arXiv subjects

Hongyi Zeng

Publications and source records attributed to Hongyi Zeng.

4 recordsLinked to original sources

arXiv2026

HCCL: Collective Communication for Meta Training and Inference Accelerators

We present HCCL, a collective communication library co-designed with Meta's MTIA 300 accelerator, the first Meta chip to integrate backend networking directly on chip package. MTIA 300 includes dedicated message engines (MEs) with near-memory compute (NMC) that fully offload collective execution from the compute grid, enabling large overlap between computation and communication. HCCL uses a compiled communication model in which the host generates a complete description of each collective including dependencies. We describe the control and data path architecture, topology-aware algorithm selection across MTIA 300's asymmetric scale-up and scale-out network, and optimizations for both training and inference workloads. For training, HCCL achieves up to 940 GB/s on intra-rack collectives while introducing less than 0.5% degradation to concurrent compute throughput. For inference, we leverage one-sided communication primitives that bypass the scheduling path to minimize collective latency and describe collective designs that improve compute-communication pipelining for latency-sensitive workloads.

cs.NI↗

arXiv2025

Collective Communication for 100k+ GPUs

The increasing scale of large language models (LLMs) necessitates highly efficient collective communication frameworks, particularly as training workloads extend to hundreds of thousands of GPUs. Traditional communication methods face significant throughput and latency limitations at this scale, hindering both the development and deployment of state-of-the-art models. This paper presents the NCCLX collective communication framework, developed at Meta, engineered to optimize performance across the full LLM lifecycle, from the synchronous demands of large-scale training to the low-latency requirements of inference. The framework is designed to support complex workloads on clusters exceeding 100,000 GPUs, ensuring reliable, high-throughput, and low-latency data exchange. Empirical evaluation on the Llama4 model demonstrates substantial improvements in communication efficiency. This research contributes a robust solution for enabling the next generation of LLMs to operate at unprecedented scales.

cs.DC↗

arXiv2025

Foveated Instance Segmentation

Instance segmentation is essential for augmented reality and virtual reality (AR/VR) as it enables precise object recognition and interaction, enhancing the integration of virtual and real-world elements for an immersive experience. However, the high computational overhead of segmentation limits its application on resource-constrained AR/VR devices, causing large processing latency and degrading user experience. In contrast to conventional scenarios, AR/VR users typically focus on only a few regions within their field of view before shifting perspective, allowing segmentation to be concentrated on gaze-specific areas. This insight drives the need for efficient segmentation methods that prioritize processing instance of interest, reducing computational load and enhancing real-time performance. In this paper, we present a foveated instance segmentation (FovealSeg) framework that leverages real-time user gaze data to perform instance segmentation exclusively on instance of interest, resulting in substantial computational savings. Evaluation results show that FSNet achieves an IoU of 0.56 on ADE20K and 0.54 on LVIS, notably outperforming the baseline. The code is available at https://github.com/SAI-

cs.CV↗

arXiv2014

Delay-Energy lower bound on Two-Way Relay Wireless Network Coding

Network coding is a novel solution that significantly improve the throughput and energy consumed of wireless networks by mixing traffic flows through algebraic operations. In conventional network coding scheme, a packet has to wait for packets from other sources to be coded before transmitting. The wait-and-code scheme will naturally result in packet loss rate in a finite buffer. We will propose Enhanced Network Coding (ENC), an extension to ONC in continuous time domain. In ENC, the relay transmits both coded and uncoded packets to reduce delay. In exchange, more energy is consumed in transmitting uncoded packets. ENC is a practical algorithm to achieve minimal average delay and zero packet-loss rate under given energy constraint. The system model for ENC on a general renewal process queuing is presented. In particular, we will show that there exists a fundamental trade-off between average delay and energy. We will also present the analytic result of lower bound for this trade-off curve, which can be achieved by ENC.

cs.IT↗