Modular RDMA Interface

Kanban board to manage your AI coding agents

A fast communication-overlapping library for tensor/expert parallelism on GPUs.

Perplexity GPU Kernels

Introduction to Machine Learning Systems

ArcticInference: vLLM plugin for high-throughput, low-latency inference

Manage your GPUs across NVIDIA, AMD, Intel, and Apple Silicon systems.

FlashInfer: Kernel Library for LLM Serving

Real-time terminal monitor for InfiniBand networks - htop for high-speed interconnects

RDMA core userspace libraries and daemons

Open MPI main development repository

LLM inference in C/C++

Notes for using Julia while learning calculus

MIT IAP short course: Matrix Calculus for Machine Learning and Beyond

Code to create the figures and solve the exercises in the textbook

Master calculus 1 using Python: derivatives and applications

Distributed View Extension for Kokkos

Examples demonstrating available options to program multiple GPUs in a single node or a cluster

NCCL Examples from Official NVIDIA NCCL Developer Guide.

Pretrained model hub for Keras 3.

example code for using DC QP for providing RDMA READ and WRITE operations to remote GPU memory

Sample examples of how to call collective operation functions on multi-GPU environments. A simple example of using broadcast, reduce, allGather, reduceScatter and sendRecv operations.

High Performance Computing Project 2021-2022.

Implementation of the allreduce algorithm using only MPI point to point communication routines (MPI_Send, MPI_Recv).

MSCCL++: A GPU-driven communication stack for scalable AI applications

Microsoft Collective Communication Library