arXiv subjects

David Sidler

Publications and source records attributed to David Sidler.

3 recordsLinked to original sources

arXiv2026

ASTRA-sim 3.0: Next-Level Distributed Machine Learning Simulations via High-Fidelity GPU and Infrastructure Modeling

Distributed machine learning (ML) is a key paradigm for today's large-scale artificial intelligence applications. As model inference arises as an important use case, faithful modeling of latency-sensitive collective communication has never been more important. Capturing the device architecture and modeling control and data paths at high fidelity is therefore a necessity today. Having a common, detailed representation for distributed ML infrastructure is also crucial. We revisit the promising open-source, community-driven simulator: ASTRA-sim. In this work, we identify limitations of the current ASTRA-sim simulator and augment it with new features. To this end, we enable fine-grained, high-fidelity simulation with a standardized infrastructure representation, opening new design space exploration opportunities. We propose the simulation at cache-line-sized load-store granularity, with a detailed graphics processing unit (GPU) execution model, to balance simulation scalability and fidelity. We also introduce InfraGraph, a standardized representation to capture distributed ML network infrastructure in detail. Using the updated ASTRA-sim 3.0 simulator, we showcase interesting design space explorations for designing optimized collective algorithms, network requirements, and GPU architectures.

cs.DC↗

arXiv2020

HyperLogLog Sketch Acceleration on FPGA

Data sketches are a set of widely used approximated data summarizing techniques. Their fundamental property is sub-linear memory complexity on the input cardinality, an important aspect when processing streams or data sets with a vast base domain (URLs, IP addresses, user IDs, etc.). Among the many data sketches available, HyperLogLog has become the reference for cardinality counting (how many distinct data items there are in a data set). Although it does not count every data item (to reduce memory consumption), it provides probabilistic guarantees on the result, and it is, thus, often used to analyze data streams. In this paper, we explore how to implement HyperLogLog on an FPGA to benefit from the parallelism available and the ability to process data streams coming from high-speed networks. Our multi-pipelined high-cardinality HyperLogLog implementation delivers 1.8x higher throughput than an optimized HyperLogLog running on a dual-socket Intel Xeon E5-2630 v3 system with a total of 16 cores and 32 hyper-threads.

cs.DC↗

arXiv2018

Feynman Diagram Description of 2D-Raman-THz Spectroscopy Applied to Water

2D-Raman-THz spectroscopy of liquid water, which has been presented recently (Proc. Natl. Acad. Sci. USA 110, 20402 (2013)), directly probes the intermolecular degrees of freedom of the hydrogen-bond network. However, being a relatively new technique, its information content is not fully explored as to date. While the spectroscopic signal can be simulated based on molecular dynamics simulation in connection with a water force field, it is difficult to relate spectroscopic signatures to the underlying microscopic features of the force field. Here, a completely different approach is taken that starts from an as simple as possible model, i.e., a single vibrational mode with electrical and mechanical anharmonicity augmented with homogeneous and inhomogeneous broadening. An intuitive Feynman diagram picture is developed for all possible pulse sequences of hybrid 2D-Raman-THz spectroscopy. It is shown that the model can explain the experimental data essentially quantitatively with a very small set of parameters, and it is tentatively concluded that the experimental signal originates from the hydrogen-bond stretching vibration around 170 cm$^{-1}$. Furthermore, the echo observed in the experimental data can be quantified by fitting the model. A dominant fraction of its linewidth is attributed to quasi-inhomogeneous broadening in the slow-modulation limit with a correlation time of 370 fs, reflecting the lifetime of the hydrogen-bond networks giving rise the absorption band.

physics.chem-ph↗