Searcharxiv⌕ Search

arXiv subjects

Eric Ding

Publications and source records attributed to Eric Ding.

8 recordsLinked to original sources
arXiv2026

CCL-Bench 1.0: A Trace-Based Benchmark for LLM Infrastructure

Evaluative claims about LLM infrastructure -- ``workload X is fastest on hardware Y with software Z'' -- depend on a complex configuration space spanning hardware accelerators, interconnect bandwidth, software frameworks, parallelism plans, and communication libraries. Current infrastructure evaluation benchmarks publish a small set of end-to-end numbers that do not explain why one configuration outperforms another. We present CCL-Bench, a trace-based benchmark that addresses the limitations of existing benchmarks by recording reusable evidence for every ML workload. Each contributed data point in CCL-Bench packages an execution trace, a YAML workload card, and the launch scripts. We have developed a community-extensible toolkit to compute fine-grained compute, memory, and communication efficiency metrics from this evidence. Using CCL-Bench, we surface three claims that summary-statistic benchmarks cannot support: (i) higher compute-communication overlap can coincide with longer training step time and reveal inefficient parallelization choices, (ii) doubling TPU interconnect bandwidth yields a much higher end-to-end improvement in step time than doubling GPU interconnect bandwidth on small and medium workloads, and (iii) the best-tuned configuration on one training framework can run up to 3$\times$ slower than the best-tuned configuration on a peer framework on identical hardware.

cs.DC↗
arXiv2026

Opus: Photonic Rail-Optimized Fabric in ML Datacenters

Rail-optimized network fabrics have become the de facto datacenter scale-out fabric for large-scale ML training. However, the use of high-radix electrical switches to provide all-to-all connectivity in rails imposes substantial power and cost. We propose a rethinking of the rail abstraction by retaining its communication semantics, but realizing it using optical circuit switches. The key challenge is that optical switches support one-to-one connectivity at a time, limiting the fan-out of traffic in ML workloads using hybrid parallelisms. We overcome this through \emph{parallelism-driven rail reconfiguration}, which exploits the non-overlapping communication phases of different parallelism dimensions. This time-multiplexes a single set of physical ports across circuit configurations tailored to each phase within a training iteration. We design and implement Opus, a control plane that orchestrates this in-job reconfiguration of photonic rails at parallelism phase boundaries, and evaluate it on a physical OCS testbed, the Perlmutter supercomputer, and in simulation at up to 2,048 GPUs. Our results show that photonic rails can achieve over $23\times$ network power reduction and $4\times$ cost savings while incurring only modest training overhead at production-relevant OCS reconfiguration latencies.

cs.NI↗
arXiv2025

Propius: A Platform for Collaborative Machine Learning across the Edge and the Cloud

Collaborative Machine Learning is a paradigm in the field of distributed machine learning, designed to address the challenges of data privacy, communication overhead, and model heterogeneity. There have been significant advancements in optimization and communication algorithm design and ML hardware that enables fair, efficient and secure collaborative ML training. However, less emphasis is put on collaborative ML infrastructure development. Developers and researchers often build server-client systems for a specific collaborative ML use case, which is not scalable and reusable. As the scale of collaborative ML grows, the need for a scalable, efficient, and ideally multi-tenant resource management system becomes more pressing. We propose a novel system, Propius, that can adapt to the heterogeneity of client machines, and efficiently manage and control the computation flow between ML jobs and edge resources in a scalable fashion. Propius is comprised of a control plane and a data plane. The control plane enables efficient resource sharing among multiple collaborative ML jobs and supports various resource sharing policies, while the data plane improves the scalability of collaborative ML model sharing and result collection. Evaluations show that Propius outperforms existing resource management techniques and frameworks in terms of resource utilization (up to $1.88\times$), throughput (up to $2.76$), and job completion time (up to $1.26\times$).

cs.DC↗
arXiv2025

Morphlux: Transforming Torus Fabrics for Efficient Multi-tenant ML

We develop Morphlux, a server-scale programmable photonic fabric to interconnect accelerators within servers. We show that augmenting state-of-the-art torus-based ML data-centers with Morphlux can improve the bandwidth of tenant compute allocations by up to 66%, reduce compute fragmentation by up to 70%, and minimize the blast radius of chip failures. We develop a novel end-to-end hardware prototype of Morphlux to demonstrate these performance benefits which translate to 1.72X improvement in training throughput of ML models. By rapidly programming the server-scale fabric in our hardware testbed, Morphlux can replace a failed accelerator chip with a healthy one in 1.2 seconds.

cs.NI↗
arXiv2025

Photonic Rails in ML Datacenters

Rail-optimized network fabrics have become the de facto datacenter scale-out fabric for large-scale ML training. However, the use of high-radix electrical switches to provide all-to-all connectivity in rails imposes massive power, cost, and complexity overheads. We propose a rethinking of the rail abstraction by retaining its communication semantics, but realizing it using optical circuit switches. The key challenge is that optical switches support only one-to-one connectivity at a time, limiting the fan-out of traffic in ML workloads using hybrid parallelisms. We introduce parallelism-driven rail reconfiguration as a solution that leverages the sequential ordering between traffic from different parallelisms. We design a control plane, Opus, to enable time-multiplexed emulation of electrical rail switches using optical switches. More broadly, our work discusses a new research agenda: datacenter fabrics that co-evolve with the model parallelism dimensions within each job, as opposed to the prevailing mindset of reconfiguring networks before a job begins.

cs.NI↗
arXiv2025

LUMION: Fast Fault Recovery for ML Jobs Using Programmable Optical Fabrics

When accelerators fail in modern ML datacenters, operators migrate the affected ML training or inference jobs to entirely new racks. This approach, while preserving network performance, is highly inefficient, requiring datacenters to reserve full racks of idle accelerators for fault tolerance. In this paper, we address this resource inefficiency by introducing LUMION, a novel reconfigurable optical fabric for connecting accelerators within a datacenter rack. Instead of migrating entire ML jobs, LUMION dynamically integrates spare accelerators into ongoing workloads as failures occur, thereby maintaining consistent performance without costly migrations. We show the benefits of LUMION by building an end-to-end hardware prototype. Our experiments fine-tune Llama 3.2 and show that LUMION swaps a failed GPU with a healthy one and restarts the ML job within ~ 1 second of the failure. LUMION achieves higher inter-GPU bandwidth compared to traditional electrical racks after replacing failed accelerators with spare ones, leading to nearly 2X improvement in fine-tuning throughput.

cs.LG↗
arXiv2025

Efficient AllReduce with Stragglers

Distributed machine learning workloads use data and tensor parallelism for training and inference, both of which rely on the AllReduce collective to synchronize gradients or activations. However, AllReduce algorithms are delayed by the slowest GPU to reach the synchronization barrier before the collective (i.e., the straggler). To address this challenge, we propose StragglAR: a parallel algorithm for AllReduce that accelerates distributed training and inference by exploiting natural variation in GPU execution times. StragglAR implements a ReduceScatter among the remaining GPUs during the straggler-induced delay, and then executes a novel collective algorithm to complete the AllReduce once the final GPU reaches the synchronization barrier. StragglAR achieves a 2x theoretical speedup over popular bandwidth-efficient algorithms for large GPU clusters, surpassing the lower bound for bandwidth-optimal synchronous AllReduce by leveraging the asymmetry in when GPUs reach the synchronization barrier. On an 8-GPU server, StragglAR provides a 25% speedup over state-of-the-art AllReduce algorithms.

cs.LG↗