Related papers: MSCCL++: Rethinking GPU Communication Abstractions…

Large-scale LLM training requires collective communication libraries to exchange data among distributed GPUs. As a company dedicated to building and operating large-scale GPU training clusters, we encounter several challenges when using…

GPU-aware collective communication has become a major bottleneck for modern computing platforms as GPU computing power rapidly rises. A traditional approach is to directly integrate lossy compression into GPU-aware collectives, which can…

HiCCL (Hierarchical Collective Communication Library) addresses the growing complexity and diversity in high-performance network architectures. As GPU systems have envolved into networks of GPUs with different multilevel communication…

FPGAs are increasingly prevalent in cloud deployments, serving as Smart NICs or network-attached accelerators. Despite their potential, developing distributed FPGA-accelerated applications remains cumbersome due to the lack of appropriate…

Machine learning models are increasingly being trained across multiple GPUs and servers. In this setting, data is transferred between GPUs using communication collectives such as AlltoAll and AllReduce, which can become a significant…

The NVIDIA Collective Communication Library (NCCL) is a critical software layer enabling high-performance collectives on large-scale GPU clusters. Despite being open source with a documented API, its internal design remains largely opaque.…

Modern distributed ML suffers from a fundamental gap between the theoretical and realized performance of collective communication algorithms due to congestion and hop-count induced dilation in practical GPU clusters. We present PCCL, a…

Collective communication is becoming increasingly important in data center and supercomputer workloads with an increase in distributed AI related jobs. However, existing libraries that provide collective support such as NCCL, RCCL, and…

The rapid growth of large language models is driving organizations to expand their GPU clusters, often with GPUs from multiple vendors. However, current deep learning frameworks lack support for collective communication across heterogeneous…

The increasing scale of large language models (LLMs) necessitates highly efficient collective communication frameworks, particularly as training workloads extend to hundreds of thousands of GPUs. Traditional communication methods face…

Modern high-end systems are increasingly becoming heterogeneous, providing users options to use general purpose Graphics Processing Units (GPU) and other accelerators for additional performance. High Performance Computing (HPC) and…

This report presents the Prime Collective Communications Library (PCCL), a novel fault-tolerant collective communication library designed for distributed ML workloads over the public internet. PCCL introduces a new programming model that…

The rapid growth of large language models (LLMs) has made GPU communication a critical bottleneck. While prior work reduces communication volume via quantization or lossy compression, these approaches introduce numerical errors that can…

Communication has emerged as a critical bottleneck in the distributed training of large language models (LLMs). While numerous approaches have been proposed to reduce communication overhead, the potential of lossless compression has…

Large language models (LLMs) training or inference across multiple nodes introduces significant pressure on GPU memory and interconnect bandwidth. The Compute Express Link (CXL) shared memory pool offers a scalable solution by enabling…

One-sided communication is a useful paradigm for irregular parallel applications, but most one-sided programming environments, including MPI's one-sided interface and PGAS programming languages, lack application level libraries to support…

Collective communication is a major bottleneck for multi-node GPU workloads in scientific computing and distributed deep learning, especially when inter-node bandwidth is limited. Although NCCL provides optimized GPU-centric collectives,…

Distributed deep neural network training necessitates efficient GPU collective communications, which are inherently susceptible to deadlocks. GPU collective deadlocks arise easily in distributed deep learning applications when multiple…

Medical image processing is often limited by the computational cost of the involved algorithms. Whereas dedicated computing devices (GPUs in particular) exist and do provide significant efficiency boosts, they have an extra cost of use in…

Collective communication algorithms are an important component of distributed computation. Indeed, in the case of deep-learning, collective communication is the Amdahl's bottleneck of data-parallel training. This paper introduces SCCL (for…