首页> 外文OA文献 >Multi-GPU maximum entropy image synthesis for radio astronomy
【2h】

Multi-GPU maximum entropy image synthesis for radio astronomy

机译:用于射电天文学的多GpU最大熵图像合成

代理获取
本网站仅为用户提供外文OA文献查询和代理获取服务,本网站没有原文。下单后我们将采用程序或人工为您竭诚获取高质量的原文,但由于OA文献来源多样且变更频繁,仍可能出现获取不到、文献不完整或与标题不符等情况,如果获取不到我们将提供退款服务。请知悉。

摘要

The maximum entropy method (MEM) is a well known deconvolution technique inradio-interferometry. This method solves a non-linear optimization problem withan entropy regularization term. Other heuristics such as CLEAN are faster buthighly user dependent. Nevertheless, MEM has the following advantages: it isunsupervised, it has a statistical basis, it has a better resolution and betterimage quality under certain conditions. This work presents a high performanceGPU version of non-gridding MEM, which is tested using real and simulated data.We propose a single-GPU and a multi-GPU implementation for single andmulti-spectral data, respectively. We also make use of the Peer-to-Peer andUnified Virtual Addressing features of newer GPUs which allows to exploittransparently and efficiently multiple GPUs. Several ALMA data sets are used todemonstrate the effectiveness in imaging and to evaluate GPU performance. Theresults show that a speedup from 1000 to 5000 times faster than a sequentialversion can be achieved, depending on data and image size. This allows toreconstruct the HD142527 CO(6-5) short baseline data set in 2.1 minutes,instead of 2.5 days that takes a sequential version on CPU.
机译:最大熵方法(MEM)是一种众所周知的反卷积技术,即放射干涉法。该方法解决了带有熵正则项的非线性优化问题。其他启发式方法(例如CLEAN)速度更快,但高度依赖用户。但是,MEM具有以下优点:不受监督,具有统计基础,在某些条件下具有更好的分辨率和更好的图像质量。这项工作提出了一种高性能GPU版本的非网格MEM,并使用真实和模拟数据进行了测试。我们分别针对单光谱和多光谱数据提出了单GPU和多GPU的实现。我们还利用了较新GPU的对等和统一虚拟寻址功能,该功能允许透明高效地利用多个GPU。几个ALMA数据集用于演示成像效果并评估GPU性能。结果表明,取决于数据和图像大小,可以实现比顺序版本快1000到5000倍的加速。这样就可以在2.1分钟内重建HD142527 CO(6-5)短基线数据集,而无需花2.5天的时间在CPU上使用顺序版本。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
代理获取

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号