Related papers: FIGARO: Improving System Performance via Fine-Grai…

Graph processing is typically considered to be a memory-bound rather than compute-bound problem. One common line of thought is that more available memory bandwidth corresponds to better graph processing performance. However, in this work we…

GPUs are critical for compute-intensive applications, yet emerging workloads such as recommender systems, graph analytics, and data analytics often exceed GPU memory capacity. Existing solutions allow GPUs to use CPU DRAM or SSDs as…

The configurable building blocks of current FPGAs -- Logic blocks (LBs), Digital Signal Processing (DSP) slices, and Block RAMs (BRAMs) -- make them efficient hardware accelerators for the rapid-changing world of Deep Learning (DL).…

This dissertation develops hardware that automatically reduces the effective latency of accessing memory in both single-core and multi-core systems. To accomplish this, the dissertation shows that all last level cache misses can be…

Graphics rendering applications increasingly leverage neural networks in tasks such as denoising, supersampling, and frame extrapolation to improve image quality while maintaining frame rates. The temporal coherence inherent in these tasks…

A key performance bottleneck when training graph neural network (GNN) models on large, real-world graphs is loading node features onto a GPU. Due to limited GPU memory, expensive data movement is necessary to facilitate the storage of these…

In the age of big data, information security has become a major issue of debate, especially with the rise of the Internet of Things (IoT), where attackers can effortlessly obtain physical access to edge devices. The hash algorithm is the…

State Space Models (SSMs) offer a promising alternative to transformers for long-sequence processing. However, their efficiency remains hindered by memory-bound operations, particularly in the prefill stage. While MARCA, a recent first…

Personalized recommendation is a ubiquitous application on the internet, with many industries and hyperscalers extensively leveraging Deep Learning Recommendation Models (DLRMs) for their personalization needs (like ad serving or movie…

FPGAs are increasingly utilized in data centers due to their capacity to exploit data parallelism in computationally intensive workloads. Furthermore, the processing of modern data center workloads requires moving vast amounts of data,…

Caching is a technique to reduce peak traffic rates by prefetching popular content into memories at the end users. Conventionally, these memories are used to deliver requested content in part from a locally cached copy rather than through…

A simultaneously transmitting and reflecting surface (STARS) enabled edge caching system is proposed for reducing backhaul traffic and ensuring the quality of service. A novel Caching-at-STARS structure, where a dedicated smart controller…

Quad-level cell (QLC) flash offers significant benefits in cost and capacity, but its limited reliability leads to frequent read retries, which severely degrade read performance. A common strategy in high-density flash storage is to program…

DRAM is the prevalent main memory technology, but its long access latency can limit the performance of many workloads. Although prior works provide DRAM designs that reduce DRAM access latency, their reduced storage capacities hinder the…

Optimizing deep learning algorithms currently requires slow, manual derivation, potentially leaving much performance untapped. Methods like FlashAttention have achieved a x6 performance improvement over native PyTorch by avoiding…

Near-data accelerators (NDAs) that are integrated with main memory have the potential for significant power and performance benefits. Fully realizing these benefits requires the large available memory capacity to be shared between the host…

Structured sparsity enables deploying large language models (LLMs) on resource-constrained systems. Approaches like dense-to-sparse fine-tuning are particularly compelling, achieving remarkable structured sparsity by reducing the model size…

Processing-in-memory (PIM) architectures have seen an increase in popularity recently, as the high internal bandwidth available within 3D-stacked memory provides greater incentive to move some computation into the logic layer of the memory.…

Retrieval-augmented generation (RAG), which combines large language models (LLMs) with retrievals from external knowledge databases, is emerging as a popular approach for reliable LLM serving. However, efficient RAG serving remains an open…

Recommendation system has gained a large popularity for a variety of personalized suggestion tasks, but the ever-increasing number of user data makes real-time processing of recommendation systems difficult. NAND flash memory-based…