Related papers: Scaling Data Center TCP to Terabits with Laminar
Reinforcement learning (RL) post-training for Large Language Models (LLMs) is now scaling to large clusters and running for extended durations to enhance model reasoning performance. However, the scalability of existing RL frameworks is…
As distributed machine learning (ML) workloads scale to thousands of GPUs connected by high-speed interconnects, tail latency in collective communication has become a major bottleneck. Existing RDMA transports, such as RoCE, IRN, SRNIC, and…
Load balancers are pervasively used inside today's clouds to scalably distribute network requests across data center servers. Given the extensive use of load balancers and their associated operating costs, several efforts have focused on…
This paper introduces Laminar, a novel serverless framework based on dispel4py, a parallel stream-based dataflow library. Laminar efficiently manages streaming workflows and components through a dedicated registry, offering a seamless…
More and more latency-sensitive services and applications are being deployed into the data center. Performance can be limited by the high latency of the network interconnect. Because the conventional network stack is designed not only for…
Modern large language model workloads put increasing demands on parallel compute capability and on-chip memory capacity, while also stressing fine-grained data movement and synchronization. These trends motivate exploring and designing…
Coordination services are a fundamental building block of modern cloud systems, providing critical functionalities like configuration management and distributed locking. The major challenge is to achieve low latency and high throughput…
Many scientific computations need multi-node parallelism for matching up both space (memory) and time (speed) ever-increasing requirements. The use of GPUs as accelerators introduces yet another level of complexity for the programmer and…
Efficient parallelization of Large Language Models (LLMs) with long sequences is essential but challenging due to their significant computational and memory demands, particularly stemming from communication bottlenecks in attention…
Load balancing, operator instance collocations and horizontal scaling are critical issues in Parallel Stream Processing Engines to achieve low data processing latency, optimized cluster utilization and minimized communication cost…
Ternary content addressable memories (TCAMs) are commonly used to implement IP lookup, but suffer from high power and area costs. Thus TCAM included in modern chips is limited and can support moderately large datasets in data centers and…
Decentralized training of large language models offers the opportunity to pool computational resources across geographically distributed participants, but is often bottlenecked by network communication, particularly under pipeline parallel…
Large model inference is shifting from cloud to edge due to concerns about the privacy of user interaction data. However, edge devices often struggle with limited computing power, memory, and bandwidth, requiring collaboration across…
Secure multi-party computation (MPC) offers a practical foundation for privacy-preserving machine learning at the edge. However, current MPC systems rely heavily on communication and computation-intensive primitives-such as secure…
In exascale-oriented GPU clusters, rigid-topology jobs leave behind a fragmented post-landing ecology in which long-resident workloads and highly transient tasks compete for unstable residual capacity. Existing centralized, hierarchical,…
Wide-area scaling trends require new approaches to Internet Protocol (IP) lookup, enabled by modern networking chips such as Intel Tofino, AMD Pensando, and Nvidia BlueField, which provide substantial ternary content-addressable memory…
Temporal reasoning and planning are essential capabilities for large language models (LLMs), yet most existing benchmarks evaluate them in isolation and under limited forms of complexity. To address this gap, we introduce the Temporal…
As chiplet-based integration and many-core architectures become the norm in high-performance computing, on-chip wireless communication has emerged as a compelling alternative to traditional interconnects. However, scalable Medium Access…
Although modern, AI-centric datacenters heavily rely on SmartNICs, existing devices impose a hard trade-off. Commercial SmartNICs provide high bandwidth and easy software integration, but offer limited support for customization and data…
TMAC is a toolbox written in C++11 that implements algorithms based on a set of modern methods for large-scale optimization. It covers a variety of optimization problems, which can be both smooth and nonsmooth, convex and nonconvex, as well…