An official website of the United States government
Abstract Next-generation gravitational wave detectors such as Cosmic Explorer, the Einstein Telescope, and LISA, demand highly accurate and extensive gravitational wave (GW) catalogs to faithfully extract physical parameters from observed signals. However, numerical relativity (NR) faces significant challenges in generating these catalogs at the required scale and accuracy on modern computers, as NR codes do not fully exploit modern GPU capabilities. In response, we extend NRPy, a Python-based NR code-generation framework, to develop NRPyEllipticGPU—a CUDA-optimized elliptic solver tailored for the binary black hole (BBH) initial data problem. NRPyEllipticGPU is the first GPU-enabled elliptic solver in the NR community, supporting a variety of coordinate systems and demonstrating substantial performance improvements on both consumer-grade and HPC-grade GPUs. We show that, when compared to a high-end CPU, NRPyEllipticGPU achieves on a high- end GPU up to a sixteenfold speedup in single precision while increasing double- precision performance by a factor of 2–4. This performance boost leverages the GPU’s superior parallelism and memory bandwidth to achieve a compute-bound application and enhancing the overall simulation efficiency. As NRPyEllipticGPU shares the core infrastructure common to NR codes, this work serves as a practical guide for developing full, CUDA-optimized NR codes.
more »
« less
A GPU-Accelerated AMR Solver for Gravitational Wave Propagation
Simulations to calculate a single gravitational waveform (GW) can take several weeks. Yet, thousands of such simulations are needed for the detection and interpretation of gravitational waves. Future detectors will require even more accurate waveforms than those currently used. We present here the first large scale, adaptive mesh, multi-GPU numerical relativity (NR) code together with performance analysis and benchmarking. While comparisons are difficult to make, our GPU extension of the Dendro-GR NR code achieves a 6x speedup over existing state-of-the-art codes. We achieve 800 GFlops/s on a single NVIDIA A100 GPU with an overall 2.5x speedup over a two-socket, 128-core AMD EPYC 7763 CPU node with an equivalent CPU implementation. We present detailed performance analyses, parallel scalability results, and accuracy assessments for GWs computed for mass ratios q=1,2,4. We also present strong scalability up to 8 A100s and weak scaling up to 229,376 ×86 cores on the Texas Advanced Computing Center's Frontera system.
more »
« less
- PAR ID:
- 10426126
- Date Published:
- Journal Name:
- SC22: International Conference for High Performance Computing, Networking, Storage and Analysis
- Issue:
- Article No.: 75
- Page Range / eLocation ID:
- 1 to 15
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract Cloud microphysics is one of the most time‐consuming components in a climate model. In this study, we port the cloud microphysics parameterization in the Community Atmosphere Model (CAM), known as Parameterization of Unified Microphysics Across Scales (PUMAS), from CPU to GPU to seek a computational speedup. The directive‐based methods (OpenACC and OpenMP target offload) are determined as the best fit specifically for our development practices, which enable a single version of source code to run either on the CPU or GPU, and yield a better portability and maintainability. Their performance is first examined in a PUMAS stand‐alone kernel and the directive‐based methods can outperform a CPU node as long as there is enough computational burden on the GPU. A consistent behavior is observed when we run PUMAS on the GPU in a practical CAM simulation. A 3.6× speedup of the PUMAS execution time, including data movement between CPU and GPU, is achieved at a coarse horizontal resolution (8 NVIDIA V100 GPUs against 36 Intel Skylake CPU cores). This speedup further increases up to 5.4× at a high resolution (24 NVIDIA V100 GPUs against 108 Intel Skylake CPU cores), which highlights the fact that GPU favors larger problem size. This study demonstrates that using GPU in a CAM simulation can save noticeable computational costs even with a small portion of code being GPU‐enabled. Therefore, we are encouraged to port more parameterizations to GPU to take advantage of its computational benefit.more » « less
-
Geographic information systems deal with spatial data and its analysis. Spatial data contains many attributes with location information. Spatial autocorrelation is a fundamental concept in spatial analysis. It suggests that similar objects tend to cluster in geographic space. Hotspots, an example of autocorrelation, are statistically significant clusters of spatial data. Other autocorrelation measures like Moran’s I are used to quantify spatial dependence. Large scale spatial autocorrelation methods are compute- intensive. Fast methods for hotspots detection and analysis are crucial in recent times of COVID-19 pandemic. Therefore, we have developed parallelization methods on heterogeneous CPU and GPU environments. To the best of our knowledge, this is the first GPU and SIMD-based design and implementation of autocorrelation kernels. Earlier methods in literature introduced cluster-based and MapReduce-based parallelization. We have used Intrinsics to exploit SIMD parallelism on x86 CPU architecture. We have used MPI Graph Topology to minimize inter-process communication. Our benchmarks for CPU/GPU optimizations gain up to 750X relative speedup with a 8 GPU setup when compared to baseline sequential implementation. Compared to the best implementation using OpenMP + R-tree data structure on a single compute node, our accelerated hotspots benchmark gains a 25X speedup. For real world US counties and COVID data evolution calculated over 500 days, we gain up to 110X speedup reducing time from 33 minutes to 0.3 minutes.more » « less
-
null (Ed.)As specialized hardware accelerators such as GPUs become increasingly popular, developers are looking for ways to target these platforms with high-level APIs. One promising approach is kernel libraries such as PyTorch or cuML, which provide interfaces that mirror CPU-only counterparts such as NumPy or Scikit-Learn. Unfortunately, these libraries are hard to develop and to adopt incrementally: they only support a subset of their CPU equivalents, only work with datasets that fit in device memory, and require developers to reason about data placement and transfers manually. To address these shortcomings, we present a new approach called offload annotations (OAs) that enables heterogeneous GPU computing in existing workloads with few or no code modifications. An annotator annotates the types and functions in a CPU library with equivalent kernel library functions and provides an offloading API to specify how the inputs and outputs of the function can be partitioned into chunks that fit in device memory and transferred between devices. A runtime then maps existing CPU functions to equivalent GPU kernels and schedules execution, data transfers and paging. In data science workloads using CPU libraries such as NumPy and Pandas, OAs enable speedups of up to 1200⇥ and a median speedup of 6.3⇥ by transparently offloading functions to a GPU using existing kernel libraries. In many cases, OAs match the performance of handwritten heterogeneous implementations. Finally, OAs can automatically page data in these workloads to scale to datasets larger than GPU memory, which would need to be done manually with most current GPU libraries.more » « less
-
Random Forests (RFs) are a commonly used machine learning method for classification and regression tasks spanning a variety of application domains, including bioinformatics, business analytics, and software optimization. While prior work has focused primarily on improving performance of the training of RFs, many applications, such as malware identification, cancer prediction, and banking fraud detection, require fast RF classification. In this work, we accelerate RF classification on GPU and FPGA. In order to provide efficient support for large datasets, we propose a hierarchical memory layout suitable to the GPU/FPGA memory hierarchy. We design three RF classification code variants based on that layout, and we investigate GPU- and FPGA-specific considerations for these kernels. Our experimental evaluation, performed on an Nvidia Xp GPU and on a Xilinx Alveo U250 FPGA accelerator card using publicly available datasets on the scale of millions of samples and tens of features, covers various aspects. First, we evaluate the performance benefits of our hierarchical data structure over the standard compressed sparse row (CSR) format. Second, we compare our GPU implementation with cuML, a machine learning library targeting Nvidia GPUs. Third, we explore the performance/accuracy tradeoff resulting from the use of different tree depths in the RF. Finally, we perform a comparative performance analysis of our GPU and FPGA implementations. Our evaluation shows that, while reporting the best performance on GPU, our code variants outperform the CSR baseline both on GPU and FPGA. For high accuracy targets, our GPU implementation yields a 5-9 × speedup over CSR, and up to a 2 × speedup over Nvidia’s cuML library.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1109/SC41404.2022.00080
