Related papers: Oases: Efficient Large-Scale Model Training on Com…

Deep learning have achieved promising results on a wide spectrum of AI applications. Larger datasets and models consistently yield better performance. However, we generally spend longer training time on more computation and communication.…

Function-as-a-Service (FaaS) has raised a growing interest in how to "tame" serverless computing to enable domain-specific use cases such as data-intensive applications and machine learning (ML), to name a few. Recently, several systems…

The rising demand for generative large language models (LLMs) poses challenges for thermal and power management in cloud datacenters. Traditional techniques often are inadequate for LLM inference due to the fine-grained, millisecond-scale…

Current main memory database system architectures are still challenged by high contention workloads and this challenge will continue to grow as the number of cores in processors continues to increase. These systems schedule transactions…

We present a new tool, GPA, that can generate key performance measures for very large systems. Based on solving systems of ordinary differential equations (ODEs), this method of performance analysis is far more scalable than stochastic…

We present a new training methodology for transformers using a multilevel, layer-parallel approach. Through a neural ODE formulation of transformers, our application of a multilevel parallel-in-time algorithm for the forward and…

Recent deep learning models are difficult to train using a large batch size, because commodity machines may not have enough memory to accommodate both the model and a large data batch size. The batch size is one of the hyper-parameters used…

Deep learning is a popular machine learning technique and has been applied to many real-world problems. However, training a deep neural network is very time-consuming, especially on big data. It has become difficult for a single machine to…

Large language models (LLMs) demand significant memory and computation resources. Wafer-scale chips (WSCs) provide high computation power and die-to-die (D2D) bandwidth but face a unique trade-off between on-chip memory and compute…

Scaling training compute, measured in FLOPs, has long been shown to improve the accuracy of large language models, yet training remains resource-intensive. Prior work shows that increasing test-time compute (TTC)-for example through…

Recent advances in large language models (LLMs) have driven impressive progress in omni-modal understanding and generation. However, training omni-modal LLMs remains a significant challenge due to the heterogeneous model architectures…

Prevailing LLM serving engines employ expert parallelism (EP) to implement multi-device inference of massive MoE models. However, the efficiency of expert parallel inference is largely bounded by inter-device communication, as EP embraces…

To reduce the computational and memory overhead of Large Language Models, various approaches have been proposed. These include a) Mixture of Experts (MoEs), where token routing affects compute balance; b) gradual pruning of model…

TAPS is a Topology-Aware intra-operator Parallelism strategy Searching algorithm that generates intra-operator parallelism strategies by considering both intra-node and inter-node bandwidth. Most of the existing auto-parallelism works use…

The recent rapid progress in pre-training Large Language Models has relied on using self-supervised language modeling objectives like next token prediction or span corruption. On the other hand, Machine Translation Systems are mostly…

The rapid growth in the size of large language models has necessitated the partitioning of computational workloads across accelerators such as GPUs, TPUs, and NPUs. However, these parallelization strategies incur substantial data…

Training large models requires a large amount of data, as well as abundant computation resources. While collaborative learning (e.g., federated learning) provides a promising paradigm to harness collective data from many participants,…

In real world industrial applications of topic modeling, the ability to capture gigantic conceptual space by learning an ultra-high dimensional topical representation, i.e., the so-called "big model", is becoming the next desideratum after…

The Mixture-of-Expert (MoE) technique plays a crucial role in expanding the size of DNN model parameters. However, it faces the challenge of extended all-to-all communication latency during the training process. Existing methods attempt to…

Pretraining large language models (LLMs) typically requires centralized clusters with thousands of high-memory GPUs (e.g., H100/A100). Recent decentralized training methods reduce communication overhead by employing federated optimization;…