Slurm: A Highly Scalable Workload Manager

Library for specialized dense and sparse matrix operations, and deep learning primitives.

The examples for the courses on advanced programming for scientific computing (AMSC and APSC), Politecnico di Milano

Utilities for Dask and CUDA interactions

Using the famous cnn model in Pytorch, we run benchmarks on various gpu.

A fast and scalable x86-64 multicore simulator

The simplest, fastest repository for training/finetuning medium-sized GPTs. Now, with kittens!

Slurm: A Highly Scalable Workload Manager

Extend the original llama.cpp repo to support redpajama model.

Slurm HPC workload manager web JS dashboard and JSON REST API

Custom Spawner for Jupyterhub to start slurm jobs when users log in

Add gpu utilization stats to Slurm batch scheduler accounting db

A Python package for exporting the weights and biases of neural networks.

A high-throughput and memory-efficient inference and serving engine for LLMs

A simple memory (or cache) bandwidth benchmark that isn't stream

Ongoing research training transformer language models at scale, including: BERT & GPT-2

Best practice for training LLaMA models in Megatron-LM

Ongoing research training transformer models at scale

vLLM: A high-throughput and memory-efficient inference and serving engine for LLMs

Fast and memory-efficient exact attention

Embedded Scalable Platforms: Heterogeneous SoC architecture and IP integration made easy

pocl - Portable Computing Language

Open source code for AlphaFold.