EGPGV: Eurographics Workshop on Parallel Graphics and Visualization
Permanent URI for this community
Browse
Browsing EGPGV: Eurographics Workshop on Parallel Graphics and Visualization by Author "Childs, Hank"
Now showing 1 - 6 of 6
Results Per Page
Sort Options
Item Dynamic I/O Budget Reallocation For In Situ Wavelet Compression(The Eurographics Association, 2019) Marsaglia, Nicole J.; Li, Shaomeng; Belcher, Kristi; Larsen, Matthew; Childs, Hank; Childs, Hank and Frey, SteffenIn situ wavelet compression is a potential solution for enabling post hoc visualization on supercomputers with slow I/O systems. While this in situ compression is typically accomplished by allocating an equal storage budget to each parallel process, we propose an adaptive approach. With our approach, we introduce an assessment step prior to compression, where each process characterizes the variation in its portion of the data, and then dynamically adapts storage budgets to the processes with the most variation. We conducted experiments comparing our adaptive approach with the traditional, non-adaptive approach, on two different simulation codes with concurrencies of 512 cores and mesh resolutions of one billion cells. Our findings show that our adaptive approach yields three orders of magnitude of improvement for one simulation and is not harmful for the other.Item HyLiPoD: Parallel Particle Advection Via a Hybrid of Lifeline Scheduling and Parallelization-Over-Data(The Eurographics Association, 2021) Binyahib, Roba; Pugmire, David; Childs, Hank; Larsen, Matthew and Sadlo, FilipPerformance characteristics of parallel particle advection algorithms can vary greatly based on workload.With this short paper, we build a new algorithm based on results from a previous bake-off study which evaluated the performance of four algorithms on a variety of workloads. Our algorithm, called HyLiPoD, is a ''meta-algorithm,'' i.e., it considers the desired workload to choose from existing algorithms to maximize performance. To demonstrate HyliPoD's benefit, we analyze results from 162 tests including concurrencies of up to 8192 cores, meshes as large as 34 billion cells, and particle counts as large as 300 million. Our findings demonstrate that HyLiPoD's adaptive approach allows it to match the best performance of existing algorithms across diverse workloads.Item An Interpolation Scheme for VDVP Lagrangian Basis Flows(The Eurographics Association, 2019) Sane, Sudhanshu; Childs, Hank; Bujack, Roxana; Childs, Hank and Frey, SteffenUsing the Eulerian paradigm, accurate flow visualization of 3D time-varying data requires a high temporal resolution resulting in large storage requirements. The Lagrangian paradigm has proven to be a viable in situ-based approach to tackle this large data visualization problem. However, previous methods constrained the generation of Lagrangian basis flows to the special case of fixed duration and fixed placement (FDFP), in part because reconstructing the flow field using these basis flows is trivial. Our research relaxes this constraint, by considering the general case of variable duration and variable placement (VDVP) with the goal of increasing the amount of information per byte stored. That said, reconstructing the flow field using VDVP basis flows is non-trivial; the primary contribution of our work is a method we call VDVP-Interpolation which solves this problem. VDVP-Interpolation reduces error propagation and limits interpolation error while using VDVP Lagrangian basis flows. As a secondary contribution of the work, we generate VDVP basis flows for multiple data sets and demonstrate improved accuracy-storage propositions compared to previous work. In some cases, we demonstrate up to 40-60% more accurate pathline calculation while using 50% less data storage.Item Machine Learning-Based Autotuning for Parallel Particle Advection(The Eurographics Association, 2021) Schwartz, Samuel D.; Childs, Hank; Pugmire, David; Larsen, Matthew and Sadlo, FilipData-parallel particle advection algorithms contain multiple controls that affect their execution characteristics and performance, in particular how often to communicate and how much work to perform between communications. Unfortunately, the optimal settings for these controls vary based on workload, and, further, it is not easy to devise straight-forward heuristics that automate calculation of these settings. To solve this problem, we investigate a machine learning-based autotuning approach for optimizing data-parallel particle advection. During a pre-processing step, we train multiple machine learning techniques using a corpus of performance data that includes results across a variety of workloads and control settings. The best performing of these techniques is then used to form an oracle, i.e., a module that can determine good algorithm control settings for a given workload immediately before execution begins. To evaluate this approach, we assessed the ability of seven machine learning models to capture particle advection performance behavior and then ran experiments for 108 particle advection workloads on 64 GPUs of a supercomputer. Our findings show that our machine learning-based oracle achieves good speedups relative to the available gains.Item PGV 2019: Frontmatter(The Eurographics Association, 2019) Childs, Hank; Frey, Steffen; Childs, Hank and Frey, SteffenItem Scalable In Situ Computation of Lagrangian Representations via Local Flow Maps(The Eurographics Association, 2021) Sane, Sudhanshu; Yenpure, Abhishek; Bujack, Roxana; Larsen, Matthew; Moreland, Kenneth; Garth, Christoph; Johnson, Chris R.; Childs, Hank; Larsen, Matthew and Sadlo, FilipIn situ computation of Lagrangian flow maps to enable post hoc time-varying vector field analysis has recently become an active area of research. However, the current literature is largely limited to theoretical settings and lacks a solution to address scalability of the technique in distributed memory. To improve scalability, we propose and evaluate the benefits and limitations of a simple, yet novel, performance optimization. Our proposed optimization is a communication-free model resulting in local Lagrangian flow maps, requiring no message passing or synchronization between processes, intrinsically improving scalability, and thereby reducing overall execution time and alleviating the encumbrance placed on simulation codes from communication overheads. To evaluate our approach, we computed Lagrangian flow maps for four time-varying simulation vector fields and investigated how execution time and reconstruction accuracy are impacted by the number of GPUs per compute node, the total number of compute nodes, particles per rank, and storage intervals. Our study consisted of experiments computing Lagrangian flow maps with up to 67M particle trajectories over 500 cycles and used as many as 2048 GPUs across 512 compute nodes. In all, our study contributes an evaluation of a communication-free model as well as a scalability study of computing distributed Lagrangian flow maps at scale using in situ infrastructure on a modern supercomputer.