An official website of the United States government
High-resolution simulations can deliver great visual quality, but they are often limited by available memory, especially on GPUs. We present a compiler for physical simulation that can achieve both high performance and significantly reduced memory costs, by enabling flexible and aggressivequantization.Low-precision (quantized) numerical data types are used and packed to represent simulation states, leading to reduced memory space and bandwidth consumption. Quantized simulation allows higher resolution simulation with less memory, which is especially attractive on GPUs. Implementing a quantized simulator that has high performance and packs the data tightly for aggressive storage reduction would be extremely labor-intensive and error-prone using a traditional programming language. To make the creation of quantized simulation practical, we have developed a new set of language abstractions and a compilation system. A suite of tailored domain-specific optimizations ensure quantized simulators often run as fast as the full-precision simulators, despite the overhead of encoding-decoding the packed quantized data types. Our programming language and compiler, based onTaichi, allow developers to effortlessly switch between different full-precision and quantized simulators, to explore the full design space of quantization schemes, and ultimately to achieve a good balance between space and precision. The creation of quantized simulation with our system has large benefits in terms of memory consumption and performance, on a variety of hardware, from mobile devices to workstations with high-end GPUs. We can simulate with levels of resolution that were previously only achievable on systems with much more memory, such as multiple GPUs. For example, on asingleGPU, we can simulate a Game of Life with 20 billion cells (8× compression per pixel), an Eulerian fluid system with 421 million active voxels (1.6× compression per voxel), and a hybrid Eulerian-Lagrangian elastic object simulation with 235 million particles (1.7× compression per particle). At the same time, quantized simulations create physically plausible results. Our quantization techniques arecomplementaryto existing acceleration approaches of physical simulation: they can be used in combination with these existing approaches, such as sparse data structures, for even higher scalability and performance.
more »
« less
Learning to simulate high energy particle collisions from unlabeled data
Abstract In many scientific fields which rely on statistical inference, simulations are often used to map from theoretical models to experimental data, allowing scientists to test model predictions against experimental results. Experimental data is often reconstructed from indirect measurements causing the aggregate transformation from theoretical models to experimental data to be poorly-described analytically. Instead, numerical simulations are used at great computational cost. We introduce Optimal-Transport-based Unfolding and Simulation (OTUS), a fast simulator based on unsupervised machine-learning that is capable of predicting experimental data from theoretical models. Without the aid of current simulation information, OTUS trains a probabilistic autoencoder to transform directly between theoretical models and experimental data. Identifying the probabilistic autoencoder’s latent space with the space of theoretical models causes the decoder network to become a fast, predictive simulator with the potential to replace current, computationally-costly simulators. Here, we provide proof-of-principle results on two particle physics examples, Z -boson and top-quark decays, but stress that OTUS can be widely applied to other fields.
more »
« less
- PAR ID:
- 10329934
- Date Published:
- Journal Name:
- Scientific Reports
- Volume:
- 12
- Issue:
- 1
- ISSN:
- 2045-2322
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
ABSTRACT The analysis of particles bound to surfaces by tethers can facilitate understanding of biophysical phenomena (e.g., DNA–protein or protein–ligand interactions and DNA extensibility). Modeling such systems theoretically aids in understanding experimentally observed motions, and the limitations of such models can provide insight into modeling complex systems. The simulation of tethered particle motion (TPM) allows for analysis of complex behaviors exhibited by such systems; however, this type of experiment is rarely taught in undergraduate science classes. We have developed a MATLAB simulation package intended to be used in academic contexts to concisely model and graphically represent the behavior of different tether–particle systems. We show how analysis of the simulation results can be used in biophysical research using single-molecule force spectroscopy (SMFS). Students in physics, engineering, and chemistry will be able to make connections with principles embedded in the field of study and understand how those principles can be used to create meaningful conclusions in a multidisciplinary context. The simulation package can model any given tether–particle system and allows the user to generate a parameter space with static and dynamic model components. Our simulation was successfully able to recreate generally observed experimental trends by using acoustic force spectroscopy (AFS). Further, the simulation was validated through consideration of the conservation of energy of the tether–bead system, trend analyses, and comparison of particle positional data from actual TPM in silico experiments conducted to simulate data with a parameter space similar to the AFS experimental setup. Overall, our TPM simulator and graphical user interface is primarily for demonstrating behaviors characteristic to TPM in a classroom setting but can serve as a template for researchers to set up TPM simulations to mimic a specific SMFS experimental setup.more » « less
-
The design of fusion devices is typically based on computationally expensive simulations. This can be alleviated using high aspect ratio models that employ a reduced number of free parameters, especially in the case of stellarator optimization where non-axisymmetric magnetic fields with a large parameter space are optimized to satisfy certain performance criteria. However, optimization is still required to find configurations with properties such as low elongation, high rotational transform, finite beta and good fast particle confinement. In this work, we train a machine learning model to construct configurations with favourable confinement properties by finding a solution to the inverse design problem, that is, obtaining a set of model input parameters for given desired properties. Since the solution of the inverse problem is non-unique, a probabilistic approach, based on mixture density networks, is used. It is shown that optimized configurations can be generated reliably using this method.more » « less
-
This work validates lumped-parameter models and cable-based models for nets against data from a parabolic flight experiment. The capabilities of a simulator based in Vortex Studio, a multibody dynamics simulation framework, are expanded by introducing i) a lumped-parameter model of the net with lumped masses placed along the threads and ii) a flexible-cable-based model, both of which enable collision detection with thin bodies. An experimental scenario is recreated in simulation, and the deployment and capture phases are analyzed. Good agreement with experiments is observed in both phases, although with differences primarily due to imperfect knowledge of experimental initial conditions. It is demonstrated that both a lumped-parameter model with inner nodes and a cable-based model can enable the detection of collisions between the net and thin geometries of the target. While both models improve notably capture realism compared to a lumped parameter model with no inner nodes, the cable-based model is found to be most computationally efficient. The effect of modeling thread-to-thread collisions (i.e., collisions among parts of the net) is analyzed and determined to be negligible during deployment and initial target wrapping. The results of this work validate the models and increase the confidence in the practicality of this simulator as a tool for research on net-based capture of debris. A cable-based model is validated for the first time in the literature.more » « less
-
Self-supervised learning (SSL) is at the core of training modern large machine learning models, providing a scheme for learning powerful representations that can be used in a variety of downstream tasks. However, SSL strategies must be adapted to the type of training data and downstream tasks required. We propose resimulation-based self-supervised representation learning (RS3L), a novel simulation-based SSL strategy that employs a method of to drive data augmentation for contrastive learning in the physical sciences, particularly, in fields that rely on stochastic simulators. By intervening in the middle of the simulation process and rerunning simulation components downstream of the intervention, we generate multiple realizations of an event, thus producing a set of augmentations covering all physics-driven variations available in the simulator. Using experiments from high-energy physics, we explore how this strategy may enable the development of a foundation model; we show how RS3L pretraining enables powerful performance in downstream tasks such as discrimination of a variety of objects and uncertainty mitigation. In addition to our results, we make the RS3L dataset publicly available for further studies on how to improve SSL strategies. Published by the American Physical Society2025more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1038/s41598-022-10966-7
