Related papers: Characterizing Off-path SmartNIC for Accelerating …
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…
Multi-tenancy is essential for unleashing SmartNIC's potential in datacenters. Our systematic analysis in this work shows that existing on-path SmartNICs have resource multiplexing limitations. For example, existing solutions lack…
Multipath communication not only allows improved throughput but can also be used to leverage different path characteristics to best fulfill each application's objective. In particular, certain delay-sensitive applications, such as real time…
Federated learning is a distributed machine learning approach where local weight parameters trained by clients locally are aggregated as global parameters by a server. The global parameters can be trained without uploading privacy-sensitive…
Overheads in Operating System kernel network stacks and sockets have been hindering OSes from managing networking operations efficiently for years. Moreover, when building Remote Procedure Calls over TCP, certain TCP features do not match…
With the growing performance requirements on networked applications, there is a new trend of offloading stateful network applications to SmartNICs to improve performance and reduce the total cost of ownership. However, offloading stateful…
Receive side scaling (RSS) is a network interface card (NIC) technology. It provides the benefits of parallel receive processing in multiprocessing environments. However, existing RSS-enabled NICs lack a critical data steering mechanism…
SmartNICs are touted as an attractive substrate for network application offloading, offering benefits in programmability, host resource saving, and energy efficiency. The current usage restricts offloading to local hosts and confines…
Deep Neural Networks have flourished at an unprecedented pace in recent years. They have achieved outstanding accuracy in fields such as computer vision, natural language processing, medicine or economics. Specifically, Convolutional Neural…
In the Fully Sharded Data Parallel (FSDP) training pipeline, collective operations can be interleaved to maximize the communication/computation overlap. In this scenario, outstanding operations such as Allgather and Reduce-Scatter can…
Why do security cameras, sensors, and siri use cloud servers instead of on-board computation? The lack of very-low-power, high-performance chips greatly limits the ability to field untethered edge devices. We present the NV-1, a new…
Unmanned aerial vehicle (UAV) communication is of crucial importance for diverse practical applications. However, it is susceptible to the severe spectrum scarcity problem and interference since it operates in the unlicensed spectrum band.…
AI training and inference impose sustained, fine-grain I/O that stresses host-mediated, TCP-based storage paths. Motivated by kernel-bypass networking and user-space storage stacks, we revisit POSIX-compatible object storage for GPU-centric…
Our analysis of recent Internet traces shows that up to 71% of flows contain suspicious behaviors indicative of low-volume network attacks such as port scans. However, distinguishing anomalous traffic in real time is challenging as each…
Large-scale graph processing has drawn great attention in recent years. Most of the modern-day datacenter workloads can be represented in the form of Graph Processing such as MapReduce etc. Consequently, a lot of designs for Domain-Specific…
Smart devices, storage and other distributed technologies have the potential to greatly improve the utilisation of network infrastructure and renewable generation. Decentralised control of these technologies overcomes many scalability and…
On-chip communication infrastructure is a central component of modern systems-on-chip (SoCs), and it continues to gain importance as the number of cores, the heterogeneity of components, and the on-chip and off-chip bandwidth continue to…
SmartNICs have been increasingly utilized across various applications to offload specific computational tasks, thereby enhancing overall system performance. However, this offloading process introduces several communication challenges that…
Emerging artificial intelligence (AI) and machine learning (ML) workloads present new challenges of managing the collective communication used in distributed training across hundreds or even thousands of GPUs. This paper presents STrack, a…
The MultiNoC system implements a programmable on-chip multiprocessing platform built on top of an efficient, low area overhead intra-chip interconnection scheme. The employed interconnection structure is a Network on Chip, or NoC. NoCs are…