Crate torsh_benches 

Benchmarking suite for ToRSh

This crate provides comprehensive benchmarks for ToRSh to measure performance against other tensor libraries and track regressions.

Architecture

The benchmarking suite is organized into specialized modules:

• core: Core comparison infrastructure and performance analysis
• ndarray_comparisons: ToRSh vs ndarray baseline comparisons
• pytorch_comparisons: PyTorch integration benchmarks (feature-gated)
• tensorflow_comparisons: TensorFlow comparison suite (feature-gated)
• jax_comparisons: JAX performance benchmarks (feature-gated)
• numpy_comparisons: NumPy baseline comparisons (feature-gated)
• reporting: Comprehensive reporting and analysis utilities

Additional specialized benchmarks include model-specific tests, hardware optimization benchmarks, edge deployment tests, and SciRS2 integration.

Re-exports

pub use benchmarks::AdvancedSystemsBenchmarkSuite;

pub use benchmarks::AutoTuningBench;

pub use benchmarks::BackwardPassBench;

pub use benchmarks::DataLoaderThroughputBench;

pub use benchmarks::ErrorDiagnosticsBench;

pub use benchmarks::GradientComputeBench;

pub use benchmarks::MatmulBench;

pub use benchmarks::SIMDGNNBench;

pub use benchmarks::TensorArithmeticBench;

pub use benchmarks::TensorCreationBench;

pub use benchmarks::VectorizedMetricsBench;

pub use core::ComparisonResult;

pub use core::ComparisonRunner;

pub use core::PerformanceAnalyzer;

pub use reporting::benchmark_and_analyze;

pub use reporting::benchmark_and_compare;

pub use reporting::generate_master_comparison_report;

pub use reporting::run_all_comparison_suites;

pub use reporting::run_comparison_benchmarks;

pub use reporting::run_extended_benchmarks;

pub use metrics::CpuStats;

pub use metrics::MemoryStats;

pub use metrics::MetricsCollector;

pub use metrics::PerformanceReport;

pub use metrics::SystemMetrics;

pub use utils::DataGenerator;

pub use utils::Distribution;

pub use utils::EnhancedBenchResult;

pub use utils::EnhancedBenchSuite;

pub use utils::Environment;

pub use utils::EnvironmentInfo;

pub use utils::Formatter;

pub use utils::MemoryMonitor;

pub use utils::ParallelBenchRunner;

pub use utils::Timer;

pub use utils::TimingStats;

pub use utils::ValidationResult;

pub use utils::Validator;

pub use model_benchmarks::ModelBenchmarkSuite;

pub use model_benchmarks::ResNetBlockBench;

pub use model_benchmarks::TransformerBlockBench;

pub use hardware_benchmarks::CPUGPUComparisonBench;

pub use hardware_benchmarks::MemoryBandwidthBench;

pub use hardware_benchmarks::MultiGPUBench;

pub use hardware_benchmarks::ThermalThrottlingBench;

pub use scalability::ScalabilityTestSuite;

pub use precision_benchmarks::MixedPrecisionTrainingBench;

pub use precision_benchmarks::PruningBench;

pub use precision_benchmarks::QuantizationBench;

pub use custom_ops_benchmarks::ConvolutionOperation;

pub use custom_ops_benchmarks::CustomOpBench;

pub use custom_ops_benchmarks::FFTOperation;

pub use edge_deployment::BatteryLifeBench;

pub use edge_deployment::EdgeInferenceBench;

pub use edge_deployment::EdgeMemoryBench;

pub use mobile_benchmarks::ARMOptimizationBench;

pub use mobile_benchmarks::MobileGPUBench;

pub use mobile_benchmarks::MobilePlatformBench;

pub use wasm_benchmarks::BrowserSpecificBench;

pub use wasm_benchmarks::WASMPerformanceBench;

pub use wasm_benchmarks::WebDeploymentBench;

pub use scirs2_benchmarks::AdvancedNeuralNetworkBench;

pub use scirs2_benchmarks::AdvancedOptimizerBench;

pub use scirs2_benchmarks::GraphNeuralNetworkBench;

pub use scirs2_benchmarks::SciRS2BenchmarkSuite;

pub use scirs2_benchmarks::SciRS2MathBench;

pub use scirs2_benchmarks::SciRS2RandomBench;

pub use scirs2_benchmarks::SpatialVisionBench;

pub use scirs2_benchmarks::TimeSeriesAnalysisBench;

pub use advanced_analysis::AdaptiveBenchmarking;

pub use advanced_analysis::AdvancedAnalyzer;

pub use benchmark_analysis::BenchmarkAnalyzer;

pub use benchmark_analysis::BottleneckAnalysis;

pub use benchmark_analysis::PerformanceAnalysis;

pub use benchmark_analysis::PerformanceRating;

pub use benchmark_validation::BenchmarkValidator;

pub use benchmark_validation::NumericalAccuracy;

pub use benchmark_validation::ValidationConfig;

pub use ci_integration::CIBenchmarkRunner;

pub use ci_integration::CIConfig;

pub use ci_integration::NotificationConfig;

pub use html_reporting::HtmlReportGenerator;

pub use html_reporting::Theme;

pub use performance_dashboards::DashboardConfig;

pub use performance_dashboards::PerformanceDashboard;

pub use regression_detection::AdvancedRegressionDetector;

pub use regression_detection::RegressionAnalysis;

pub use system_info::BenchmarkEnvironment;

pub use system_info::SystemInfo;

pub use system_info::SystemInfoCollector;

pub use visualization::ChartType;

pub use visualization::VisualizationGenerator;

Modules

advanced_analysis

Advanced performance analysis and micro-benchmarking tools

benchmark_analysis

Enhanced benchmark analysis and reporting

benchmark_cache

Benchmark result caching system

benchmark_comparison

Advanced benchmark comparison and result aggregation utilities

benchmark_validation

Benchmark correctness validation and cross-architecture verification

benchmarks

Modular Benchmark System for ToRSh

cached_runner

Cached benchmark runner utilities

ci_integration

CI Integration for ToRSh Benchmarks

comparisons

Legacy comparison interface that delegates to specialized modules

core

Core infrastructure for ToRSh performance comparisons

custom_ops_benchmarks

Custom operations benchmarks for ToRSh

distributed_training

Distributed Training Benchmarks

edge_deployment

Edge deployment benchmarks for ToRSh

hardware_benchmarks

Hardware-specific benchmarks

html_reporting

Advanced HTML reporting for ToRSh benchmarks

metrics

Comprehensive system metrics collection for benchmarks

mobile_benchmarks

Mobile performance benchmarks for ToRSh

model_benchmarks

Model architecture benchmarks

ndarray_comparisons

ndarray library comparisons

performance_dashboards

Performance Dashboards for ToRSh Benchmarks

precision_benchmarks

Mixed precision and quantization benchmarks

prelude

Common benchmark utilities

regression_detection

Advanced Regression Detection System for ToRSh Benchmarks

reporting

General benchmark reporting and analysis utilities for ToRSh

scalability

Scalability testing framework for ToRSh operations

scirs2_benchmarks

SciRS2 Integration Benchmarks

system_info

Enhanced system information gathering for benchmarking

utils

Utility functions for benchmarking

visualization

Advanced Visualization Tools for ToRSh Benchmarks

wasm_benchmarks

WebAssembly benchmarks for ToRSh

Macros

benchmark

Macro to create a simple benchmark

profile

Macro for scoped profiling

Structs

BenchConfig

Benchmark configuration

BenchResult

Benchmark result

BenchRunner

Benchmark runner

Traits

Benchmarkable

Trait for benchmarkable operations