Multi-Node Multi-GPU Diffeomorphic Image Registration for Large-Scale Imaging Problems.

Multi-Node Multi-GPU Diffeomorphic Image Registration for Large-Scale Imaging Problems.
复制标题

DOI:
10.1109/sc41405.2020.00042
复制
发表时间:
2020-11
期刊:
International Conference for High Performance Computing, Networking, Storage and Analysis : [proceedings]. SC (Conference : Supercomputing)
影响因子:
--
通讯作者:
Mang A
Mang A
中科院分区:
其他
文献类型:
--
作者:
Brunn M;Himthani N;Biros G;Mehl M;Mang A

文献摘要

参考文献

被引文献

相似文献

我们提出了一种用于大变形微分同胚图像配准的高斯-牛顿-克雷洛夫求解器。我们将公开可用的 CLAIRE 库扩展到多节点多图形处理单元 (GPU) 系统,并引入新颖的算法修改,可显着提高性能。我们的贡献包括(i)用于缩减空间高斯-牛顿 Hessian 系统的新预处理器,(ii)高度优化的多节点多 GPU 实现,利用主要计算内核(插值、高阶有限差分算子和快速傅里叶变换)的设备直接通信,以及(iii)与最先进的 CPU 和 GPU 实现的比较。我们在单个 NVIDIA Tesla V100 上仅用 5 秒就解决了 2563 分辨率的图像配准问题,与最先进的技术相比,性能加速了 70%。在最大的一次运行中,我们在 TACC 的 Longhorn 系统上的 64 个节点和 256 个 GPU 上注册了 20483 个分辨率图像(25 B 未知数;大约比最先进的 GPU 实现中解决的最大问题大 152 倍)。
We present a Gauss-Newton-Krylov solver for large deformation diffeomorphic image registration. We extend the publicly available CLAIRE library to multi-node multi-graphics processing unit (GPUs) systems and introduce novel algorithmic modifications that significantly improve performance. Our contributions comprise (i) a new preconditioner for the reduced-space Gauss-Newton Hessian system, (ii) a highly-optimized multi-node multi-GPU implementation exploiting device direct communication for the main computational kernels (interpolation, high-order finite difference operators and Fast-Fourier-Transform), and (iii) a comparison with state-of-the-art CPU and GPU implementations. We solve a 2563-resolution image registration problem in five seconds on a single NVIDIA Tesla V100, with a performance speedup of 70% compared to the state-of-the-art. In our largest run, we register 20483 resolution images (25 B unknowns; approximately 152× larger than the largest problem solved in state-of-the-art GPU implementations) on 64 nodes with 256 GPUs on TACC’s Longhorn system.
DOI: 10.1137/140992564
发表时间: 2016-01-01
影响因子: 3.1
作者:
Alexanderian, Alen;Petra, Noemi;Ghattas, Omar
通讯作者: Ghattas, Omar
DOI: 10.1023/b:visi.0000043755.93987.aa
发表时间: 2005-02-01
影响因子: 19.5
作者:
Beg, MF;Miller, MI;Younes, L
通讯作者: Younes, L
DOI: 10.1038/nature12107
发表时间: 2013-05-16
期刊: Nature
影响因子: 64.8
作者:
通讯作者: --
DOI: 10.1016/j.neuroimage.2010.09.025
发表时间: 2011-02-01
期刊: NeuroImage
影响因子: 5.7
作者:
Avants BB;Tustison NJ;Song G;Cook PA;Klein A;Gee JC
通讯作者: Gee JC
DOI: 10.1142/s0219467808002988
发表时间: 2008-01-01
影响因子: 1.6
作者:
Courty, Nicolas;Hellier, Pierre
通讯作者: Hellier, Pierre