Calculate speedup between baseline and evolved solution. Speedup = (Baseline Time) / (Evolved Time) Higher is better. Args:

Calculate speedup between baseline and evolved solution. Speedup = (Baseline Time) / (Evolved Time) Higher is better. Args:

Calculate speedup between baseline and evolved solution. Speedup = (Baseline Time) / (Evolved Time) Higher is better. Args:

Calculate speedup between baseline and evolved solution. Speedup = (Baseline Time) / (Evolved Time) Higher is better. Args:

Calculate speedup between baseline and evolved solution. Speedup = (Baseline Time) / (Evolved Time) Higher is better. Args:

Calculate speedup between baseline and evolved solution. Speedup = (Baseline Time) / (Evolved Time) Higher is better. Args:

Calculate speedup factor based on applied optimizations

Checks that all points are inside the triangle with vertices (0,0), (1,0), (0.5, sqrt(3)/2). Args: points: Array of 2D points to check

Places 21 non-overlapping circles inside a rectangle of perimeter 4 in order to maximize the sum of their radii. Returns: circles: n

Collect all config/*config*.yaml and examples/**/*config*.yaml files

Apply optimizations and compile MLIR

Compile TSP.cpp in `dir_path`, creating an executable `dir_path/bin/runner` Raises: FileNotFoundError, TimeoutError, RuntimeError

Compute the maximum possible radii for each circle position such that they don't overlap and stay within the unit square. Args:

Compute the maximum possible radii for each circle position such that they don't overlap and stay within the unit square. Args:

Computes base metrics after filtering NaNs from predictions. Ensures inputs y_pred and y are treated as 1D arrays.

Configure a data processing pipeline with 4 independent modules. Each module choice is independent - changing one doesn't affect what's

Construct a specific arrangement of 26 circles in a unit square that attempts to maximize the sum of their radii. Returns: Tuple

Construct a specific arrangement of 26 circles in a unit square that attempts to maximize the sum of their radii. Returns: Tuple

Construct an optimized arrangement of 26 circles in a unit square using mathematical principles and optimization techniques. Returns:

Create a simple baseline MLIR attention implementation

Create a YAML configuration file for the symbolic regression task. Args: problem: Dictionary containing problem data Returns:

Create an evaluator script for the symbolic regression problem. The evaluator assesses model performance using BFGS optimization and com

Create prompt asking LLM to improve the program.

Create a fully lowered MLIR file

Create a Python script with a naive linear model for symbolic regression. The generated script contains a `func(x, params)` that computes pr

Create and track a task in the pool Args: coro: Coroutine function to run *args: Arguments to pass to the co

Demonstrate how different models get different parameters

Discover all AlgoTune task directories. Args: algotune_dir: Path to examples/algotune directory Returns: Li

Enhanced version with trend preservation using weighted moving average. Args: x: Input signal (1D array of real-valued samples)

Estimate performance based on IR analysis

Main evaluation function that tests the signal processing algorithm on multiple test signals and calculates the composite performance metric.

🛡️ BULLETPROOF evaluation function called by OpenEvolve

Evaluate the program module located in the specified directory.

Evaluate the program by running it once and checking the sum of radii Args: program_path: Path to the program file Returns:

Enhanced evaluation with baseline comparison for fft_cmplx_scipy_fftpack task. This evaluator: 1. Loads the evolved solve method fro

Enhanced evaluation with baseline comparison for eigenvectors_complex task. This evaluator: 1. Loads the evolved solve method from i

Enhanced evaluation with baseline comparison for convolve2d_full_fill task. This evaluator: 1. Loads the evolved solve method from i

Enhanced evaluation with baseline comparison for affine_transform_2d task. This evaluator: 1. Loads the evolved solve method from in

Enhanced evaluation with baseline comparison for polynomial_real task. This evaluator: 1. Loads the evolved solve method from initia

Enhanced evaluation with baseline comparison for psd_cone_projection task. This evaluator: 1. Loads the evolved solve method from in

Enhanced evaluation with baseline comparison for fft_convolution task. This evaluator: 1. Loads the evolved solve method from initia

Enhanced evaluation with baseline comparison for lu_factorization task. This evaluator: 1. Loads the evolved solve method from initi

Entry point for OpenEvolve

Evaluate the program by running it once and checking the sum of radii Args: program_path: Path to the program file Returns:

Evaluate the program by running it multiple times and checking how close it gets to the known global minimum. Args: program_path

Evaluate an R program implementing robust regression. Tests the program on synthetic data with outliers to measure: - Accuracy (MSE, MAE

High-level, single-call evaluation function. This orchestrates the entire process: 1. Infers the task name. 2. Fits the model on tra

BULLETPROOF evaluation that handles ALL Metal kernel failures: 1. Enhanced program extraction with syntax validation 2. Pre-e

Evaluate the accuracy of extracted documentation. Args: docs: Extracted documentation expected: Expected results Return

Stage 2 evaluation: Full evaluation with all samples Args: prompt_path: Path to the prompt file Returns: Dictionary wit

Evaluates a model by loading it, optimizing its parameters, and testing it. The model function from program_path is expected to be named 'f

Evolve an algorithm class based on a benchmark Args: algorithm_class: Initial algorithm class to evolve benchmark: Function

Export traces to HDF5 format Args: traces: List of trace objects with to_dict() method output_path: Path to save the HDF5 fi

Export traces to JSONL format (one JSON object per line) Args: traces: List of trace objects with to_dict() method output_pa

Extract Python code from LLM response.

Try to determine the language of a code snippet Args: code: Code snippet Returns: Detected language or "unknown"

Extract data for a specific problem from an initialized dataset. Args: initialized_dataset: Pre-initialized and setup dataset object

Find what lowering passes are available

Find the best_program.py file in the expected location

Find a free port starting from start_port

Format extracted documentation into a readable string. Args: api_docs: List of API documentation dictionaries Returns:

Safely format improvement metrics for logging. Args: parent_metrics: Parent program metrics child_metrics: Child program met

Format rich feedback if available (RICH_FEEDBACK=1).

Create from dictionary representation

Generate text using a all available models and average their returned metrics

Generate detailed feedback for the LLM to improve the scraper. This feedback will be included in the evolution prompt to guide the LLM t

Generate a random pipeline configuration.

Generate summary statistics from task results. Args: task_results: Dictionary of task results Returns: Summ

Generate benchmark summary statistics

Generate synthetic test signal with known characteristics. Args: length: Length of the signal noise_level: Standard deviatio

Generate multiple test signals with different characteristics

Generate text using a system message and conversational context

Get list of artifact keys

Calculates the precise C4 bound from a final set of parameters.

Returns a ConfigParser object for the .kattisrc file(s)