
|
+
SYCL_PLOG(cl::sycl::cl_double2)
|
40
|
+
#undef SYCL_PLOG
|
41
|
+
|
42
|
+
#define SYCL_PLOG1P(packet_type) \
|
43
|
+
template <> \
|
44
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type plog1p<packet_type>( \
|
45
|
+
const packet_type& a) { \
|
46
|
+
return cl::sycl::log1p(a); \
|
47
|
+
}
|
48
|
+
|
49
|
+
SYCL_PLOG1P(cl::sycl::cl_float4)
|
50
|
+
SYCL_PLOG1P(cl::sycl::cl_double2)
|
51
|
+
#undef SYCL_PLOG1P
|
52
|
+
|
53
|
+
#define SYCL_PLOG10(packet_type) \
|
54
|
+
template <> \
|
55
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type plog10<packet_type>( \
|
56
|
+
const packet_type& a) { \
|
57
|
+
return cl::sycl::log10(a); \
|
58
|
+
}
|
59
|
+
|
60
|
+
SYCL_PLOG10(cl::sycl::cl_float4)
|
61
|
+
SYCL_PLOG10(cl::sycl::cl_double2)
|
62
|
+
#undef SYCL_PLOG10
|
63
|
+
|
64
|
+
#define SYCL_PEXP(packet_type) \
|
65
|
+
template <> \
|
66
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pexp<packet_type>( \
|
67
|
+
const packet_type& a) { \
|
68
|
+
return cl::sycl::exp(a); \
|
69
|
+
}
|
70
|
+
|
71
|
+
SYCL_PEXP(cl::sycl::cl_float4)
|
72
|
+
SYCL_PEXP(cl::sycl::cl_float)
|
73
|
+
SYCL_PEXP(cl::sycl::cl_double2)
|
74
|
+
#undef SYCL_PEXP
|
75
|
+
|
76
|
+
#define SYCL_PEXPM1(packet_type) \
|
77
|
+
template <> \
|
78
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pexpm1<packet_type>( \
|
79
|
+
const packet_type& a) { \
|
80
|
+
return cl::sycl::expm1(a); \
|
81
|
+
}
|
82
|
+
|
83
|
+
SYCL_PEXPM1(cl::sycl::cl_float4)
|
84
|
+
SYCL_PEXPM1(cl::sycl::cl_double2)
|
85
|
+
#undef SYCL_PEXPM1
|
86
|
+
|
87
|
+
#define SYCL_PSQRT(packet_type) \
|
88
|
+
template <> \
|
89
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type psqrt<packet_type>( \
|
90
|
+
const packet_type& a) { \
|
91
|
+
return cl::sycl::sqrt(a); \
|
92
|
+
}
|
93
|
+
|
94
|
+
SYCL_PSQRT(cl::sycl::cl_float4)
|
95
|
+
SYCL_PSQRT(cl::sycl::cl_double2)
|
96
|
+
#undef SYCL_PSQRT
|
97
|
+
|
98
|
+
#define SYCL_PRSQRT(packet_type) \
|
99
|
+
template <> \
|
100
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type prsqrt<packet_type>( \
|
101
|
+
const packet_type& a) { \
|
102
|
+
return cl::sycl::rsqrt(a); \
|
103
|
+
}
|
104
|
+
|
105
|
+
SYCL_PRSQRT(cl::sycl::cl_float4)
|
106
|
+
SYCL_PRSQRT(cl::sycl::cl_double2)
|
107
|
+
#undef SYCL_PRSQRT
|
108
|
+
|
109
|
+
/** \internal \returns the hyperbolic sine of \a a (coeff-wise) */
|
110
|
+
#define SYCL_PSIN(packet_type) \
|
111
|
+
template <> \
|
112
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type psin<packet_type>( \
|
113
|
+
const packet_type& a) { \
|
114
|
+
return cl::sycl::sin(a); \
|
115
|
+
}
|
116
|
+
|
117
|
+
SYCL_PSIN(cl::sycl::cl_float4)
|
118
|
+
SYCL_PSIN(cl::sycl::cl_double2)
|
119
|
+
#undef SYCL_PSIN
|
120
|
+
|
121
|
+
/** \internal \returns the hyperbolic cosine of \a a (coeff-wise) */
|
122
|
+
#define SYCL_PCOS(packet_type) \
|
123
|
+
template <> \
|
124
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pcos<packet_type>( \
|
125
|
+
const packet_type& a) { \
|
126
|
+
return cl::sycl::cos(a); \
|
127
|
+
}
|
128
|
+
|
129
|
+
SYCL_PCOS(cl::sycl::cl_float4)
|
130
|
+
SYCL_PCOS(cl::sycl::cl_double2)
|
131
|
+
#undef SYCL_PCOS
|
132
|
+
|
133
|
+
/** \internal \returns the hyperbolic tan of \a a (coeff-wise) */
|
134
|
+
#define SYCL_PTAN(packet_type) \
|
135
|
+
template <> \
|
136
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type ptan<packet_type>( \
|
137
|
+
const packet_type& a) { \
|
138
|
+
return cl::sycl::tan(a); \
|
139
|
+
}
|
140
|
+
|
141
|
+
SYCL_PTAN(cl::sycl::cl_float4)
|
142
|
+
SYCL_PTAN(cl::sycl::cl_double2)
|
143
|
+
#undef SYCL_PTAN
|
144
|
+
|
145
|
+
/** \internal \returns the hyperbolic sine of \a a (coeff-wise) */
|
146
|
+
#define SYCL_PASIN(packet_type) \
|
147
|
+
template <> \
|
148
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pasin<packet_type>( \
|
149
|
+
const packet_type& a) { \
|
150
|
+
return cl::sycl::asin(a); \
|
151
|
+
}
|
152
|
+
|
153
|
+
SYCL_PASIN(cl::sycl::cl_float4)
|
154
|
+
SYCL_PASIN(cl::sycl::cl_double2)
|
155
|
+
#undef SYCL_PASIN
|
156
|
+
|
157
|
+
/** \internal \returns the hyperbolic cosine of \a a (coeff-wise) */
|
158
|
+
#define SYCL_PACOS(packet_type) \
|
159
|
+
template <> \
|
160
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pacos<packet_type>( \
|
161
|
+
const packet_type& a) { \
|
162
|
+
return cl::sycl::acos(a); \
|
163
|
+
}
|
164
|
+
|
165
|
+
SYCL_PACOS(cl::sycl::cl_float4)
|
166
|
+
SYCL_PACOS(cl::sycl::cl_double2)
|
167
|
+
#undef SYCL_PACOS
|
168
|
+
|
169
|
+
/** \internal \returns the hyperbolic tan of \a a (coeff-wise) */
|
170
|
+
#define SYCL_PATAN(packet_type) \
|
171
|
+
template <> \
|
172
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type patan<packet_type>( \
|
173
|
+
const packet_type& a) { \
|
174
|
+
return cl::sycl::atan(a); \
|
175
|
+
}
|
176
|
+
|
177
|
+
SYCL_PATAN(cl::sycl::cl_float4)
|
178
|
+
SYCL_PATAN(cl::sycl::cl_double2)
|
179
|
+
#undef SYCL_PATAN
|
180
|
+
|
181
|
+
/** \internal \returns the hyperbolic sine of \a a (coeff-wise) */
|
182
|
+
#define SYCL_PSINH(packet_type) \
|
183
|
+
template <> \
|
184
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type psinh<packet_type>( \
|
185
|
+
const packet_type& a) { \
|
186
|
+
return cl::sycl::sinh(a); \
|
187
|
+
}
|
188
|
+
|
189
|
+
SYCL_PSINH(cl::sycl::cl_float4)
|
190
|
+
SYCL_PSINH(cl::sycl::cl_double2)
|
191
|
+
#undef SYCL_PSINH
|
192
|
+
|
193
|
+
/** \internal \returns the hyperbolic cosine of \a a (coeff-wise) */
|
194
|
+
#define SYCL_PCOSH(packet_type) \
|
195
|
+
template <> \
|
196
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pcosh<packet_type>( \
|
197
|
+
const packet_type& a) { \
|
198
|
+
return cl::sycl::cosh(a); \
|
199
|
+
}
|
200
|
+
|
201
|
+
SYCL_PCOSH(cl::sycl::cl_float4)
|
202
|
+
SYCL_PCOSH(cl::sycl::cl_double2)
|
203
|
+
#undef SYCL_PCOSH
|
204
|
+
|
205
|
+
/** \internal \returns the hyperbolic tan of \a a (coeff-wise) */
|
206
|
+
#define SYCL_PTANH(packet_type) \
|
207
|
+
template <> \
|
208
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type ptanh<packet_type>( \
|
209
|
+
const packet_type& a) { \
|
210
|
+
return cl::sycl::tanh(a); \
|
211
|
+
}
|
212
|
+
|
213
|
+
SYCL_PTANH(cl::sycl::cl_float4)
|
214
|
+
SYCL_PTANH(cl::sycl::cl_double2)
|
215
|
+
#undef SYCL_PTANH
|
216
|
+
|
217
|
+
#define SYCL_PCEIL(packet_type) \
|
218
|
+
template <> \
|
219
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pceil<packet_type>( \
|
220
|
+
const packet_type& a) { \
|
221
|
+
return cl::sycl::ceil(a); \
|
222
|
+
}
|
223
|
+
|
224
|
+
SYCL_PCEIL(cl::sycl::cl_float4)
|
225
|
+
SYCL_PCEIL(cl::sycl::cl_double2)
|
226
|
+
#undef SYCL_PCEIL
|
227
|
+
|
228
|
+
#define SYCL_PROUND(packet_type) \
|
229
|
+
template <> \
|
230
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pround<packet_type>( \
|
231
|
+
const packet_type& a) { \
|
232
|
+
return cl::sycl::round(a); \
|
233
|
+
}
|
234
|
+
|
235
|
+
SYCL_PROUND(cl::sycl::cl_float4)
|
236
|
+
SYCL_PROUND(cl::sycl::cl_double2)
|
237
|
+
#undef SYCL_PROUND
|
238
|
+
|
239
|
+
#define SYCL_PRINT(packet_type) \
|
240
|
+
template <> \
|
241
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type print<packet_type>( \
|
242
|
+
const packet_type& a) { \
|
243
|
+
return cl::sycl::rint(a); \
|
244
|
+
}
|
245
|
+
|
246
|
+
SYCL_PRINT(cl::sycl::cl_float4)
|
247
|
+
SYCL_PRINT(cl::sycl::cl_double2)
|
248
|
+
#undef SYCL_PRINT
|
249
|
+
|
250
|
+
#define SYCL_FLOOR(packet_type) \
|
251
|
+
template <> \
|
252
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pfloor<packet_type>( \
|
253
|
+
const packet_type& a) { \
|
254
|
+
return cl::sycl::floor(a); \
|
255
|
+
}
|
256
|
+
|
257
|
+
SYCL_FLOOR(cl::sycl::cl_float4)
|
258
|
+
SYCL_FLOOR(cl::sycl::cl_double2)
|
259
|
+
#undef SYCL_FLOOR
|
260
|
+
|
261
|
+
#define SYCL_PMIN(packet_type, expr) \
|
262
|
+
template <> \
|
263
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pmin<packet_type>( \
|
264
|
+
const packet_type& a, const packet_type& b) { \
|
265
|
+
return expr; \
|
266
|
+
}
|
267
|
+
|
268
|
+
SYCL_PMIN(cl::sycl::cl_float4, cl::sycl::fmin(a, b))
|
269
|
+
SYCL_PMIN(cl::sycl::cl_double2, cl::sycl::fmin(a, b))
|
270
|
+
#undef SYCL_PMIN
|
271
|
+
|
272
|
+
#define SYCL_PMAX(packet_type, expr) \
|
273
|
+
template <> \
|
274
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pmax<packet_type>( \
|
275
|
+
const packet_type& a, const packet_type& b) { \
|
276
|
+
return expr; \
|
277
|
+
}
|
278
|
+
|
279
|
+
SYCL_PMAX(cl::sycl::cl_float4, cl::sycl::fmax(a, b))
|
280
|
+
SYCL_PMAX(cl::sycl::cl_double2, cl::sycl::fmax(a, b))
|
281
|
+
#undef SYCL_PMAX
|
282
|
+
|
283
|
+
#define SYCL_PLDEXP(packet_type) \
|
284
|
+
template <> \
|
285
|
+
EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE packet_type pldexp( \
|
286
|
+
const packet_type& a, const packet_type& exponent) { \
|
287
|
+
return cl::sycl::ldexp( \
|
288
|
+
a, exponent.template convert<cl::sycl::cl_int, \
|
289
|
+
cl::sycl::rounding_mode::automatic>()); \
|
290
|
+
}
|
291
|
+
|
292
|
+
SYCL_PLDEXP(cl::sycl::cl_float4)
|
293
|
+
SYCL_PLDEXP(cl::sycl::cl_double2)
|
294
|
+
#undef SYCL_PLDEXP
|
295
|
+
|
296
|
+
#endif
|
297
|
+
} // end namespace internal
|
298
|
+
|
299
|
+
} // end namespace Eigen
|
300
|
+
|
301
|
+
#endif // EIGEN_MATH_FUNCTIONS_SYCL_H
|