Related papers: Demystifying NCCL: An In-depth Analysis of GPU Com…

The Continuous Learning (CL) paradigm consists of continuously evolving the parameters of the Deep Neural Network (DNN) model to progressively learn to perform new tasks without reducing the performance on previous tasks, i.e., avoiding the…

The emergence of Large Language Models (LLMs) has necessitated the adoption of distributed training techniques, involving the deployment of thousands of GPUs to train a single model. Unfortunately, the efficiency of large-scale distributed…

RAPID-LLM is a unified performance modeling framework for large language model (LLM) training and inference on GPU clusters. It couples a DeepFlow-based frontend that generates hardware-aware, operator-level Chakra execution traces from an…

Concurrent computation and communication (C3) is a pervasive paradigm in ML and other domains, making its performance optimization crucial. In this paper, we carefully characterize C3 in ML on GPUs, which are most widely deployed for ML…

Distributed inference of large language models (LLMs) using tensor parallelism can introduce communication overheads of $20$% even over GPUs connected via NVLink, a high-speed GPU interconnect. Several techniques have been proposed to…

Convolutional neural networks (CNNs) have achieved great success on grid-like data such as images, but face tremendous challenges in learning from more generic data such as graphs. In CNNs, the trainable local filters enable the automatic…

Large Language Models (LLMs) have resulted in a surging demand for planet-scale serving systems, where tens of thousands of GPUs continuously serve hundreds of millions of users. Consequently, throughput has emerged as a key metric that…

Spatial computing architectures pose an attractive alternative to mitigate control and data movement overheads typical of load-store architectures. In practice, these devices are rarely considered in the HPC community due to the steep…

Graphs are essential for modeling complex interactions across domains such as social networks, biology, and recommendation systems. Traditional Graph Neural Networks, particularly Message Passing Neural Networks (MPNNs), rely heavily on…

Distributed ML workloads rely heavily on collective communication across multi-GPU, multi-node systems. Emerging scale-up fabrics, such as NVLink and UALink, enable direct memory access across nodes but introduce a critical destination-side…

Training convolutional neural networks (CNNs) requires intense compute throughput and high memory bandwidth. Especially, convolution layers account for the majority of the execution time of CNN training, and GPUs are commonly used to…

Although recent works try to improve collective communication in grid systems by separating intra and inter-cluster communication, the optimisation of communications focus only on inter-cluster communications. We believe, instead, that the…

The widespread adoption of convolutional neural networks (CNNs) in resource-constrained scenarios has driven the development of Machine Learning as a Service (MLaaS) system. However, this approach is susceptible to privacy leakage, as the…

Large Language Models (LLMs) built on transformer architectures have transformed natural language processing, achieving remarkable performance across diverse applications. While distributed inference frameworks enable practical deployment…

Monolithic serving with chunked prefill improves GPU utilization by batching prefill and decode together, but suffers from fine-grained phase interference. Engine-level prefill-decode (PD) disaggregation avoids interference but incurs…

NPCs in traditional games are often limited by static dialogue trees and a single platform for interaction. To overcome these constraints, this study presents a prototype system that enables large language model (LLM)-powered NPCs to…

The IBM Neural Computer (INC) is a highly flexible, re-configurable parallel processing system that is intended as a research and development platform for emerging machine intelligence algorithms and computational neuroscience. It consists…

Large-scale Transformer language models (LMs) trained solely on next-token prediction with web-scale data can solve a wide range of tasks after seeing just a few examples. The mechanism behind this capability, known as in-context learning…

Characterizing and predicting the training performance of modern machine learning (ML) workloads on compute systems with compute and communication spread between CPUs, GPUs, and network devices is not only the key to optimization and…

The pervasive adoption of Deep Learning (DL) and Graph Processing (GP) makes it a de facto requirement to build large-scale clusters of heterogeneous accelerators including GPUs and FPGAs. The OpenCL programming framework can be used on the…