Related papers: Vilamb: Low Overhead Asynchronous Redundancy for D…
Production LLM serving must simultaneously deliver high throughput, low latency, and sufficient context capacity under non-stationary traffic and mixed request requirements. Data parallelism (DP) maximizes throughput by running independent…
Predictable execution time upon accessing shared memories in multi-core real-time systems is a stringent requirement. A plethora of existing works focus on the analysis of Double Data Rate Dynamic Random Access Memories (DDR DRAMs), or…
Parallel processing is considered as todays and future trend for improving performance of computers. Computing devices ranging from small embedded systems to big clusters of computers rely on parallelizing applications to reduce execution…
Diffusion Large Language Models (dLLMs) have emerged as a promising alternative to autoregressive generation by enabling parallel token prediction. However, practical dLLM decoding still suffers from high inference latency, which limits…
The modern semiconductor industry requires memory solutions that can keep pace with the high-speed demands of high-performance computing. Embedded non-volatile memories (eNVMs) address these requirements by offering faster access to stored…
Video diffusion models (VDMs) perform attention computation over the 3D spatio-temporal domain. Compared to large language models (LLMs) processing 1D sequences, their memory consumption scales cubically, necessitating parallel serving…
To improve system performance, modern operating systems (OSes) often undertake activities that require modification of virtual-to-physical page translation mappings. For example, the OS may migrate data between physical frames to defragment…
Modern operating systems are typically POSIX-compliant with major system calls specified decades ago. The next generation of non-volatile memory (NVM) technologies raise concerns about the efficiency of the traditional POSIX-based systems.…
With the surge in cloud storage adoption, enterprises face challenges managing data duplication and exponential data growth. Deduplication mitigates redundancy, yet maintaining redundancy ensures high availability, incurring storage costs.…
Vision-language models (VLMs) could power real-time assistants and autonomous agents, but they face a critical challenge: understanding near-infinite video streams without escalating latency and memory usage. Processing entire videos with…
Deep Learning Recommendation Models (DLRMs) underpin personalized services but face a critical freshness-accuracy tradeoff due to massive parameter synchronization overheads. Production DLRMs deploy decoupled training/inference clusters,…
In-memory computing (IMC) with non-volatile memories (NVMs) has emerged as a promising approach to address the rapidly growing computational demands of Deep Neural Networks (DNNs). Mapping DNN layers spatially onto NVM-based IMC…
Redox Flow Batteries are a promising option for large-scale stationary energy storage. The vanadium redox flow battery is the most widely commercialized system thanks to its chemical stability and performance. This work aims to optimize the…
At the end of Silicon roadmap, keeping the leakage power in tolerable limit and bridging the bandwidth gap between processor and memory have become some of the biggest challenges. Several promising Non-Volatile Memories (NVMs) such as,…
We design and analyze the performance of a redundancy management mechanism for Peer-to-Peer backup applications. Armed with the realization that a backup system has peculiar requirements -- namely, data is read over the network only during…
We can use a hybrid memory system consisting of DRAM and Intel Optane DC Persistent Memory (We call it DCPM in this paper) as DCPM is now commercially available since April 2019. Even if the latency for DCPM is several times higher than…
Demands for implementing original OSs that can achieve high I/O performance on PC/AT compatible hardware have recently been increasing, but conventional OS debugging environments have not been able to simultaneously assure their stability,…
Variable length coding for Non-Volatile Memory (NVM) technologies is a promising method to improve memory capacity and system performance through compressing memory blocks. However, compression techniques used to improve capacity or…
Intermittent computing requires custom programming models to ensure the correct execution of applications despite power failures. However, existing programming models lead to programs that are hardware-dependent and not reusable. This paper…
Artificial Neural Network computation relies on intensive vector-matrix multiplications. Recently, the emerging nonvolatile memory (NVM) crossbar array showed a feasibility of implementing such operations with high energy efficiency, thus…