#vec #signed #performance

signvec

Vector implementation for fast, sign-based manipulation of dynamic collections

4 releases (2 breaking)

0.3.0 Mar 23, 2024
0.2.0 Mar 22, 2024
0.1.1 Mar 22, 2024
0.1.0 Mar 14, 2024

#335 in Data structures

22 downloads per month

MIT license

125KB
2K SLoC

SignVec

Crates.io docs.rs License GitHub Workflow Status

SignVec extends the capabilities of the traditional Vec by providing additional functionalities to efficiently track and manipulate elements based on their sign (positive or negative) using the Signable trait.

Features

  • Tracks the sign of elements for optimized sign-specific operations.
  • Provides methods for element counting, access, and manipulation based on sign.
  • Integrates with user-defined types via the Signable trait.

Usage: Basic operations

use nanorand::{Rng, WyRand};
use signvec::{svec, Sign, SignVec};

fn main() {
    let mut vector = svec![1, -2, 3, -4];
    assert_eq!(vector.len(), 4);
    assert_eq!(vector[0], 1);
    assert_eq!(vector[1], -2);

    // Count positive and negative elements
    assert_eq!(vector.count(Sign::Plus), 2);
    assert_eq!(vector.count(Sign::Minus), 2);

    // Get indices of positive and negative elements
    assert_eq!(
        vector.indices(Sign::Plus).iter().collect::<Vec<_>>(),
        vec![&0, &2]
    );
    assert_eq!(
        vector.indices(Sign::Minus).iter().collect::<Vec<_>>(),
        vec![&1, &3]
    );

    // Retrieve values based on their sign
    assert_eq!(vector.values(Sign::Plus).collect::<Vec<_>>(), vec![&1, &3]);
    assert_eq!(
        vector.values(Sign::Minus).collect::<Vec<_>>(),
        vec![&-2, &-4]
    );

    // Modify individual elements
    vector.set(1, 5);
    assert_eq!(vector[1], 5);

    // Randomly select an element based on its sign
    let mut rng = WyRand::new();
    if let Some(random_positive) = vector.random(Sign::Plus, &mut rng) {
        println!("Random positive value: {}", random_positive);
    }
}

Usage: Monte Carlo simulations

This demonstrates a simple Monte Carlo simulation where site energies in a SignVec are updated based on simulated dynamics and system energy distribution.

use signvec::{SignVec, svec, Sign, Signable};
use nanorand::{WyRand, Rng};

fn main() {
    let mut energies = svec![1.0, -1.0, 1.5, -1.5, 0.5, -0.5];
    let mut rng = WyRand::new();

    // Simulation loop for multiple Monte Carlo steps
    for _step in 0..100 {
        let site = rng.generate_range(0..energies.len());
        let dE = rng.generate::<f64>() - 0.5; // Change in energy

        let new_energy = energies[site] + dE; // Update site energy
        
        // Make decisions based on system's energy distribution
        if energies.count(Sign::Minus) > energies.count(Sign::Plus) {
            if energies[site].sign() == Sign::Minus && rng.generate::<f64>() < 0.5 {
                energies.set(site, -new_energy); // Flip energy sign
            } else {
                energies.set(site, new_energy);
            }
        } else {
            energies.set(site, new_energy); // Balanced distribution
        }
    }

    println!("Final energy distribution: {:?}", energies);
}

Usage: Portfolio management

Demonstrates how SignVec can be used for managing a financial portfolio, simulating market conditions, and making decisions based on the sign-aware characteristics of assets and liabilities.

use signvec::{SignVec, Sign, svec};
use nanorand::WyRand;

fn main() {
    let mut portfolio = svec![150.0, -200.0, 300.0, -50.0, 400.0];
    let market_conditions = vec![1.05, 0.95, 1.10, 1.00, 1.03];

    // Apply market conditions to adjust portfolio balances
    for (index, &factor) in market_conditions.iter().enumerate() {
        portfolio.set(index, portfolio[index] * factor);
    }

    // Decision making based on portfolio's sign-aware characteristics
    if portfolio.count(Sign::Minus) > 2 {
        println!("Consider rebalancing your portfolio to manage risk.");
    } else {
        println!("Your portfolio is well-balanced and diversified.");
    }

    // Calculate 10% of total liabilities for debt reduction
    let debt_reduction = portfolio.values(Sign::Minus).sum::<f64>() * 0.1;
    println!("Plan for a debt reduction of ${:.2} to strengthen your financial position.", debt_reduction.abs());

    // Identify a high-performing asset for potential investment
    let mut rng = WyRand::new();
    if let Some(lucky_asset) = portfolio.random(Sign::Plus, &mut rng) {
        println!("Consider investing more in an asset valued at ${:.2}.", lucky_asset);
    } else {
        println!("No standout assets for additional investment at the moment.");
    }
}

Benchmarks

The table below is a summary of benchmark results for the specialized functionality of SignVec.

Operation SignVec Vec Speedup
set 1.3922 ns - -
set_unchecked 1.3873 ns - -
random (Plus) 822.90 ps - -
random (Minus) 829.92 ps - -
random_pos 652.96 ps - -
random_neg 687.77 ps - -
count (Plus) 453.21 ps - -
count (Minus) 458.70 ps - -
count_pos 229.73 ps - -
count_neg 228.44 ps - -
indices (Plus) 465.04 ps - -
indices (Minus) 461.85 ps - -
indices_pos 226.99 ps - -
indices_neg 225.83 ps - -
sync 61.208 µs - -
count[^1] 225.74 ps 153.38 ns ~679x faster
indices[^2] 86.42 ns 1.11 µs ~12.8x faster
values[^3] 579.37 ns 1.13 µs ~1.95x faster
random[^4] 857.86 ps 950.84 ns ~1106x faster

[^1]: The count_pos and count_neg benchmarks are used here for comparison as they represent the optimized paths for counting elements by sign. [^2]: The indices_pos and indices_neg benchmarks are presented as the optimized methods for retrieving indices by sign. [^3]: The values operation does not have a direct counterpart in the provided benchmarks but is included for context. [^4]: The random_pos and random_neg benchmarks provide context for the random operation's performance when the sign is predetermined.

Benchmarks were conducted on a machine with the following specifications:

  • Processor: AMD Ryzen™ 5 5600G with Radeon™ Graphics x 12
  • Memory: 58.8 GiB
  • Operating System: Guix System
  • OS Type: 64-bit

Dependencies

~0.6–1.1MB
~25K SLoC