biren-surand

What problem does it solve?

This Skill provides a GPU-accelerated library for generating high-quality random numbers, essential for AI training, simulations, and probabilistic modeling.

Core Features & Use Cases

• Multiple Algorithms: Supports XORWOW, Philox, and MTGP32 for diverse randomness needs.
• Flexible APIs: Includes Host API for batch generation and Device API for in-kernel random number generation.
• Use Case: Enable AI model training with stochastic layers or perform Monte Carlo simulations efficiently using GPU-accelerated sampling.

Quick Start

Call the library functions to initialize the generator, generate random data in your application, and clean up resources afterward.