Related papers: JSPIM: A Skew-Aware PIM Accelerator for High-Perfo…

Processing-in-memory (PIM) architectures bring computation closer to data, reducing the processor-memory transfer bottleneck in traditional processor-centric designs. Novel hardware solutions, such as UPMEM's in-memory processing…

As large language models (LLMs) continue to advance, retrieval-augmented generation (RAG) has become the key mechanism for expanding model knowledge and reducing hallucinations. Central to RAG is approximate nearest neighbor search (ANNS),…

In order to boost the performance of data-intensive computing on HPC systems, in-memory computing frameworks, such as Apache Spark and Flink, use local DRAM for data storage. Optimizing the memory allocation to data storage is critical to…

Phase-change memory (PCM) devices have multiple banks to serve memory requests in parallel. Unfortunately, if two requests go to the same bank, they have to be served one after another, leading to lower system performance. We observe that a…

The exponential growth of artificial intelligence (AI) applications has exposed the inefficiency of conventional von Neumann architectures, where frequent data transfers between compute units and memory create significant energy and latency…

Decoder-only Transformer models such as GPT have demonstrated exceptional performance in text generation, by autoregressively predicting the next token. However, the efficacy of running GPT on current hardware systems is bounded by low…

The rapid advancement of Large Language Models (LLMs) has revolutionized various aspects of human life, yet their immense computational and energy demands pose significant challenges for efficient inference. The memory wall, the growing…

High-performance computing systems are moving towards 2.5D and 3D memory hierarchies, based on High Bandwidth Memory (HBM) and Hybrid Memory Cube (HMC) to mitigate the main memory bottlenecks. This trend is also creating new opportunities…

Modern online large language model (LLM) services, such as Retrieval-Augmented Generation (RAG) and agent systems, increasingly expose two prominent characteristics: prompt segmentation (e.g., system instructions, retrieved passages, tool…

The widespread adoption of Large Language Models (LLMs) has exponentially increased the demand for efficient serving systems. With growing requests and context lengths, key-value (KV)-related operations, including attention computation and…

Co-locating latency-critical services (LCSs) and best-effort jobs (BEJs) constitute the principal approach for enhancing resource utilization in production. Nevertheless, the co-location practice hurts the performance of LCSs due to…

Large Language Models (LLMs) increasingly require processing long text sequences, but GPU memory limitations force difficult trade-offs between memory capacity and bandwidth. While HBM-based acceleration offers high bandwidth, its capacity…

Large Language Models (LLMs) are increasingly deployed on edge devices with Neural Processing Units (NPUs), yet the decode phase remains memory-intensive, limiting performance. Processing-in-Memory (PIM) offers a promising solution, but…

Large language model (LLM) inference has been a prevalent demand in daily life and industries. The large tensor sizes and computing complexities in LLMs have brought challenges to memory, computing, and databus. This paper proposes a…

Sparse tensors are the most used representation of sparse multidimensional data. Operations that decompose them, selecting their most important features while reducing their dimension, have become prevalent procedures in machine learning.…

Sparse deep learning has reduced computation significantly, but its irregular non-zero data distribution complicates the data flow and hinders data reuse, increasing on-chip SRAM access and thus power consumption of the chip. This paper…

Spiking Neural Networks (SNNs) have emerged as a biologically inspired alternative to conventional deep networks, offering event-driven and energy-efficient computation. However, their throughput remains constrained by the serial update of…

We show in this work that memory intensive computations can result in severe performance problems due to off-chip memory access and CPU-GPU context switch overheads in a wide range of deep learning models. For this problem, current…

Evaluating the relational join is one of the central algorithmic and most well-studied problems in database systems. A staggering number of variants have been considered including Block-Nested loop join, Hash-Join, Grace, Sort-merge for…

The expansion of long-context Large Language Models (LLMs) creates significant memory system challenges. While Processing-in-Memory (PIM) is a promising accelerator, we identify that it suffers from critical inefficiencies when scaled to…