NanoVector 1.0.2

⏏NanoVector⚡

A high-performance, SIMD-accelerated vector library for .NET. Minimalist, zero-allocation, and engineered for raw speed.

✨ Key Features

• 🚀 SIMD Accelerated — Uses hardware intrinsics (AVX/AVX2) to process multiple numbers in a single CPU cycle.
• 💎 Zero Allocations — Engineered to work with Span<T> and unsafe pointers. No Garbage Collector overhead.
• 🛠 Raw Speed — Bypasses standard .NET bounds checking for maximum throughput.
• 🧩 Simple API — Easy to use for AI, game engines, and math-heavy applications.

📥 Installation

Install via NuGet: dotnet add package NanoVector

💻 Usage

🔧 Basic Operations

• ⏺ Using

  using NanoVector;
  

• 🚀 Create & Sum — Create a vector from an array and calculate the total sum using SIMD power.

  var vector = new MultiVector(new float[] { 1, 2, 3, 4 });
  float result = vector.Sum();
  

• 💎 Sum of Squares — Optimized calculation for AI and statistical variance.

  float squares = vector.SumOfSquares();
  

• 🤖 AI Comparison — Measure cosine similarity between two vectors for neural networks.

  float dotProduct = NanoComparator.DotProduct(vector1, vector2);
  float similarity = NanoComparator.Compare(vector1, vector2);
  

• 🛡️ Read-Only Safety — Use ReadOnlyMultiVector for safe data processing without accidental modifications.

  ReadOnlyMultiVector readOnlyVector = vector.AsReadOnly();
  float safeSum = readOnlyVector.Sum();
  float safeSumSquares = readOnlyVector.SumOfSquares();
  

🧠 Example: Simple Neuron Calculation

This is a conceptual example of how to implement a high-speed neural network layer using NanoVector:

// Let's represent the input signals and the weight of the neuron
// Input signals and trained weights
var inputs = new MultiVector(new float[] { 0.5f, -0.2f, 0.8f, 0.1f });
var weights = new MultiVector(new float[] { 0.15f, 0.45f, -0.1f, 0.9f });
float bias = 0.05f;

// 1. Compute the dot product using SIMD
// Calculate dot product with hardware acceleration
float dot = NanoComparator.DotProduct(inputs, weights);

// 2. Add bias and apply activation (ReLU)
// Add bias and apply ReLU activation
float neuronOutput = MathF.Max(0, dot + bias);

//Showing the result
Console.WriteLine($"Neuron result: {neuronOutput}");

☕ Support NanoVector ⚡

If this library saved your nanoseconds, consider a tip:

ETH / EVM	Bitcoin	Litecoin
0xCCDD9...4AB4	bc1qjmw...uhcd	LgSVCqN...VP3o

• ETH / Polygon / BSC / Base (EVM): 0xCCDD90984C6e9BC4a84e10c402CF85e69f9A4AB4
• Bitcoin (BTC): bc1qjmwjjcqnphrc3qmy6x9pl5qk06a28msdw8uhcd
• Litecoin (LTC): LgSVCqNyt44Rgg7jPTkbiKzskedy5AVP3o

🛡️ License

Distributed under the MIT License. See LICENSE for more information.

Built for nanoseconds. 🧬