Search Results for "parallel computing"

Xtuner is a large-scale training engine designed for efficient training and fine-tuning of modern large language models, particularly mixture-of-experts architectures. The framework focuses on enabling scalable training for extremely large models while maintaining efficiency across distributed computing environments. Unlike traditional 3D parallel training strategies, XTuner introduces optimized parallelism techniques that simplify scaling and reduce system complexity when training massive models. The engine supports training models with hundreds of billions of parameters and enables long-context training with sequence lengths reaching tens of thousands of tokens. ...

Parallax is a decentralized inference framework designed to run large language models across distributed computing resources. Instead of relying on centralized GPU clusters in data centers, the system allows multiple heterogeneous machines to collaborate in serving AI inference workloads. Parallax divides model layers across different nodes and dynamically coordinates them to form a complete inference pipeline. A two-stage scheduling architecture determines how model layers are allocated to...

Mars is a distributed computing framework designed to scale scientific computing and data science workloads across large clusters while preserving the familiar programming interfaces of common Python libraries. The project provides a tensor-based execution model that extends the capabilities of tools such as NumPy, pandas, and scikit-learn so that large datasets can be processed in parallel without rewriting code for distributed environments.

Genetic Programming in OpenCL is a parallel implementation of genetic programming targeted at heterogeneous devices, such as CPU and GPU. It is written in OpenCL, an open standard for portable parallel programming across many computing platforms.

For the purpose of policy simulation in coupled social-ecological systems (e.g. energy supply), a credible modelling of actors – especially citizens – and their decision processes is needed. This requires a framework capable of handling high numbers of heterogeneous agents (several hundreds of thousands). LARA (Lightweight Architecture for boundedly Rational Agents) meets these requirements and fills the gap between frameworks without built-in psychological foundations and full-fledged...

MapPSO is a tool for Ontology Alignment, which uses Discrete Particle Swarm Optimisation. A particle swarm is used to search for the optimal alignment. The algorithm is massively parallel and adapts naturally on parallel architectures.

Platform for parallel computation in the Amazon cloud, including machine learning ensembles written in R for computational biology and other areas of scientific research. Home to MR-Tandem, a hadoop-enabled fork of X!Tandem peptide search engine.

GAP (Grid Agents Platform) Toolkit for Modeling and Simulation of Mobile Agents in Grid Environments. GAP is an abstraction over GridSim, a Grid Simulation Toolkit for Resource Modelling and Application Scheduling for Parallel and Distributed Computing.