/external/tensorflow/tensorflow/stream_executor/ |
D | blas.h | 62 namespace blas { 423 virtual bool DoBlasGbmv(Stream *stream, blas::Transpose trans, uint64 m, 428 virtual bool DoBlasGbmv(Stream *stream, blas::Transpose trans, uint64 m, 433 virtual bool DoBlasGbmv(Stream *stream, blas::Transpose trans, uint64 m, 440 virtual bool DoBlasGbmv(Stream *stream, blas::Transpose trans, uint64 m, 459 virtual bool DoBlasGemv(Stream *stream, blas::Transpose trans, uint64 m, 463 virtual bool DoBlasGemv(Stream *stream, blas::Transpose trans, uint64 m, 467 virtual bool DoBlasGemv(Stream *stream, blas::Transpose trans, uint64 m, 473 virtual bool DoBlasGemv(Stream *stream, blas::Transpose trans, uint64 m, 481 Stream *stream, blas::Transpose trans, uint64 m, uint64 n, float alpha, [all …]
|
D | stream.cc | 74 string ToVlogString(blas::Transpose t) { return blas::TransposeString(t); } in ToVlogString() 76 string ToVlogString(blas::UpperLower ul) { return blas::UpperLowerString(ul); } in ToVlogString() 78 string ToVlogString(blas::Diagonal d) { return blas::DiagonalString(d); } in ToVlogString() 80 string ToVlogString(blas::Side s) { return blas::SideString(s); } in ToVlogString() 82 string ToVlogString(blas::ComputationType ty) { in ToVlogString() 83 return blas::ComputationTypeString(ty); in ToVlogString() 1882 bool (blas::BlasSupport::*blas_func)(Stream *, Args...), in operator ()() 1890 bool (blas::BlasSupport::*blas_func)(Stream *, Args...), 1896 Stream *stream, bool (blas::BlasSupport::*blas_func)(Stream *, Args...), in Run() 1900 if (blas::BlasSupport *blas = stream->parent_->AsBlas()) { in Run() local [all …]
|
D | stream.h | 895 Stream &ThenBlasGbmv(blas::Transpose trans, uint64 m, uint64 n, uint64 kl, 899 Stream &ThenBlasGbmv(blas::Transpose trans, uint64 m, uint64 n, uint64 kl, 903 Stream &ThenBlasGbmv(blas::Transpose trans, uint64 m, uint64 n, uint64 kl, 909 Stream &ThenBlasGbmv(blas::Transpose trans, uint64 m, uint64 n, uint64 kl, 917 Stream &ThenBlasGemv(blas::Transpose trans, uint64 m, uint64 n, float alpha, 921 Stream &ThenBlasGemv(blas::Transpose trans, uint64 m, uint64 n, double alpha, 925 Stream &ThenBlasGemv(blas::Transpose trans, uint64 m, uint64 n, 931 Stream &ThenBlasGemv(blas::Transpose trans, uint64 m, uint64 n, 938 Stream &ThenBlasGemvWithProfiling(blas::Transpose trans, uint64 m, uint64 n, 943 blas::ProfileResult *output_profile_result); [all …]
|
D | BUILD | 60 "blas.h", 115 "blas.h", 180 "blas.h", 216 ":blas", 246 "blas.h", 326 name = "blas", 327 srcs = ["blas.cc"], 328 hdrs = ["blas.h"], 378 "blas.h", 433 ":blas", [all …]
|
D | plugin_registry.cc | 44 blas(kNullPlugin), dnn(kNullPlugin), fft(kNullPlugin), rng(kNullPlugin) { } in DefaultFactories() 125 default_factories_[platform_id].blas = plugin_id; in SetDefaultFactory() 150 return factories.blas.find(plugin_id) != factories.blas.end(); in HasFactory() 242 EMIT_PLUGIN_SPECIALIZATIONS(BlasFactory, blas, "BLAS");
|
D | plugin_registry.h | 53 typedef blas::BlasSupport* (*BlasFactory)(internal::StreamExecutorInterface*); 108 std::map<PluginId, BlasFactory> blas; member 118 PluginId blas, dnn, fft, rng; member
|
D | plugin.cc | 33 PluginConfig& PluginConfig::SetBlas(PluginId blas) { in SetBlas() argument 34 blas_ = blas; in SetBlas()
|
D | plugin.h | 71 PluginConfig& SetBlas(PluginId blas); 76 PluginId blas() const { return blas_; } in blas() function
|
/external/tensorflow/tensorflow/stream_executor/rocm/ |
D | rocm_blas.cc | 333 rocblas_operation ROCMBlasTranspose(blas::Transpose trans) { in ROCMBlasTranspose() 335 case blas::Transpose::kNoTranspose: in ROCMBlasTranspose() 337 case blas::Transpose::kTranspose: in ROCMBlasTranspose() 339 case blas::Transpose::kConjugateTranspose: in ROCMBlasTranspose() 346 rocblas_fill ROCMBlasUpperLower(blas::UpperLower uplo) { in ROCMBlasUpperLower() 348 case blas::UpperLower::kUpper: in ROCMBlasUpperLower() 350 case blas::UpperLower::kLower: in ROCMBlasUpperLower() 357 rocblas_diagonal ROCMBlasDiagonal(blas::Diagonal diag) { in ROCMBlasDiagonal() 359 case blas::Diagonal::kUnit: in ROCMBlasDiagonal() 361 case blas::Diagonal::kNonUnit: in ROCMBlasDiagonal() [all …]
|
D | rocm_blas.h | 50 class ROCMBlas : public blas::BlasSupport { 104 FuncT rocblas_func, Stream *stream, blas::Transpose transa, 105 blas::Transpose transb, uint64 m, uint64 n, uint64 k, T alpha, 119 Stream *stream, blas::Transpose transa, blas::Transpose transb, uint64 m, 122 DeviceMemory<OutT> *c, int ldc, blas::ComputationType computation_type, 123 blas::AlgorithmType algorithm, 124 blas::ProfileResult *output_profile_result); 129 Stream *stream, blas::Transpose transa, blas::Transpose transb, uint64 m, 132 DeviceMemory<T> *c, int ldc, blas::ProfileResult *output_profile_result); 136 bool DoBlasGemvWithProfilingImpl(Stream *stream, blas::Transpose trans, [all …]
|
/external/tensorflow/tensorflow/stream_executor/cuda/ |
D | cuda_blas.cc | 273 cublasOperation_t CUDABlasTranspose(blas::Transpose trans) { in CUDABlasTranspose() 275 case blas::Transpose::kNoTranspose: in CUDABlasTranspose() 277 case blas::Transpose::kTranspose: in CUDABlasTranspose() 279 case blas::Transpose::kConjugateTranspose: in CUDABlasTranspose() 286 cublasFillMode_t CUDABlasUpperLower(blas::UpperLower uplo) { in CUDABlasUpperLower() 288 case blas::UpperLower::kUpper: in CUDABlasUpperLower() 290 case blas::UpperLower::kLower: in CUDABlasUpperLower() 297 cublasDiagType_t CUDABlasDiagonal(blas::Diagonal diag) { in CUDABlasDiagonal() 299 case blas::Diagonal::kUnit: in CUDABlasDiagonal() 301 case blas::Diagonal::kNonUnit: in CUDABlasDiagonal() [all …]
|
D | cuda_blas.h | 53 class CUDABlas : public blas::BlasSupport { 111 FuncT cublas_func, Stream *stream, blas::Transpose transa, 112 blas::Transpose transb, uint64 m, uint64 n, uint64 k, Scalar alpha, 121 Stream *stream, blas::Transpose transa, blas::Transpose transb, uint64 m, 125 blas::ComputationType computation_type, blas::AlgorithmType algorithm, 126 blas::ProfileResult *output_profile_result); 131 Stream *stream, blas::Transpose transa, blas::Transpose transb, uint64 m, 134 DeviceMemory<T> *c, int ldc, blas::ProfileResult *output_profile_result); 138 bool DoBlasGemvWithProfilingImpl(Stream *stream, blas::Transpose trans, 143 blas::ProfileResult *output_profile_result);
|
/external/tensorflow/tensorflow/compiler/xla/service/gpu/ |
D | triangular_solve_thunk.cc | 41 uplo_(options.lower() ? se::blas::UpperLower::kLower in TriangularSolveThunk() 42 : se::blas::UpperLower::kUpper), in TriangularSolveThunk() 43 side_(options.left_side() ? se::blas::Side::kLeft in TriangularSolveThunk() 44 : se::blas::Side::kRight), in TriangularSolveThunk() 45 unit_diagonal_(options.unit_diagonal() ? se::blas::Diagonal::kUnit in TriangularSolveThunk() 46 : se::blas::Diagonal::kNonUnit), in TriangularSolveThunk() 58 return se::blas::Transpose::kNoTranspose; in TriangularSolveThunk() 60 return se::blas::Transpose::kTranspose; in TriangularSolveThunk() 62 return se::blas::Transpose::kConjugateTranspose; in TriangularSolveThunk() 66 return se::blas::Transpose::kNoTranspose; in TriangularSolveThunk() [all …]
|
D | gemm_thunk.cc | 66 auto lhs_transpose = lhs_matrix.transpose ? se::blas::Transpose::kTranspose in DoGemm() 67 : se::blas::Transpose::kNoTranspose; in DoGemm() 68 auto rhs_transpose = rhs_matrix.transpose ? se::blas::Transpose::kTranspose in DoGemm() 69 : se::blas::Transpose::kNoTranspose; in DoGemm() 118 se::blas::ComputationType computation_type, in DoGemmWithAlgorithm() 119 se::blas::AlgorithmType algorithm, se::Stream* stream, in DoGemmWithAlgorithm() 120 se::blas::ProfileResult* output_profile_result) { in DoGemmWithAlgorithm() 131 auto lhs_transpose = lhs_matrix.transpose ? se::blas::Transpose::kTranspose in DoGemmWithAlgorithm() 132 : se::blas::Transpose::kNoTranspose; in DoGemmWithAlgorithm() 133 auto rhs_transpose = rhs_matrix.transpose ? se::blas::Transpose::kTranspose in DoGemmWithAlgorithm() [all …]
|
D | cusolver_context.h | 50 Status Potrf(se::blas::UpperLower uplo, int n, se::DeviceMemory<float> dev_A, 53 Status Potrf(se::blas::UpperLower uplo, int n, se::DeviceMemory<double> dev_A, 56 Status Potrf(se::blas::UpperLower uplo, int n, 60 Status Potrf(se::blas::UpperLower uplo, int n, 67 StatusOr<int64> PotrfBufferSize(se::blas::UpperLower uplo, int n, 69 StatusOr<int64> PotrfBufferSize(se::blas::UpperLower uplo, int n, 71 StatusOr<int64> PotrfBufferSize(se::blas::UpperLower uplo, int n, 74 StatusOr<int64> PotrfBufferSize(se::blas::UpperLower uplo, int n,
|
D | triangular_solve_thunk.h | 56 const se::blas::UpperLower uplo_; 57 const se::blas::Side side_; 58 const se::blas::Diagonal unit_diagonal_; 59 se::blas::Transpose transpose_a_;
|
D | cusolver_context.cc | 44 cublasFillMode_t CUDABlasUpperLower(se::blas::UpperLower uplo) { in CUDABlasUpperLower() 46 case se::blas::UpperLower::kUpper: in CUDABlasUpperLower() 48 case se::blas::UpperLower::kLower: in CUDABlasUpperLower() 136 se::blas::UpperLower uplo, int n, se::DeviceMemory<T> A, int lda) { \ 148 se::blas::UpperLower uplo, int n, se::DeviceMemory<T> A, int lda, \
|
D | cholesky_thunk.cc | 41 uplo_(options.lower() ? se::blas::UpperLower::kLower in CholeskyThunk() 42 : se::blas::UpperLower::kUpper), in CholeskyThunk() 57 << " uplo=" << se::blas::UpperLowerString(uplo_) in ExecuteOnStream()
|
D | cusolver_rewriter.cc | 67 se::blas::UpperLower uplo = options.lower() ? se::blas::UpperLower::kLower in CreateCholesky() 68 : se::blas::UpperLower::kUpper; in CreateCholesky()
|
/external/tensorflow/tensorflow/core/kernels/ |
D | matrix_triangular_solve_op.cc | 205 se::blas::UpperLower upper_lower_matrix; in ComputeMatrix() 206 se::blas::Transpose transpose_matrix; in ComputeMatrix() 208 upper_lower_matrix = se::blas::UpperLower::kUpper; in ComputeMatrix() 210 upper_lower_matrix = se::blas::UpperLower::kLower; in ComputeMatrix() 213 transpose_matrix = se::blas::Transpose::kConjugateTranspose; in ComputeMatrix() 215 transpose_matrix = se::blas::Transpose::kNoTranspose; in ComputeMatrix() 224 se::blas::Side::kRight /*side*/, upper_lower_matrix /*uplo*/, in ComputeMatrix() 226 se::blas::Diagonal::kNonUnit /*diag*/, colmajor_rows /*m*/, in ComputeMatrix()
|
D | matmul_op.cc | 115 typedef se::blas::AlgorithmType AlgorithmType; 166 se::blas::ProfileResult* output_profile) { in Compute() 167 const auto blas_trans = trans ? se::blas::Transpose::kTranspose in Compute() 168 : se::blas::Transpose::kNoTranspose; in Compute() 201 se::blas::ProfileResult* output_profile) { in Compute() 219 se::blas::ComputationType* compute_type) { in GetCublasAutotuneComputationType() 220 using se::blas::ComputationType; in GetCublasAutotuneComputationType() 248 se::blas::AlgorithmConfig> 257 using se::blas::AlgorithmConfig; in launch() 258 using se::blas::ComputationType; in launch() [all …]
|
D | batch_matmul_op_impl.h | 294 constexpr se::blas::Transpose kTranspose = 295 is_complex<Scalar>::value ? se::blas::Transpose::kConjugateTranspose 296 : se::blas::Transpose::kTranspose; 297 se::blas::Transpose trans[] = {se::blas::Transpose::kNoTranspose, 346 blas_transpose_b != se::blas::Transpose::kConjugateTranspose && 347 blas_transpose_a != se::blas::Transpose::kConjugateTranspose) { 353 auto gemv_trans_a = blas_transpose_a == se::blas::Transpose::kTranspose 354 ? se::blas::Transpose::kNoTranspose 355 : se::blas::Transpose::kTranspose; 410 constexpr perftools::gputools::blas::Transpose kTranspose = [all …]
|
/external/tensorflow/tensorflow/contrib/rnn/kernels/ |
D | blas_gemm.cc | 45 se::blas::Transpose trans[] = {se::blas::Transpose::kNoTranspose, in operator ()() 46 se::blas::Transpose::kTranspose}; in operator ()()
|
/external/eigen/blas/testing/ |
D | CMakeLists.txt | 17 …add_test(${testname} "${Eigen_SOURCE_DIR}/blas/testing/runblastest.sh" "${testname}" "${Eigen_SOUR…
|
/external/eigen/blas/ |
D | CMakeLists.txt | 15 add_custom_target(blas) target 39 add_dependencies(blas eigen_blas eigen_blas_static)
|