Optimizing layout of an application on a massively parallel supercomputer
    11.
    发明授权
    Optimizing layout of an application on a massively parallel supercomputer 失效
    在大型并行超级计算机上优化应用程序的布局

    公开(公告)号:US08117288B2

    公开(公告)日:2012-02-14

    申请号:US10963101

    申请日:2004-10-12

    IPC分类号: G06F15/177

    CPC分类号: G06F9/5066

    摘要: A general computer-implement method and apparatus to optimize problem layout on a massively parallel supercomputer is described. The method takes as input the communication matrix of an arbitrary problem in the form of an array whose entries C(i, j) are the amount to data communicated from domain i to domain j. Given C(i, j), first implement a heuristic map is implemented which attempts sequentially to map a domain and its communications neighbors either to the same supercomputer node or to near-neighbor nodes on the supercomputer torus while keeping the number of domains mapped to a supercomputer node constant (as much as possible). Next a Markov Chain of maps is generated from the initial map using Monte Carlo simulation with Free Energy (cost function) F=Σi,jC(i,j)H(i,j)− where H(i,j) is the smallest number of hops on the supercomputer torus between domain i and domain j. On the cases tested, found was that the method produces good mappings and has the potential to be used as a general layout optimization tool for parallel codes. At the moment, the serial code implemented to test the method is un-optimized so that computation time to find the optimum map can be several hours on a typical PC. For production implementation, good parallel code for our algorithm would be required which could itself be implemented on supercomputer.

    摘要翻译: 描述了在大型并行超级计算机上优化问题布局的通用计算机实现方法和装置。 该方法采用数组形式的任意问题的通信矩阵作为输入,其条目C(i,j)是从域i到域j传送的数据量。 给定C(i,j),首先实现启发式映射,其尝试顺序地将域及其通信邻居映射到超级计算机节点或超级计算机环面上的近邻节点,同时保持域的数量映射到 超级计算机节点常数(尽可能多)。 接下来,使用具有自由能的蒙特卡罗模拟(成本函数)F =&Sgr; i,jC(i,j)H(i,j),从初始映射生成马尔科夫链映射。其中H(i,j) 域i和域j之间的超级计算机环面上的最小跳数。 在测试的情况下,发现该方法产生良好的映射,并且有可能被用作并行代码的通用布局优化工具。 此时,实现测试方法的序列号未优化,以便在典型的PC上找到最佳映射的计算时间可以为几个小时。 对于生产实现,将需要我们的算法的良好的并行代码,这本身可以在超级计算机上实现。

    Plasma constituent analysis by interferometric techniques
    13.
    发明授权
    Plasma constituent analysis by interferometric techniques 失效
    通过干涉技术进行等离子体成分分析

    公开(公告)号:US5225888A

    公开(公告)日:1993-07-06

    申请号:US633811

    申请日:1990-12-26

    CPC分类号: G01J3/26 G01N21/62

    摘要: An interferometer (18 or 40) is used to identify trace constituents in a plasma during processing semiconductor devices such as transistors. Light emissions collected from the processing chamber (10) are collimated by lens (14) and transmitted to the interferometer (18 or 40) which selectively allows therethrough particular wavelengths of light which are characteristic of the excitation emissions of certain atoms such as sodium and copper. The light intensity at the selected wavelengths is sensed by a photomultiplier tube (30). In one embodiment, the interferometer (18) is a Fabry-Perot type interferometer and the separation of the plates (20 and 22) which form the Fabry-Perot etalon is controlled using a piezoelectric driver (26). A signal processor (34) correlates the sensed light emissions from the photomultiplier tube (30) with the selected wavelength that is determined by the piezoelectric driver (26). In another embodiment, the interferometer (40) is a narrow bandpass interferometric filter which is tiltable with respect to the collimated incident light from the processing chamber (10). Tilting a narrow bandpass interferometric filter (42) with respect to incident light changes the path length through the filter (42) and allows for the selective transmission of certain wavelengths of light. By rapidly tilting the narrow bandpass interferometric filter (42) at a rate between 5-300 Hz with respect to the incident light, a narrow range of wavelengths on the order of 3 nm can be scanned.

    摘要翻译: 在处理诸如晶体管的半导体器件期间,使用干涉仪(18或40)来识别等离子体中的微量成分。 从处理室(10)收集的光发射被透镜(14)准直并传输到干涉仪(18或40),干涉仪选择性地允许特定波长的光,这些特定波长的光是某些原子的激发发射的特征,例如钠和铜 。 所选波长处的光强度由光电倍增管(30)感测。 在一个实施例中,干涉仪(18)是法布里 - 珀罗型干涉仪,并且使用压电驱动器(26)来控制形成法布里 - 珀罗标准具的板(20和22)的分离。 信号处理器(34)使来自光电倍增管(30)的感测光发射与由压电驱动器(26)确定的所选择的波长相关联。 在另一个实施例中,干涉仪(40)是相对于来自处理室(10)的准直入射光可倾斜的窄带通干涉滤光器。 相对于入射光倾斜窄带通干涉滤光器(42)改变通过滤光器(42)的路径长度,并且允许选择性地传输某些波长的光。 通过以相对于入射光的5-300Hz之间的速率快速倾斜窄带通干涉滤光器(42),可以扫描约3nm量级的窄范围的波长。