diff options
Diffstat (limited to 'tests/validation/NEON/ConvolutionLayer.cpp')
| -rw-r--r-- | tests/validation/NEON/ConvolutionLayer.cpp | 1123 |
1 files changed, 1016 insertions, 107 deletions
diff --git a/tests/validation/NEON/ConvolutionLayer.cpp b/tests/validation/NEON/ConvolutionLayer.cpp index 9e00da16ae..d739d4e1a4 100644 --- a/tests/validation/NEON/ConvolutionLayer.cpp +++ b/tests/validation/NEON/ConvolutionLayer.cpp @@ -1,5 +1,5 @@ /* - * Copyright (c) 2017-2021 Arm Limited. + * Copyright (c) 2017-2024 Arm Limited. * * SPDX-License-Identifier: MIT * @@ -28,11 +28,16 @@ #include "arm_compute/runtime/NEON/functions/NEWinogradConvolutionLayer.h" #include "arm_compute/runtime/Tensor.h" #include "arm_compute/runtime/TensorAllocator.h" + +#include "src/core/CPP/Validate.h" +#include "src/core/helpers/MemoryHelpers.h" +#include "src/cpu/operators/CpuGemmConv2d.h" +#include "src/cpu/operators/CpuGemmDirectConv2d.h" +#include "src/cpu/operators/CpuWinogradConv2d.h" + #include "tests/NEON/Accessor.h" -#include "tests/PaddingCalculator.h" #include "tests/datasets/LargeConvolutionLayerDataset.h" #include "tests/datasets/SmallConvolutionLayerDataset.h" -#include "tests/datasets/TinyConvolutionLayerDataset.h" #include "tests/framework/Asserts.h" #include "tests/framework/Macros.h" #include "tests/framework/datasets/Datasets.h" @@ -46,6 +51,8 @@ namespace test { namespace validation { +using framework::dataset::make; + namespace detail { template <> @@ -77,10 +84,17 @@ const RelativeTolerance<half_float::half> rel_tolerance_f16(half_float::half(0.2 const AbsoluteTolerance<float> abs_tolerance_f16(0.2f); /**< Absolute tolerance for FP16 types */ constexpr float tolerance_num = 0.07f; /**< Tolerance number for the FP16 implementation */ #endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ -constexpr AbsoluteTolerance<float> tolerance_qasymm8(0.0); /**< Tolerance value for comparing reference's output against implementation's output for quantized data types */ + +#ifdef ARM_COMPUTE_ENABLE_SME +// TODO(COMPMID-6011): SME kernels and the reference model use different rounding mode. +// Temporarily increase the tolerance for quantized data. +constexpr AbsoluteTolerance<float> tolerance_qasymm8(1.0); /**< Tolerance value for comparing reference's output against implementation's output for quantized data types */ +#else // ARM_COMPUTE_ENABLE_SME +constexpr AbsoluteTolerance<float> tolerance_qasymm8(0.0); /**< Tolerance value for comparing reference's output against implementation's output for quantized data types */ +#endif // ARM_COMPUTE_ENABLE_SME /** CNN data types */ -const auto CNNDataTypes = framework::dataset::make("DataType", +const auto CNNDataTypes = make("DataType", { #ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC DataType::F16, @@ -88,14 +102,41 @@ const auto CNNDataTypes = framework::dataset::make("DataType", DataType::F32, DataType::QASYMM8, }); -const auto ActivationFunctionsDataset = framework::dataset::make("ActivationInfo", +const auto ActivationFunctionsDataset = make("ActivationInfo", { ActivationLayerInfo(), ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU), ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::BOUNDED_RELU, 0.5f) }); -const auto QuantizationData = framework::dataset::make("QuantizationInfo", +const auto NoActivation = make("ActivationInfo", +{ + ActivationLayerInfo(), +}); + +const auto ActivationFunctionsDatasetNightly = make("ActivationInfo", +{ + ActivationLayerInfo(), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::RELU), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::BOUNDED_RELU, 0.5f), + + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, 0.5f, -0.5f), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LEAKY_RELU, 0.1f), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::SOFT_RELU), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::ELU), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::ABS), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LOGISTIC), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::TANH), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::SQUARE), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::SWISH), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::HARD_SWISH), + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LINEAR, 2.f, 1.f), +#ifdef __aarch64__ + ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::GELU), +#endif // __aarch64__ +}); + +const auto QuantizationData = make("QuantizationInfo", { QuantizationInfo(0.5f, 10), QuantizationInfo(0.3f, 3), @@ -110,32 +151,32 @@ TEST_SUITE(ConvolutionLayer) // *INDENT-OFF* // clang-format off DATA_TEST_CASE(ValidateConvolutionMethod, framework::DatasetMode::ALL, zip(zip(zip(zip(zip( - framework::dataset::make("InputInfo", { TensorInfo(TensorShape(18U, 18U, 32U), 1, DataType::F32), + make("InputInfo", { TensorInfo(TensorShape(18U, 18U, 32U), 1, DataType::F32), TensorInfo(TensorShape(23U, 27U, 32U, 4U), 1, DataType::F32), TensorInfo(TensorShape(3U, 3U, 2U, 1U), 1, DataType::F32), TensorInfo(TensorShape(33U, 27U, 7U, 4U), 1, DataType::F32) }), - framework::dataset::make("WeightsInfo", { TensorInfo(TensorShape(3U, 3U, 32U, 21U), 1, DataType::F32), + make("WeightsInfo", { TensorInfo(TensorShape(3U, 3U, 32U, 21U), 1, DataType::F32), TensorInfo(TensorShape(5U, 5U, 32U, 21U), 1, DataType::F32), TensorInfo(TensorShape(3U, 3U, 5U, 21U), 1, DataType::F32), TensorInfo(TensorShape(5U, 5U, 7U, 16U), 1, DataType::F16) })), - framework::dataset::make("OutputInfo", { TensorInfo(TensorShape(16U, 16U, 21U), 1, DataType::F32), + make("OutputInfo", { TensorInfo(TensorShape(16U, 16U, 21U), 1, DataType::F32), TensorInfo(TensorShape(19U, 23U, 21U, 4U), 1, DataType::F32), TensorInfo(TensorShape(11U, 25U, 21U), 1, DataType::F32), TensorInfo(TensorShape(11U, 12U, 16U, 4U), 1, DataType::F32) })), - framework::dataset::make("ConvInfo", { PadStrideInfo(1, 1, 0, 0), + make("ConvInfo", { PadStrideInfo(1, 1, 0, 0), PadStrideInfo(1, 1, 0, 0), PadStrideInfo(2, 1, 0, 0), PadStrideInfo(3, 2, 1, 0) })), - framework::dataset::make("FastMath", { true, + make("FastMath", { true, true, false, false })), - framework::dataset::make("Expected", { ConvolutionMethod::WINOGRAD, ConvolutionMethod::WINOGRAD, ConvolutionMethod::GEMM, ConvolutionMethod::GEMM })), + make("Expected", { ConvolutionMethod::WINOGRAD, ConvolutionMethod::WINOGRAD, ConvolutionMethod::GEMM, ConvolutionMethod::GEMM })), input_info, weights_info, output_info, conv_info, fast_math, expected) { ConvolutionMethod is_valid = NEConvolutionLayer::get_convolution_method(&input_info.clone()->set_is_resizable(true), @@ -147,6 +188,14 @@ DATA_TEST_CASE(ValidateConvolutionMethod, framework::DatasetMode::ALL, zip(zip(z // *INDENT-ON* TEST_SUITE_END() // ConvolutionLayer +/* + Testing Strategy of Neon Winograd: + - There is no need to thoroughly test nchw cases because winograd kernels accept + nhwc and the tensors are permuted before and after if they're nchw. + - Except relu and bounded relu, testing activations for a single input + combination is enough because activation is not fused into winograd and called + separately. +*/ TEST_SUITE(WinogradLayer) template <typename T> using NEWinogradConvolutionLayerFixture = WinogradConvolutionLayerFastMathValidationFixture<Tensor, Accessor, NEWinogradConvolutionLayer, T>; @@ -156,38 +205,250 @@ using NEWinogradConvolutionLayerMixedDataLayoutFixture = WinogradConvolutionLaye template <typename T> using NEWinogradConvolutionLayerNoBiasFixture = WinogradConvolutionLayerFastMathValidationFixture<Tensor, Accessor, NEWinogradConvolutionLayer, T, T, false>; +/** Test case for memory injection in @ref cpu::CpuWinogradConv2d. + * + * Configure the operator once and inject memory at run-time in multiple executions. + * + * Checks performed in order: + * - Both runs compute the same output + */ +TEST_CASE(MemoryInjection, framework::DatasetMode::ALL) +{ + auto winograd = std::make_unique<cpu::CpuWinogradConv2d>(); + const auto src_info = TensorInfo(TensorShape(8U, 8U, 32U), 1, DataType::F32); + const auto w_info = TensorInfo(TensorShape(1U), 1, DataType::F32); + const auto b_info = TensorInfo(TensorShape(1U, 3U, 32U, 1U), 1, DataType::F32); + auto dst_info = TensorInfo(TensorShape(8U, 6U, 1U), 1, DataType::F32); + const PadStrideInfo pad_info{}; + + winograd->configure(&src_info, &b_info, &w_info, &dst_info, pad_info); + + // telhs are newly created every call of this lambda function + auto a = create_tensor<Tensor>(src_info); + auto b = create_tensor<Tensor>(b_info); + auto c = create_tensor<Tensor>(w_info); + a.allocator()->allocate(); + b.allocator()->allocate(); + c.allocator()->allocate(); + + ITensorPack run_pack{ { TensorType::ACL_SRC_0, &a }, { TensorType::ACL_SRC_1, &b }, { TensorType::ACL_SRC_2, &c } }; + ITensorPack prep_pack{ { TensorType::ACL_SRC_1, &b }, { TensorType::ACL_SRC_2, &c } }; + + auto mg = MemoryGroup{}; + auto ws = manage_workspace<Tensor>(winograd->workspace(), mg, run_pack, prep_pack); + auto run_conv = [&]() -> Tensor + { + auto dst = create_tensor<Tensor>(dst_info); + dst.allocator()->allocate(); + + run_pack.add_tensor(TensorType::ACL_DST, &dst); + library->fill_tensor_value(Accessor(a), 1.f); + library->fill_tensor_value(Accessor(b), 2.f); + library->fill_tensor_value(Accessor(c), 3.f); + + // This operator is configured once and captured by this lambda. + winograd->prepare(prep_pack); + winograd->run(run_pack); + return dst; + }; + + auto result_0 = run_conv(); + auto result_1 = run_conv(); + + for(size_t i = 0; i < result_0.info()->tensor_shape().total_size(); ++i) + { + ARM_COMPUTE_EXPECT(((float *)result_0.buffer())[i] == ((float *)result_1.buffer())[i], framework::LogLevel::ERRORS); + } +} + +/** Test case for memory injection in @ref NEWinogradConvolutionLayer. + * + * Make sure @ref NEWinogradConvolutionLayer still works through injecting the memory at configure time using the old API. + * + * Checks performed in order: + * - Both runs compute the same output + */ +TEST_CASE(MultipleExecutionWithConfigure, framework::DatasetMode::ALL) +{ + auto gemm = std::make_unique<NEWinogradConvolutionLayer>(); + const auto src_info = TensorInfo(TensorShape(8U, 8U, 32U), 1, DataType::F32); + const auto w_info = TensorInfo(TensorShape(1U), 1, DataType::F32); + const auto b_info = TensorInfo(TensorShape(1U, 3U, 32U, 1U), 1, DataType::F32); + auto dst_info = TensorInfo(TensorShape(8U, 6U, 1U), 1, DataType::F32); + const PadStrideInfo pad_info{}; + + auto run_conv = [&]() + { + auto src = create_tensor<Tensor>(src_info); + auto w = create_tensor<Tensor>(w_info); + auto b = create_tensor<Tensor>(b_info); + auto dst = create_tensor<Tensor>(dst_info); + + gemm->configure(&src, &b, &w, &dst, pad_info); + + src.allocator()->allocate(); + b.allocator()->allocate(); + w.allocator()->allocate(); + dst.allocator()->allocate(); + + library->fill_tensor_value(Accessor(src), 1.f); + library->fill_tensor_value(Accessor(b), 2.f); + library->fill_tensor_value(Accessor(w), 3.f); + gemm->run(); + return dst; + }; + + auto result_0 = run_conv(); + auto result_1 = run_conv(); + + for(size_t i = 0; i < result_0.info()->tensor_shape().total_size(); ++i) + { + ARM_COMPUTE_EXPECT(((float *)result_0.buffer())[i] == ((float *)result_1.buffer())[i], framework::LogLevel::ERRORS); + } +} + +DATA_TEST_CASE(SupportedKernels, framework::DatasetMode::ALL, zip( + make("WeightsInfo", +{ + // Shapes are always in NCHW format. When layout is NHWC, the shape is permuted + + // Fp32, NCHW/NHWC (layout does not matter as it's ) + // 3x1, 1x3, 3x3 --> all TRUE + TensorInfo(TensorShape(3U, 3U, 2U, 8U), 1, DataType::F32, DataLayout::NHWC), + TensorInfo(TensorShape(1U, 3U, 2U, 8U), 1, DataType::F32, DataLayout::NHWC), + TensorInfo(TensorShape(3U, 1U, 2U, 8U), 1, DataType::F32, DataLayout::NCHW), + + // 5x1, 1x5, 5x5 --> all TRUE + TensorInfo(TensorShape(5U, 5U, 2U, 8U), 1, DataType::F32, DataLayout::NCHW), + TensorInfo(TensorShape(1U, 5U, 2U, 8U), 1, DataType::F32, DataLayout::NHWC), + TensorInfo(TensorShape(5U, 1U, 2U, 8U), 1, DataType::F32, DataLayout::NCHW), + + // 7x1, 1x7, 7x7 + // --> all FALSE + TensorInfo(TensorShape(7U, 7U, 2U, 8U), 1, DataType::F32, DataLayout::NCHW), + TensorInfo(TensorShape(1U, 7U, 2U, 8U), 1, DataType::F32, DataLayout::NHWC), + TensorInfo(TensorShape(7U, 1U, 2U, 8U), 1, DataType::F32, DataLayout::NHWC), + + // unsupported kernel sizes + TensorInfo(TensorShape(2U, 2U, 2U, 8U), 1, DataType::F32, DataLayout::NHWC), + TensorInfo(TensorShape(5U, 2U, 2U, 8U), 1, DataType::F32, DataLayout::NHWC), + TensorInfo(TensorShape(3U, 6U, 2U, 8U), 1, DataType::F32, DataLayout::NCHW), + + // Fp16 + TensorInfo(TensorShape(3U, 3U, 2U, 8U), 1, DataType::F16, DataLayout::NHWC), + TensorInfo(TensorShape(1U, 3U, 2U, 8U), 1, DataType::F16, DataLayout::NHWC), + TensorInfo(TensorShape(3U, 1U, 2U, 8U), 1, DataType::F16, DataLayout::NCHW), + + // 5x1, 1x5, 5x5 --> all TRUE + TensorInfo(TensorShape(5U, 5U, 2U, 8U), 1, DataType::F16, DataLayout::NCHW), + TensorInfo(TensorShape(1U, 5U, 2U, 8U), 1, DataType::F16, DataLayout::NHWC), + TensorInfo(TensorShape(5U, 1U, 2U, 8U), 1, DataType::F16, DataLayout::NCHW), + + // 7x1, 1x7, 7x7 + // --> all FALSE + TensorInfo(TensorShape(7U, 7U, 2U, 8U), 1, DataType::F16, DataLayout::NCHW), + TensorInfo(TensorShape(1U, 7U, 2U, 8U), 1, DataType::F16, DataLayout::NHWC), + TensorInfo(TensorShape(7U, 1U, 2U, 8U), 1, DataType::F16, DataLayout::NHWC), + + // unsupported kernel sizes + TensorInfo(TensorShape(2U, 2U, 2U, 8U), 1, DataType::F16, DataLayout::NHWC), + TensorInfo(TensorShape(5U, 2U, 2U, 8U), 1, DataType::F16, DataLayout::NHWC), + TensorInfo(TensorShape(3U, 6U, 2U, 8U), 1, DataType::F16, DataLayout::NCHW), + +}), +make("Expected", +{ + // fp32 + true, true, true, // 3x3, 1x3, 3x1 + true, true, true, // 5x5, 1x5, 5x1 + false, true, true, // 7x7, 1x7, 7x1 + false, false, false, // random unsupported kernels + + // fp16 + true, false, false, // 3x3, 1x3, 3x1 + false, false, false, // 5x5, 1x5, 5x1 + false, false, false, // 7x7, 1x7, 7x1 + false, false, false, // random unsupported kernels +})), +weights_info_const, expected_const) +{ + DataType data_type = weights_info_const.data_type(); + DataLayout data_layout = weights_info_const.data_layout(); + + TensorInfo input_info = TensorInfo(TensorShape(17U, 31U, 2U), 1, data_type); + TensorInfo bias_info = TensorInfo(TensorShape(8U), 1, data_type); + TensorInfo weights_info = weights_info_const; + + if(data_layout == DataLayout::NHWC) + { + // Convert to NHWC + PermutationVector perm = PermutationVector(2U, 0U, 1U); + + TensorShape input_shape = input_info.tensor_shape(); + TensorShape weights_shape = weights_info.tensor_shape(); + permute(input_shape, perm); + permute(weights_shape, perm); + + input_info.set_tensor_shape(input_shape); + weights_info.set_tensor_shape(weights_shape); + + input_info.set_data_layout(data_layout); + weights_info.set_data_layout(data_layout); + bias_info.set_data_layout(data_layout); + } + + PadStrideInfo conv_info(1, 1, 0, 0); + + TensorShape output_shape = compute_deep_convolution_shape(input_info, weights_info, conv_info); + TensorInfo output_info = TensorInfo(output_shape, 1, data_type, data_layout); + + Status status = NEWinogradConvolutionLayer::validate( + &input_info, + &weights_info, + &bias_info, + &output_info, + conv_info, + ActivationLayerInfo(), + true /* fast math */); + + Status fp16_supported = ::arm_compute::error_on_unsupported_cpu_fp16("N/A", "N/A", 0, &input_info); + bool expected = expected_const && static_cast<bool>(fp16_supported); + + ARM_COMPUTE_EXPECT(bool(status) == expected, framework::LogLevel::ERRORS); +} + TEST_SUITE(FP32) TEST_SUITE(Conv1x3) FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer1x3Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer1x3Dataset(), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); } FIXTURE_DATA_TEST_CASE(RunMixedDataLayout, NEWinogradConvolutionLayerMixedDataLayoutFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(combine(combine(combine(combine(combine( - framework::dataset::make("Input", TensorShape(8U, 8U, 32U)), - framework::dataset::make("Weight", TensorShape(1U, 3U, 32U, 1U))), - framework::dataset::make("Bias", TensorShape(1U))), - framework::dataset::make("Output", TensorShape(8U, 6U, 1U))), - framework::dataset::make("PadStrideInfo", PadStrideInfo(1, 1, 0, 0))), - framework::dataset::make("Dilation", Size2D(1U, 1U))), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine( + make("Input", TensorShape(8U, 8U, 32U)), + make("Weight", TensorShape(1U, 3U, 32U, 1U)), + make("Bias", TensorShape(1U)), + make("Output", TensorShape(8U, 6U, 1U)), + make("PadStrideInfo", PadStrideInfo(1, 1, 0, 0)), + make("Dilation", Size2D(1U, 1U)), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); } FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, - combine(combine(combine(datasets::LargeWinogradConvolutionLayer1x3Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::LargeWinogradConvolutionLayer1x3Dataset(), + make("DataType", { DataType::F32 }), + make("ActivationInfo", { ActivationLayerInfo() }), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_1xN_f32); @@ -197,19 +458,19 @@ TEST_SUITE_END() // Conv1x3 TEST_SUITE(Conv3x1) FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer3x1Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer3x1Dataset(), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); } FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, - combine(combine(combine(datasets::LargeWinogradConvolutionLayer3x1Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::LargeWinogradConvolutionLayer3x1Dataset(), + make("DataType", { DataType::F32 }), + make("ActivationInfo", { ActivationLayerInfo() }), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_1xN_f32); @@ -219,19 +480,19 @@ TEST_SUITE_END() // Conv3x1 TEST_SUITE(Conv1x5) FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer1x5Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer1x5Dataset(), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); } FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, - combine(combine(combine(datasets::LargeWinogradConvolutionLayer1x5Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::LargeWinogradConvolutionLayer1x5Dataset(), + make("DataType", { DataType::F32 }), + make("ActivationInfo", { ActivationLayerInfo() }), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_1xN_f32); @@ -241,19 +502,19 @@ TEST_SUITE_END() // Conv1x5 TEST_SUITE(Conv5x1) FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer5x1Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer5x1Dataset(), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); } FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, - combine(combine(combine(datasets::LargeWinogradConvolutionLayer5x1Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::LargeWinogradConvolutionLayer5x1Dataset(), + make("DataType", { DataType::F32 }), + make("ActivationInfo", { ActivationLayerInfo() }), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_1xN_f32); @@ -263,10 +524,10 @@ TEST_SUITE_END() // Conv5x1 TEST_SUITE(Conv7x1) FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer7x1Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer7x1Dataset(), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); @@ -274,9 +535,9 @@ FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, frame FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, combine(combine(combine(datasets::LargeWinogradConvolutionLayer7x1Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + make("DataType", { DataType::F32 })), + make("ActivationInfo", { ActivationLayerInfo() })), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_1xN_f32); @@ -285,20 +546,20 @@ TEST_SUITE_END() // Conv7x1 TEST_SUITE(Conv1x7) FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer1x7Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer1x7Dataset(), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); } FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, - combine(combine(combine(datasets::LargeWinogradConvolutionLayer7x1Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::LargeWinogradConvolutionLayer7x1Dataset(), + make("DataType", { DataType::F32 }), + make("ActivationInfo", { ActivationLayerInfo() }), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_1xN_f32); @@ -307,20 +568,40 @@ TEST_SUITE_END() // Conv1x7 TEST_SUITE(Conv3x3) FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer3x3Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer3x3Dataset(), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + +{ + // Validate output + validate(Accessor(_target), _reference, abs_tolerance_f32); +} +/// It's enough to run the activations for a single weight/input combination and data type because +/// activation function is called on top of the winograd output as a separate operator +/// TODO: Enable after COMPMID-6573 is resolved +FIXTURE_DATA_TEST_CASE(RunActivations, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::DISABLED, + combine( + make("Input", TensorShape(3U, 3U, 32U)), + make("Weight", TensorShape(3U, 3U, 32U, 4U)), + make("Bias", TensorShape(4U)), + make("Output", TensorShape(1U, 1U, 4U)), + make("PadStrideInfo", PadStrideInfo(1, 1, 0, 0)), + make("Dilation", Size2D(1U, 1U)), + make("DataType", { DataType::F32 }), + ActivationFunctionsDatasetNightly, + make("DataLayout", { DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); } + FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, - combine(combine(combine(datasets::LargeWinogradConvolutionLayer3x3Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::LargeWinogradConvolutionLayer3x3Dataset(), + make("DataType", { DataType::F32 }), + make("ActivationInfo", { ActivationLayerInfo() }), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output @@ -331,20 +612,20 @@ TEST_SUITE_END() // Conv3x3 TEST_SUITE(Conv5x5) FIXTURE_DATA_TEST_CASE(RunSmall, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer5x5Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer5x5Dataset(), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); } FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, - combine(combine(combine(datasets::LargeWinogradConvolutionLayer5x5Dataset(), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::LargeWinogradConvolutionLayer5x5Dataset(), + make("DataType", { DataType::F32 }), + make("ActivationInfo", { ActivationLayerInfo() }), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output @@ -354,12 +635,12 @@ FIXTURE_DATA_TEST_CASE(RunLarge, NEWinogradConvolutionLayerFixture<float>, frame TEST_SUITE_END() // Conv5x5 FIXTURE_DATA_TEST_CASE(RunSmallNoBias, NEWinogradConvolutionLayerNoBiasFixture<float>, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(framework::dataset::concat(datasets::SmallWinogradConvolutionLayer3x3Dataset(), - datasets::SmallWinogradConvolutionLayer5x5Dataset()), - framework::dataset::make("DataType", { DataType::F32 })), - ActivationFunctionsDataset), - - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(framework::dataset::concat( + datasets::SmallWinogradConvolutionLayer3x3Dataset(), + datasets::SmallWinogradConvolutionLayer5x5Dataset()), + make("DataType", { DataType::F32 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output validate(Accessor(_target), _reference, abs_tolerance_f32); @@ -371,12 +652,39 @@ TEST_SUITE_END() // FP32 TEST_SUITE(FP16) using CLWinogradConvolutionLayerFastMathFixture16 = WinogradConvolutionLayerFastMathValidationFixture<Tensor, Accessor, NEWinogradConvolutionLayer, half, float>; +DATA_TEST_CASE(ValidateConvolutionMethod, framework::DatasetMode::ALL, zip( + make("InputInfo", { TensorInfo(TensorShape(18U, 18U, 32U), 1, DataType::F16), + TensorInfo(TensorShape(18U, 18U, 32U), 1, DataType::F16) + }), + make("WeightsInfo", { TensorInfo(TensorShape(3U, 3U, 32U, 21U), 1, DataType::F16), + TensorInfo(TensorShape(3U, 3U, 32U, 21U), 1, DataType::F16) + }), + make("OutputInfo", { TensorInfo(TensorShape(16U, 16U, 21U), 1, DataType::F32), + TensorInfo(TensorShape(16U, 16U, 21U), 1, DataType::F16) + }), + make("ConvInfo", { PadStrideInfo(1, 1, 0, 0), + PadStrideInfo(1, 1, 0, 0) + }), + make("FastMath", +{ + false, // case fp16 and fast_math False then disable Winograd + true // case fp16 and fast_math True then enable Winograd +}), +make("Expected", { ConvolutionMethod::GEMM, ConvolutionMethod::WINOGRAD })), +input_info, weights_info, output_info, conv_info, fast_math, expected) +{ + ConvolutionMethod is_valid = NEConvolutionLayer::get_convolution_method(&input_info.clone()->set_is_resizable(true), + &weights_info.clone()->set_is_resizable(true), + &output_info.clone()->set_is_resizable(true), conv_info, WeightsInfo(), Size2D(1U, 1U), ActivationLayerInfo(), fast_math); + ARM_COMPUTE_EXPECT(is_valid == expected, framework::LogLevel::ERRORS); +} + TEST_SUITE(Conv3x3) FIXTURE_DATA_TEST_CASE(RunSmall, CLWinogradConvolutionLayerFastMathFixture16, framework::DatasetMode::PRECOMMIT, - combine(combine(combine(datasets::SmallWinogradConvolutionLayer3x3Dataset(), - framework::dataset::make("DataType", { DataType::F16 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::SmallWinogradConvolutionLayer3x3Dataset(), + make("DataType", { DataType::F16 }), + ActivationFunctionsDataset, + make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) { // Validate output @@ -384,10 +692,10 @@ FIXTURE_DATA_TEST_CASE(RunSmall, CLWinogradConvolutionLayerFastMathFixture16, fr } FIXTURE_DATA_TEST_CASE(RunLarge, CLWinogradConvolutionLayerFastMathFixture16, framework::DatasetMode::NIGHTLY, - combine(combine(combine(datasets::LargeWinogradConvolutionLayer3x3Dataset(), - framework::dataset::make("DataType", { DataType::F16 })), - ActivationFunctionsDataset), - framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }))) + combine(datasets::LargeWinogradConvolutionLayer3x3Dataset(), + make("DataType", { DataType::F16 }), + make("ActivationInfo", { ActivationLayerInfo() }), + make("DataLayout", { DataLayout::NHWC }))) { // Validate output @@ -398,18 +706,470 @@ TEST_SUITE_END() // FP16 #endif /* __ARM_FEATURE_FP16_VECTOR_ARITHMETIC */ TEST_SUITE_END() // WinogradLayer +#ifdef ARM_COMPUTE_ENABLE_FIXED_FORMAT_KERNELS +TEST_SUITE(FIXED_FORMAT_KERNELS) +TEST_SUITE(VariableWeightUtils) + +// UC2_1_* tests: the user requests a specific fixed format, but there is no kernel that supports it. + +template <typename ConvolutionClass> +using HasOptImplFixtureNoFastMath = HasOptImplFixture<ConvolutionClass, /*enable_fast_math*/ false>; + +template <typename ConvolutionClass> +using HasOptImplFixtureFastMath = HasOptImplFixture<ConvolutionClass, /*enable_fast_math*/ true>; + +// UC2_1 + +FIXTURE_DATA_TEST_CASE(UC2_1_CpuGemmConv2d, HasOptImplFixtureNoFastMath<cpu::CpuGemmConv2d>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::OHWIo2 }))) +{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); +} +FIXTURE_DATA_TEST_CASE(UC2_1_NEGEMMConvolutionLayer, HasOptImplFixtureNoFastMath<NEGEMMConvolutionLayer>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::OHWIo2 }))) +{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); +} + +FIXTURE_DATA_TEST_CASE(UC2_1_CpuGemmConv2d_FastMath, HasOptImplFixtureFastMath<cpu::CpuGemmConv2d>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::OHWIo2 }))) +{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); +} + +FIXTURE_DATA_TEST_CASE(UC2_1_NEGEMMConvolutionLayer_FastMath, HasOptImplFixtureFastMath<NEGEMMConvolutionLayer>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::OHWIo2 }))) +{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); +} + +// UC2_2_* tests: the user requests a specific fixed format, and a +// kernel that support that fixed format is found. + +FIXTURE_DATA_TEST_CASE(UC2_2_CpuGemmConv2d, HasOptImplFixtureNoFastMath<cpu::CpuGemmConv2d>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::OHWIo4 }))) +{ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format == arm_compute::WeightFormat::OHWIo4, framework::LogLevel::ERRORS); +} + +FIXTURE_DATA_TEST_CASE(UC2_2_NEGEMMConvolutionLayer, HasOptImplFixtureNoFastMath<NEGEMMConvolutionLayer>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::OHWIo4 }))) +{ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format == arm_compute::WeightFormat::OHWIo4, framework::LogLevel::ERRORS); +} + +#if defined(ARM_COMPUTE_ENABLE_BF16) +// These tests currently only works with SVE length 256 +// If other SVE length is used a kernel will fail to be found +// This needs to be addressed in order to ensure it doesn't revert to FP32 kernels for systems with SVE length other than 256 +FIXTURE_DATA_TEST_CASE(UC2_2_CpuGemmConv2d_FastMath, HasOptImplFixtureFastMath<cpu::CpuGemmConv2d>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::OHWIo8i4_bf16 }))) +{ + if(Scheduler::get().cpu_info().has_bf16() && (arm_gemm::utils::get_vector_length<float>() == 8)){ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT_EQUAL(_computed_weight_format, arm_compute::WeightFormat::OHWIo8i4_bf16, framework::LogLevel::ERRORS); + } + else{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); + } +} + +FIXTURE_DATA_TEST_CASE(UC2_2_NEGEMMConvolutionLayer_FastMath, HasOptImplFixtureFastMath<NEGEMMConvolutionLayer>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::OHWIo8i4_bf16 }))) +{ + if(Scheduler::get().cpu_info().has_bf16() && (arm_gemm::utils::get_vector_length<float>() == 8)){ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format == arm_compute::WeightFormat::OHWIo8i4_bf16, framework::LogLevel::ERRORS); + } + else{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); + } +} + +#endif // ARM_COMPUTE_ENABLE_BF16 + +// UC3_1_* tests: the user queries for ANY fixed format, but there is +// no kernel that support the use case specified by the user (for +// example, there is no fixed format kernel for the datatype of the +// problem). + +FIXTURE_DATA_TEST_CASE(UC3_1_CpuGemmConv2d, HasOptImplFixtureNoFastMath<cpu::CpuGemmConv2d>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::S32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::ANY }))) +{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); +} + +FIXTURE_DATA_TEST_CASE(UC3_1_NEGEMMConvolutionLayer, HasOptImplFixtureNoFastMath<NEGEMMConvolutionLayer>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::S32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::ANY }))) +{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); +} + +FIXTURE_DATA_TEST_CASE(UC3_1_CpuGemmConv2d_FastMath, HasOptImplFixtureFastMath<cpu::CpuGemmConv2d>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::S32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::ANY }))) +{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); +} + +FIXTURE_DATA_TEST_CASE(UC3_1_NEGEMMConvolutionLayer_FastMath, HasOptImplFixtureFastMath<NEGEMMConvolutionLayer>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::S32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::ANY }))) +{ + ARM_COMPUTE_EXPECT(!_kernel_found, framework::LogLevel::ERRORS); +} + +// UC3_2_* tests: the user queries for ANY fixed format. The search +// succeeded and the fixed format found is prompted back for +// consumption by the user. Note that we just test the +// _computed_weight_format to be anything but not the formats that are +// not fixed formats (ANY and UNSPECIFIED). This is because the weight +// format that the runtime produces depends on the size of the vector +// units of the hardware where the tests is executed. For example, a +// format like OHWIo4 for FP32 data returned for 128-bit NEON hardware +// is replaced by OHWIo8 when running on 256-bit SVE. + +FIXTURE_DATA_TEST_CASE(UC3_2_CpuGemmConv2d, HasOptImplFixtureNoFastMath<cpu::CpuGemmConv2d>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::ANY }))) +{ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::ANY, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::UNSPECIFIED, framework::LogLevel::ERRORS); +} + +FIXTURE_DATA_TEST_CASE(UC3_2_NEGEMMConvolutionLayer, HasOptImplFixtureNoFastMath<NEGEMMConvolutionLayer>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::ANY }))) +{ + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::ANY, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::UNSPECIFIED, framework::LogLevel::ERRORS); +} + +#if defined(ARM_COMPUTE_ENABLE_BF16) + +FIXTURE_DATA_TEST_CASE(UC3_2_CpuGemmConv2d_FastMath, HasOptImplFixtureFastMath<cpu::CpuGemmConv2d>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::ANY }))) +{ + if(Scheduler::get().cpu_info().has_bf16()){ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::ANY, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::UNSPECIFIED, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(arm_compute::is_fixed_format_fast_math(_computed_weight_format), framework::LogLevel::ERRORS); + } + else{ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::ANY, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::UNSPECIFIED, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(!arm_compute::is_fixed_format_fast_math(_computed_weight_format), framework::LogLevel::ERRORS); + } +} + +FIXTURE_DATA_TEST_CASE(UC3_2_NEGEMMConvolutionLayer_FastMath, HasOptImplFixtureFastMath<NEGEMMConvolutionLayer>, framework::DatasetMode::ALL, + combine(framework::dataset::make("DataType", { DataType::F32 }), + framework::dataset::make("QueryWeightFormat", { arm_compute::WeightFormat::ANY }))) +{ + if(Scheduler::get().cpu_info().has_bf16()){ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::ANY, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::UNSPECIFIED, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(arm_compute::is_fixed_format_fast_math(_computed_weight_format), framework::LogLevel::ERRORS); + } + else{ + ARM_COMPUTE_EXPECT(_kernel_found, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::ANY, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(_computed_weight_format != arm_compute::WeightFormat::UNSPECIFIED, framework::LogLevel::ERRORS); + ARM_COMPUTE_EXPECT(!arm_compute::is_fixed_format_fast_math(_computed_weight_format), framework::LogLevel::ERRORS); + } +} + +#endif // ARM_COMPUTE_ENABLE_BF16 + +namespace +{ +using TestCaseType = std::tuple<TensorShape, TensorShape, arm_compute::WeightFormat>; +auto prepare_weights_shapes = framework::dataset::make("TensorShape", +{ + // OHWIo<interleave_by>i<block_by> + // + // OHWI --> O'HWI', where: + // + // O'= smallest multiple of <interleave_by> such that O<=O' + // I'= smallest multiple of <block_by> such that I<=I' + // + + // Change N for OHWIo4 + TestCaseType({ { 1U, 1U, 1U, 1U }, { 1U, 1U, 1U, 4U }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1U, 1U, 1U, 2U }, { 1U, 1U, 1U, 4U }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1U, 1U, 1U, 3U }, { 1U, 1U, 1U, 4U }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1U, 1U, 1U, 4U }, { 1U, 1U, 1U, 4U }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1U, 1U, 1U, 5U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1U, 1U, 1U, 6U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1U, 1U, 1U, 7U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1U, 1U, 1U, 8U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1U, 1U, 1U, 9U }, { 1U, 1U, 1U, 12U }, arm_compute::WeightFormat::OHWIo4 }), + // // Change N for OHWIo8 + TestCaseType({ { 1U, 1U, 1U, 1U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1U, 1U, 1U, 2U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1U, 1U, 1U, 3U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1U, 1U, 1U, 4U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1U, 1U, 1U, 5U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1U, 1U, 1U, 6U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1U, 1U, 1U, 7U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1U, 1U, 1U, 8U }, { 1U, 1U, 1U, 8U }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1U, 1U, 1U, 9U }, { 1U, 1U, 1U, 16U }, arm_compute::WeightFormat::OHWIo8 }), + // // Change N for OHWIo4 when H, W and C are not 1 + TestCaseType({ { 3U, 4U, 2U, 1U }, { 3, 4, 2, 4 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 2U }, { 3, 4, 2, 4 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 3U }, { 3, 4, 2, 4 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 4U }, { 3, 4, 2, 4 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 5U }, { 3, 4, 2, 8 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 6U }, { 3, 4, 2, 8 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 7U }, { 3, 4, 2, 8 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 8U }, { 3, 4, 2, 8 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 9U }, { 3, 4, 2, 12 }, arm_compute::WeightFormat::OHWIo4 }), + + // // Fix N and move HWI around, with different data layouts and formats + TestCaseType({ { 2U, 4U, 3U, 5U }, { 2, 4, 3, 8 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 3U, 4U, 2U, 5U }, { 3, 4, 2, 8 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 2U, 4U, 3U, 9U }, { 2, 4, 3, 16 }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 3U, 4U, 2U, 9U }, { 3, 4, 2, 16 }, arm_compute::WeightFormat::OHWIo8 }), + TestCaseType({ { 1024U, 1U, 1U, 1001U }, { 1024, 1, 1, 1008 }, arm_compute::WeightFormat::OHWIo8 }), + + // // Adding <block_by> on I (=C) + TestCaseType({ { 1U, 4U, 3U, 5U }, { 2, 4, 3, 8 }, arm_compute::WeightFormat::OHWIo4i2 }), + TestCaseType({ { 2U, 4U, 3U, 5U }, { 2, 4, 3, 8 }, arm_compute::WeightFormat::OHWIo4i2 }), + TestCaseType({ { 3U, 4U, 3U, 5U }, { 4, 4, 3, 8 }, arm_compute::WeightFormat::OHWIo4i2 }), + + // --------- + TestCaseType({ { 2, 2, 1, 5 }, { 2, 2, 1, 8 }, arm_compute::WeightFormat::OHWIo4 }), + TestCaseType({ { 1, 2, 2, 5 }, { 1, 2, 2, 8 }, arm_compute::WeightFormat::OHWIo4 }), + +}); +} // unnamed namespace + +DATA_TEST_CASE(PrepareWeightShape, framework::DatasetMode::ALL, + prepare_weights_shapes, shapes) +{ + const TensorShape input_shape = std::get<0>(shapes); + const TensorShape expected_shape = std::get<1>(shapes); + const arm_compute::WeightFormat wf = std::get<2>(shapes); + const DataType DT = DataType::F32; + const DataLayout DL = DataLayout::NHWC; + const auto TI = TensorInfo(input_shape, 1 /*num_channels, deprecated*/, DT, DL); + const TensorInfo computed_info = ::arm_compute::test::validation::prepare_weights(TI, wf); + ARM_COMPUTE_EXPECT_EQUAL(computed_info.tensor_shape(), expected_shape, framework::LogLevel::ERRORS); +} + +TEST_SUITE_END() // VariableWeightUtils + +TEST_SUITE(ExperimentalCpuAPIVariableWeightWithFixtures) + +template <typename ScalarType> +using VarWidth = VariableWeightsFixture<cpu::CpuGemmConv2d, Tensor, Accessor, ScalarType, /*enable_fast_math*/ false>; + +FIXTURE_DATA_TEST_CASE(RunSmallFloat, VarWidth<float>, framework::DatasetMode::ALL, + combine(combine(datasets::SmallConvolutionLayerDataset(), + framework::dataset::make("DataLayout", { DataLayout::NHWC })), + framework::dataset::make("ACL Scalar type", { DataType::F32 }))) +{ + // Validate output + validate(Accessor(_target), _reference, rel_tolerance_f32, 0.f, float(abs_tolerance_f32)); +} + +#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) +FIXTURE_DATA_TEST_CASE(RunSmallHalf, VarWidth<half>, framework::DatasetMode::ALL, + combine(combine(datasets::SmallConvolutionLayerDataset(), + framework::dataset::make("DataLayout", { DataLayout::NHWC })), + framework::dataset::make("ACL Scalar type", { DataType::F16 }))) +{ + // Validate output + validate(Accessor(_target), _reference, rel_tolerance_f16, 0.f, half(abs_tolerance_f16)); +} +#endif // __ARM_FEATURE_FP16_VECTOR_ARITHMETIC + +#if defined(ARM_COMPUTE_ENABLE_BF16) +template <typename ScalarType> +using VarWidthFastMath = VariableWeightsFixture<cpu::CpuGemmConv2d, Tensor, Accessor, ScalarType, /*enable_fast_math*/ true>; + +FIXTURE_DATA_TEST_CASE(RunSmallFloatFastMath, VarWidthFastMath<float>, framework::DatasetMode::ALL, + combine(combine(datasets::SmallConvolutionLayerDataset(), + framework::dataset::make("DataLayout", { DataLayout::NHWC })), + framework::dataset::make("ACL Scalar type", { DataType::F32 }))) +{ + // Validate output + validate(Accessor(_target), _reference, rel_tolerance_f32, 0.f, float(abs_tolerance_f32)); +} +#endif // ARM_COMPUTE_ENABLE_BF16 + +TEST_SUITE_END() // ExperimentalCpuAPIVariableWeightWithFixtures + +TEST_SUITE(ExperimentalNEAPIVariableWeightWithFixtures) + +template <typename ScalarType> +using NEGEMMVarWidth = VariableWeightsFixtureNEInterface<NEGEMMConvolutionLayer, Tensor, Accessor, ScalarType, /*enable_fast_math*/ false>; + +FIXTURE_DATA_TEST_CASE(NEGEMMRunSmallFloat, NEGEMMVarWidth<float>, framework::DatasetMode::ALL, + combine(combine(datasets::SmallConvolutionLayerDataset(), + framework::dataset::make("DataLayout", { DataLayout::NHWC })), + framework::dataset::make("ACL Scalar type", { DataType::F32 }))) +{ + // Validate output + validate(Accessor(_target), _reference, rel_tolerance_f32, 0.f, float(abs_tolerance_f32)); +} + +#if defined(__ARM_FEATURE_FP16_VECTOR_ARITHMETIC) +FIXTURE_DATA_TEST_CASE(NEGEMMRunSmallHalf, NEGEMMVarWidth<half>, framework::DatasetMode::ALL, + combine(combine(datasets::SmallConvolutionLayerDataset(), + framework::dataset::make("DataLayout", { DataLayout::NHWC })), + framework::dataset::make("ACL Scalar type", { DataType::F16 }))) +{ + // Validate output + validate(Accessor(_target), _reference, rel_tolerance_f16, 0.f, half(abs_tolerance_f16)); +} +#endif // __ARM_FEATURE_FP16_VECTOR_ARITHMETIC + +#if defined(ARM_COMPUTE_ENABLE_BF16) +template <typename ScalarType> +using NEGEMMVarWidthFastMath = VariableWeightsFixtureNEInterface<NEGEMMConvolutionLayer, Tensor, Accessor, ScalarType, /*enable_fast_math*/ true>; + +FIXTURE_DATA_TEST_CASE(NEGEMMRunSmallFloatFastMath, NEGEMMVarWidthFastMath<float>, framework::DatasetMode::ALL, + combine(combine(datasets::SmallConvolutionLayerDataset(), + framework::dataset::make("DataLayout", { DataLayout::NHWC })), + framework::dataset::make("ACL Scalar type", { DataType::F32 }))) +{ + // Validate output + validate(Accessor(_target), _reference, rel_tolerance_f32, 0.f, float(abs_tolerance_f32)); +} +#endif // ARM_COMPUTE_ENABLE_BF16 + +TEST_SUITE_END() // ExperimentalNEAPIVariableWeightWithFixtures +TEST_SUITE_END() // FIXED_FORMAT_KERNELS + +#endif // ARM_COMPUTE_ENABLE_FIXED_FORMAT_KERNELS + TEST_SUITE(GEMMConvolutionLayer) template <typename T> using NEGEMMConvolutionLayerFixture = ConvolutionValidationFixture<Tensor, Accessor, NEConvolutionLayer, T>; template <typename T> +using NEGEMMConvolutionLayerPaddedWeightsFixture = ConvolutionValidationPaddedWeightsFixture<Tensor, Accessor, NEConvolutionLayer, T>; +template <typename T> using NEGEMMConvolutionLayerMixedDataLayoutFixture = ConvolutionValidationFixture<Tensor, Accessor, NEConvolutionLayer, T, true>; +/** Test case for memory injection in @ref cpu::CpuGemmConv2d. + * + * Configure the operator once and inject memory at run-time in multiple executions. + * + * Checks performed in order: + * - Both runs compute the same output + */ +TEST_CASE(MemoryInjection, framework::DatasetMode::ALL) +{ + auto conv = std::make_unique<cpu::CpuGemmConv2d>(); + const auto src_info = TensorInfo(TensorShape(1U, 5U, 2U), 1, DataType::F32, DataLayout::NCHW); + const auto weight_info = TensorInfo(TensorShape(1U, 3U, 2U, 3U), 1, DataType::F32, DataLayout::NCHW); + const auto bias_info = TensorInfo(TensorShape(3U), 1, DataType::F32, DataLayout::NCHW); + auto dst_info = TensorInfo(TensorShape(1U, 7U, 3U), 1, DataType::F32, DataLayout::NCHW); + const auto conv_info = PadStrideInfo(1, 1, 0, 0, 2, 2, DimensionRoundingType::FLOOR); + WeightsInfo weights_info(false, 3U, 3U, 1U); + conv->configure(&src_info, &weight_info, &bias_info, &dst_info, conv_info, weights_info); + + // tensors are newly created every call of this lambda function + auto src = create_tensor<Tensor>(src_info); + auto weight = create_tensor<Tensor>(weight_info); + auto bias = create_tensor<Tensor>(bias_info); + src.allocator()->allocate(); + weight.allocator()->allocate(); + bias.allocator()->allocate(); + + ITensorPack run_pack{ { TensorType::ACL_SRC_0, &src }, { TensorType::ACL_SRC_1, &weight }, { TensorType::ACL_SRC_2, &bias } }; + ITensorPack prep_pack{ { TensorType::ACL_SRC_1, &weight }, { TensorType::ACL_SRC_2, &bias } }; + + auto mg = MemoryGroup{}; + auto ws = manage_workspace<Tensor>(conv->workspace(), mg, run_pack, prep_pack); + + auto run_conv = [&]() -> Tensor + { + auto dst = create_tensor<Tensor>(dst_info); + dst.allocator()->allocate(); + run_pack.add_tensor(TensorType::ACL_DST, &dst); + + library->fill_tensor_value(Accessor(src), 1.f); + library->fill_tensor_value(Accessor(weight), 2.f); + library->fill_tensor_value(Accessor(bias), 3.f); + // This operator is configured once and captured by this lambda. + conv->prepare(prep_pack); + conv->run(run_pack); + return dst; + }; + auto result_0 = run_conv(); + auto result_1 = run_conv(); + for(size_t i = 0; i < result_0.info()->tensor_shape().total_size(); ++i) + { + ARM_COMPUTE_EXPECT(((float *)result_0.buffer())[i] == ((float *)result_1.buffer())[i], framework::LogLevel::ERRORS); + } +} + +/** Test case for memory injection in @ref NEGEMMConvolutionLayer. + * + * Make sure @ref NEGEMMConvolutionLayer still works through injecting the memory at configure time using the old API. + * + * Checks performed in order: + * - Both runs compute the same output + */ +TEST_CASE(MultipleExecutionWithConfigure, framework::DatasetMode::ALL) +{ + auto conv = std::make_unique<NEGEMMConvolutionLayer>(); + const auto src_info = TensorInfo(TensorShape(1U, 5U, 2U), 1, DataType::F32, DataLayout::NCHW); + const auto weight_info = TensorInfo(TensorShape(1U, 3U, 2U, 3U), 1, DataType::F32, DataLayout::NCHW); + const auto bias_info = TensorInfo(TensorShape(3U), 1, DataType::F32, DataLayout::NCHW); + auto dst_info = TensorInfo(TensorShape(1U, 7U, 3U), 1, DataType::F32, DataLayout::NCHW); + const auto conv_info = PadStrideInfo(1, 1, 0, 0, 2, 2, DimensionRoundingType::FLOOR); + WeightsInfo weights_info(false, 3U, 3U, 1U); + auto run_conv = [&]() + { + auto src = create_tensor<Tensor>(src_info); + auto weight = create_tensor<Tensor>(weight_info); + auto bias = create_tensor<Tensor>(bias_info); + auto dst = create_tensor<Tensor>(dst_info); + conv->configure(&src, &weight, &bias, &dst, conv_info, weights_info); + src.allocator()->allocate(); + weight.allocator()->allocate(); + bias.allocator()->allocate(); + dst.allocator()->allocate(); + library->fill_tensor_value(Accessor(src), 1.f); + library->fill_tensor_value(Accessor(weight), 2.f); + library->fill_tensor_value(Accessor(bias), 3.f); + conv->run(); + return dst; + }; + auto result_0 = run_conv(); + auto result_1 = run_conv(); + for(size_t i = 0; i < result_0.info()->tensor_shape().total_size(); ++i) + { + ARM_COMPUTE_EXPECT(((float *)result_0.buffer())[i] == ((float *)result_1.buffer())[i], framework::LogLevel::ERRORS); + } +} + TEST_SUITE(Float) -#if defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) || defined(ARM_COMPUTE_FORCE_BF16) +#if defined(ARM_COMPUTE_ENABLE_BF16) TEST_SUITE(BFLOAT16) FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMConvolutionLayerFixture<float>, framework::DatasetMode::ALL, combine(combine(combine(combine(datasets::SmallConvolutionLayerDataset(), framework::dataset::make("ReshapeWeights", { true })), - framework::dataset::make("DataType", DataType::BFLOAT16)), + framework::dataset::make("DataType", Scheduler::get().cpu_info().has_bf16() ? DataType::BFLOAT16 : DataType::F32)), framework::dataset::make("DataLayout", { DataLayout::NHWC })), ActivationFunctionsDataset)) { @@ -417,7 +1177,7 @@ FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMConvolutionLayerFixture<float>, framework validate(Accessor(_target), _reference, rel_tolerance_f32, 0.f, float(abs_tolerance_f32)); } TEST_SUITE_END() // BFLOAT16 -#endif /* defined(__ARM_FEATURE_BF16_VECTOR_ARITHMETIC) || defined(ARM_COMPUTE_FORCE_BF16) */ +#endif /* defined(ARM_COMPUTE_ENABLE_BF16) */ #ifdef __ARM_FEATURE_FP16_VECTOR_ARITHMETIC TEST_SUITE(FP16) @@ -459,9 +1219,39 @@ FIXTURE_DATA_TEST_CASE(RunMixedDataLayout, NEGEMMConvolutionLayerMixedDataLayout // Validate output validate(Accessor(_target), _reference, rel_tolerance_f32, 0.f, float(abs_tolerance_f32)); } +/** Padded weights + * CpuGemmConv2d uses two different paths for reshaping the weights based on if the weight tensor has holes (a common + * way to have "holes" in tensor is via extended paddings) + * + * We only need to test the padded weight path here on a single floating data type and a single layout, because the fallback path is agnostic of them + */ +FIXTURE_DATA_TEST_CASE(RunPaddedWeights, NEGEMMConvolutionLayerPaddedWeightsFixture<float>, framework::DatasetMode::ALL, combine(datasets::SmallConvolutionLayerDataset(), + framework::dataset::make("ReshapeWeights", { true }), + framework::dataset::make("DataType", DataType::F32), + framework::dataset::make("DataLayout", { DataLayout::NHWC }) + )) +{ + // Validate output + validate(Accessor(_target), _reference, rel_tolerance_f32, 0.f, float(abs_tolerance_f32)); +} + +// This very large shape test is required to test heuristic paths where the tensor size is > 1e7 bytes +// and weight dimensions larger than 7 +FIXTURE_DATA_TEST_CASE(RunVeryLarge, NEGEMMConvolutionLayerFixture<float>, framework::DatasetMode::NIGHTLY, + combine(datasets::VeryLargeConvolutionLayerDataset(), + framework::dataset::make("ReshapeWeights", { true }), + framework::dataset::make("DataType", DataType::F32), + framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC }), + NoActivation)) +{ + // Validate output + validate(Accessor(_target), _reference, rel_tolerance_f32, 0.f, float(abs_tolerance_f32)); +} + TEST_SUITE_END() // FP32 TEST_SUITE_END() // Float +// TODO: COMPMID-6596 Extend quantized tests with at least one suite where the weight is padded (the legacy case, see floating point's RunPaddedWeights) template <typename T> using NEGEMMConvolutionLayerQuantizedFixture = ConvolutionValidationQuantizedFixture<Tensor, Accessor, NEConvolutionLayer, T>; template <typename T> @@ -477,12 +1267,17 @@ const auto QuantizedActivationFunctionsDataset = framework::dataset::make("Activ ActivationLayerInfo(ActivationLayerInfo::ActivationFunction::LU_BOUNDED_RELU, 6.f) }); TEST_SUITE(Quantized) +/// @note: Every asymmetric quantized test where there's no fused activation will have its quantization info ignored +/// This is because instead of using the same quantization information for all the tensors, the fixture generates +/// separate quantization info for each input and the output tensor. +/// When we can also support dynamic quantization with the presence of activation, these two versions should be merged +/// again, with the explicitly specified quantization info removed TEST_SUITE(QASYMM8) FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMConvolutionLayerQuantizedFixture<uint8_t>, framework::DatasetMode::ALL, combine(combine(combine(combine(combine(datasets::SmallConvolutionLayerDataset(), framework::dataset::make("ReshapeWeights", { true })), framework::dataset::make("DataType", DataType::QASYMM8)), framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC })), - framework::dataset::make("QuantizationInfo", { QuantizationInfo(2.f / 255.f, 10) })), + framework::dataset::make("QuantizationInfoIfActivationEnabled", { QuantizationInfo(2.f / 255.f, 10) })), QuantizedActivationFunctionsDataset)) { // Validate output @@ -499,7 +1294,7 @@ FIXTURE_DATA_TEST_CASE(RunMixedDataLayout, NEGEMMConvolutionLayerQuantizedFixtur framework::dataset::make("ReshapeWeights", { true })), framework::dataset::make("DataType", DataType::QASYMM8)), framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC })), - framework::dataset::make("QuantizationInfo", { QuantizationInfo(2.f / 255.f, 10) })), + framework::dataset::make("QuantizationInfoIfActivationEnabled", { QuantizationInfo(2.f / 255.f, 10) })), QuantizedActivationFunctionsDataset)) { // Validate output @@ -512,7 +1307,7 @@ FIXTURE_DATA_TEST_CASE(RunSmall, NEGEMMConvolutionLayerQuantizedFixture<int8_t>, framework::dataset::make("ReshapeWeights", { true })), framework::dataset::make("DataType", DataType::QASYMM8_SIGNED)), framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC })), - framework::dataset::make("QuantizationInfo", { QuantizationInfo(0.01f, -10) })), + framework::dataset::make("QuantizationInfoIfActivationEnabled", { QuantizationInfo(0.01f, -10) })), QuantizedActivationFunctionsDataset)) { // Validate output @@ -529,7 +1324,7 @@ FIXTURE_DATA_TEST_CASE(RunMixedDataLayout, NEGEMMConvolutionLayerQuantizedFixtur framework::dataset::make("ReshapeWeights", { true })), framework::dataset::make("DataType", DataType::QASYMM8_SIGNED)), framework::dataset::make("DataLayout", { DataLayout::NCHW, DataLayout::NHWC })), - framework::dataset::make("QuantizationInfo", { QuantizationInfo(2.f / 255.f, 10) })), + framework::dataset::make("QuantizationInfoIfActivationEnabled", { QuantizationInfo(2.f / 255.f, 10) })), QuantizedActivationFunctionsDataset)) { // Validate output @@ -562,6 +1357,27 @@ FIXTURE_DATA_TEST_CASE(RunSmallSigned, NEGEMMConvolutionLayerQuantizedPerChannel // Validate output validate(Accessor(_target), _reference, tolerance_qasymm8); } + +FIXTURE_DATA_TEST_CASE(MemoryStressLargeChannels, NEGEMMConvolutionLayerQuantizedPerChannelFixture<int8_t>, + framework::DatasetMode::ALL, + combine( + make("In", TensorShape(1U)), + make("Weights", TensorShape(1U, 1U, 1U, 17000U)), + make("Biases", TensorShape(17000U)), + make("Out", TensorShape(1U, 1U, 17000U)), + make("Info", PadStrideInfo(1, 1, 0, 0)), + make("Dilation", Size2D(1, 1)), + make("ReshapeWeights", { true }), + make("DataType", { DataType::QASYMM8_SIGNED }), + make("DataLayout", { DataLayout::NHWC }), + make("QuantizationInfo", QuantizationInfo(0.5f, 10)), + make("ActivationInfo", ActivationLayerInfo()), + make("WeightsDataType", { DataType::QSYMM8_PER_CHANNEL }))) +{ + // Validate output + validate(Accessor(_target), _reference, tolerance_qasymm8); +} + TEST_SUITE_END() // QSYMM8_PER_CHANNEL TEST_SUITE_END() // Quantized @@ -571,6 +1387,99 @@ TEST_SUITE(DirectGEMMConv2d) template <typename T> using NEDirectGEMMConv2dLayerFixture = ConvolutionValidationFixture<Tensor, Accessor, NEGEMMConv2d, T>; +/** Test case for memory injection in @ref cpu::CpuGemmDirectConv2d. + * + * Configure the operator once and inject memory at run-time in multiple executions. + * + * Checks performed in order: + * - Both runs compute the same output + */ +TEST_CASE(MemoryInjection, framework::DatasetMode::ALL) +{ + auto conv = std::make_unique<cpu::CpuGemmDirectConv2d>(); + const auto src_info = TensorInfo(TensorShape(1U, 5U, 2U), 1, DataType::F32, DataLayout::NHWC); + const auto weight_info = TensorInfo(TensorShape(1U, 3U, 2U, 3U), 1, DataType::F32, DataLayout::NHWC); + const auto bias_info = TensorInfo(TensorShape(3U), 1, DataType::F32, DataLayout::NHWC); + auto dst_info = TensorInfo(TensorShape(1U, 7U, 3U), 1, DataType::F32, DataLayout::NHWC); + const auto conv_info = Conv2dInfo{}; + conv->configure(&src_info, &weight_info, &bias_info, &dst_info, conv_info); + + // tensors are newly created every call of this lambda function + auto src = create_tensor<Tensor>(src_info); + auto weight = create_tensor<Tensor>(weight_info); + auto bias = create_tensor<Tensor>(bias_info); + src.allocator()->allocate(); + weight.allocator()->allocate(); + bias.allocator()->allocate(); + + ITensorPack run_pack{ { TensorType::ACL_SRC_0, &src }, { TensorType::ACL_SRC_1, &weight }, { TensorType::ACL_SRC_2, &bias } }; + ITensorPack prep_pack{ { TensorType::ACL_SRC_1, &weight }, { TensorType::ACL_SRC_2, &bias } }; + + auto mg = MemoryGroup{}; + auto ws = manage_workspace<Tensor>(conv->workspace(), mg, run_pack, prep_pack); + + auto run_conv = [&]() -> Tensor + { + auto dst = create_tensor<Tensor>(dst_info); + dst.allocator()->allocate(); + run_pack.add_tensor(TensorType::ACL_DST, &dst); + + library->fill_tensor_value(Accessor(src), 1.f); + library->fill_tensor_value(Accessor(weight), 2.f); + library->fill_tensor_value(Accessor(bias), 3.f); + // This operator is configured once and captured by this lambda. + conv->prepare(prep_pack); + conv->run(run_pack); + return dst; + }; + auto result_0 = run_conv(); + auto result_1 = run_conv(); + for(size_t i = 0; i < result_0.info()->tensor_shape().total_size(); ++i) + { + ARM_COMPUTE_EXPECT(((float *)result_0.buffer())[i] == ((float *)result_1.buffer())[i], framework::LogLevel::ERRORS); + } +} + +/** Test case for memory injection in @ref NEGEMMConv2d. + * + * Make sure @ref NEGEMMConv2d still works through injecting the memory at configure time using the old API. + * + * Checks performed in order: + * - Both runs compute the same output + */ +TEST_CASE(MultipleExecutionWithConfigure, framework::DatasetMode::ALL) +{ + auto conv = std::make_unique<NEGEMMConv2d>(); + const auto src_info = TensorInfo(TensorShape(1U, 5U, 2U), 1, DataType::F32, DataLayout::NHWC); + const auto weight_info = TensorInfo(TensorShape(1U, 3U, 2U, 3U), 1, DataType::F32, DataLayout::NHWC); + const auto bias_info = TensorInfo(TensorShape(3U), 1, DataType::F32, DataLayout::NHWC); + auto dst_info = TensorInfo(TensorShape(1U, 7U, 3U), 1, DataType::F32, DataLayout::NHWC); + const auto conv_info = Conv2dInfo{}; + auto run_conv = [&]() + { + auto src = create_tensor<Tensor>(src_info); + auto weight = create_tensor<Tensor>(weight_info); + auto bias = create_tensor<Tensor>(bias_info); + auto dst = create_tensor<Tensor>(dst_info); + conv->configure(&src, &weight, &bias, &dst, conv_info); + src.allocator()->allocate(); + weight.allocator()->allocate(); + bias.allocator()->allocate(); + dst.allocator()->allocate(); + library->fill_tensor_value(Accessor(src), 1.f); + library->fill_tensor_value(Accessor(weight), 2.f); + library->fill_tensor_value(Accessor(bias), 3.f); + conv->run(); + return dst; + }; + auto result_0 = run_conv(); + auto result_1 = run_conv(); + for(size_t i = 0; i < result_0.info()->tensor_shape().total_size(); ++i) + { + ARM_COMPUTE_EXPECT(((float *)result_0.buffer())[i] == ((float *)result_1.buffer())[i], framework::LogLevel::ERRORS); + } +} + TEST_SUITE(Float) TEST_SUITE(FP32) FIXTURE_DATA_TEST_CASE(RunSmall, NEDirectGEMMConv2dLayerFixture<float>, framework::DatasetMode::ALL, combine(combine(combine(combine(datasets::SmallConvolutionLayerDataset(), |