API Reference

This page contains the complete API reference for HPCSeries Core, automatically generated from docstrings.

Core Module

HPCSeries Core v0.9 - Python Bindings

High-performance statistical computing library with SIMD vectorization, OpenMP parallelization, adaptive auto-tuning, and composable pipelines.

Examples

>>> import hpcs
>>> import numpy as np
>>> x = np.random.randn(1000000)

# Basic reductions (SIMD-accelerated) >>> hpcs.sum(x) >>> hpcs.mean(x) >>> hpcs.std(x)

# Rolling operations (fast C++ implementation) >>> hpcs.rolling_mean(x, window=50) >>> hpcs.rolling_median(x, window=100)

# Robust statistics (MAD-based outlier detection) >>> hpcs.median(x) >>> hpcs.mad(x) >>> hpcs.robust_zscore(x)

# Anomaly detection >>> anomalies = hpcs.detect_anomalies(x, threshold=3.0)

# v0.8.0: Composable pipelines >>> pipe = hpcs.pipeline(mode=’fast’) >>> pipe.diff(order=1).ewma(alpha=0.2).robust_zscore() >>> result = pipe.execute(x) >>> print(pipe.summary())

Basic Reductions

SIMD-accelerated reduction operations.

`hpcs.sum`
`hpcs.mean`
`hpcs.var`
`hpcs.std`
`hpcs.min`
`hpcs.max`

Robust Statistics

Robust statistical operations using Median Absolute Deviation (MAD).

`hpcs.median`
`hpcs.mad`
`hpcs.quantile`

Execution Mode API (v0.8)

Control execution mode for safety/performance trade-offs.

`hpcs.set_execution_mode`
`hpcs.get_execution_mode`
`hpcs.MODE_SAFE`
`hpcs.MODE_FAST`
`hpcs.MODE_DETERMINISTIC`

Exponential Weighted Statistics (v0.8)

Streaming statistics with exponential decay weighting (15-60x faster than pandas).

`hpcs.ewma`
`hpcs.ewvar`
`hpcs.ewstd`

Time Series Transforms (v0.8)

Differencing, cumulative operations, and convolution.

`hpcs.diff`
`hpcs.cumulative_min`
`hpcs.cumulative_max`
`hpcs.convolve_valid`

Advanced Robust Statistics (v0.8)

Outlier-resistant descriptive statistics (10-15x faster than SciPy).

`hpcs.trimmed_mean`
`hpcs.winsorized_mean`

Transforms & Normalization

Data transformation and normalization functions.

`hpcs.zscore`
`hpcs.robust_zscore`
`hpcs.normalize_minmax`
`hpcs.clip`

Anomaly Detection

Statistical and robust anomaly detection methods.

`hpcs.detect_anomalies`
`hpcs.detect_anomalies_robust`

Rolling Operations

Fast sliding window computations (50-100x faster than Pandas).

`hpcs.rolling_sum`
`hpcs.rolling_mean`
`hpcs.rolling_std`
`hpcs.rolling_var`
`hpcs.rolling_median`
`hpcs.rolling_mad`
`hpcs.rolling_zscore`
`hpcs.rolling_robust_zscore`

Axis Operations

2D array operations along specified axes.

`hpcs.axis_sum`
`hpcs.axis_mean`
`hpcs.axis_median`
`hpcs.axis_mad`
`hpcs.axis_min`
`hpcs.axis_max`
`hpcs.anomaly_axis`
`hpcs.anomaly_robust_axis`

Masked Operations

Operations on arrays with missing data (masked values).

`hpcs.sum_masked`
`hpcs.mean_masked`
`hpcs.var_masked`
`hpcs.median_masked`
`hpcs.mad_masked`
`hpcs.rolling_mean_masked`
`hpcs.rolling_mean_batched`

SIMD & CPU Information

Query SIMD capabilities and CPU topology.

`hpcs.simd_info`
`hpcs.get_simd_width`
`hpcs.get_cpu_info`

Calibration

Performance calibration and configuration management.

`hpcs.calibrate`
`hpcs.save_calibration_config`
`hpcs.load_calibration_config`

Detailed API

hpcs.sum()

hpcs.mean()

hpcs.std()

hpcs.var()

hpcs.min()

hpcs.max()

hpcs.median()

hpcs.mad()

hpcs.quantile()

hpcs.zscore()

hpcs.robust_zscore()

hpcs.normalize_minmax()

hpcs.clip()

hpcs.detect_anomalies()

hpcs.detect_anomalies_robust()

hpcs.rolling_sum()

hpcs.rolling_mean()

hpcs.rolling_std()

hpcs.rolling_var()

hpcs.rolling_median()

hpcs.rolling_mad()

hpcs.rolling_zscore()

hpcs.rolling_robust_zscore()

hpcs.axis_sum()

hpcs.axis_mean()

hpcs.axis_median()

hpcs.axis_mad()

hpcs.axis_min()

hpcs.axis_max()

hpcs.anomaly_axis()

hpcs.anomaly_robust_axis()

hpcs.sum_masked()

hpcs.mean_masked()

hpcs.var_masked()

hpcs.median_masked()

hpcs.mad_masked()

hpcs.rolling_mean_masked()

hpcs.rolling_mean_batched()

hpcs.simd_info()

hpcs.get_simd_width()

hpcs.get_cpu_info()

hpcs.calibrate()

hpcs.save_calibration_config()

hpcs.load_calibration_config()

hpcs.set_execution_mode()

hpcs.get_execution_mode()

Safe execution mode (IEEE 754 compliant, full error checking).

hpcs.MODE_SAFE: int = 0: Use SAFE mode for maximum numerical stability and error detection.

Fast execution mode (relaxed math optimizations, minimal validation).

hpcs.MODE_FAST: int = 1: Use FAST mode for performance-critical code where data is pre-validated.

Deterministic execution mode (bit-exact reproducibility, no OpenMP).

hpcs.MODE_DETERMINISTIC: int = 2: Use DETERMINISTIC mode when exact reproducibility is required across runs.

hpcs.ewma()

hpcs.ewvar()

hpcs.ewstd()

hpcs.diff()

hpcs.cumulative_min()

hpcs.cumulative_max()

hpcs.convolve_valid()

hpcs.trimmed_mean()

hpcs.winsorized_mean()