Zobrazeno 1 - 10
of 986
pro vyhledávání: '"Hwu, Wen"'
Autor:
Wu, Kun, Park, Jeongmin Brian, Zhang, Xiaofan, Hidayetoğlu, Mert, Mailthody, Vikram Sharma, Huang, Sitao, Lumetta, Steven Sam, Hwu, Wen-mei
The growth rate of the GPU memory capacity has not been able to keep up with that of the size of large language models (LLMs), hindering the model training process. In particular, activations -- the intermediate tensors produced during forward propag
Externí odkaz:
http://arxiv.org/abs/2408.10013
Autor:
Hidayetoglu, Mert, de Gonzalo, Simon Garcia, Slaughter, Elliott, Surana, Pinku, Hwu, Wen-mei, Gropp, William, Aiken, Alex
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 hierarchies, op
Externí odkaz:
http://arxiv.org/abs/2408.05962
Autor:
Park, Jeongmin Brian, Wu, Kun, Mailthody, Vikram Sharma, Quresh, Zaid, Mahlke, Scott, Hwu, Wen-mei
Graph Neural Networks (GNNs) are widely used today in recommendation systems, fraud detection, and node/link classification tasks. Real world GNNs continue to scale in size and require a large memory footprint for storing graphs and embeddings that o
Externí odkaz:
http://arxiv.org/abs/2407.15264
Neighborhood attention reduces the cost of self attention by restricting each token's attention span to its nearest neighbors. This restriction, parameterized by a window size and dilation factor, draws a spectrum of possible attention patterns betwe
Externí odkaz:
http://arxiv.org/abs/2403.04690
Autor:
Reidys, Benjamin, Xue, Yuqi, Li, Daixuan, Sukhwani, Bharat, Hwu, Wen-mei, Chen, Deming, Asaad, Sameh, Huang, Jian
Software-defined networking (SDN) and software-defined flash (SDF) have been serving as the backbone of modern data centers. They are managed separately to handle I/O requests. At first glance, this is a reasonable design by following the rack-scale
Externí odkaz:
http://arxiv.org/abs/2309.06513
Autor:
Park, Jeongmin, Qureshi, Zaid, Mailthody, Vikram, Gacek, Andrew, Shao, Shunfan, AlMasri, Mohammad, Gelado, Isaac, Xiong, Jinjun, Newburn, Chris, Chung, I-hsin, Garland, Michael, Sakharnykh, Nikolay, Hwu, Wen-mei
Data compression and decompression have become vital components of big-data applications to manage the exponential growth in the amount of data collected and stored. Furthermore, big-data applications have increasingly adopted GPUs due to their high
Externí odkaz:
http://arxiv.org/abs/2307.03760
Graph Neural Networks (GNNs) are emerging as a powerful tool for learning from graph-structured data and performing sophisticated inference tasks in various application domains. Although GNNs have been shown to be effective on modest-sized graphs, tr
Externí odkaz:
http://arxiv.org/abs/2306.16384
Autor:
Khatua, Arpandeep, Mailthody, Vikram Sharma, Taleka, Bhagyashree, Ma, Tengfei, Song, Xiang, Hwu, Wen-mei
Publikováno v:
KDD 2023
Graph neural networks (GNNs) have shown high potential for a variety of real-world, challenging applications, but one of the major obstacles in GNN research is the lack of large-scale flexible datasets. Most existing public datasets for GNNs are rela
Externí odkaz:
http://arxiv.org/abs/2302.13522
Relational graph neural networks (RGNNs) are graph neural networks with dedicated structures for modeling the different types of nodes and edges in heterogeneous graphs. While RGNNs have been increasingly adopted in many real-world applications due t
Externí odkaz:
http://arxiv.org/abs/2301.06284