首页> 外国专利> scale process to the construction of the interleaved grinding in a more processor computers.

scale process to the construction of the interleaved grinding in a more processor computers.

机译:在更多处理器计算机中将过程扩展到交错研磨的构造。

摘要

ABSTRACTA compiler for generating code for enabling multipleprocessors to process programs in parallel. The code enablesthe multiple processor system to operate in the followingmanner: one interation of an outer loop in a set of nestedloops is assigned to each processor. If the outer loopcontains more iterations than processor in the system, theprocessors are initially assigned an earlier iteration, and theremaining iterations are assigned to the processor one as theyfinish their earlier iterations. Each processor runs the innerloop iterations serially. In order to enforce dependencies inthe loops, each processor reports its progress in itsiterations of the inner loop to the processor executing thesucceeding outer loop iteration and the waits until theprocessor computing the preceding outer loop is ahead or behindin processing its inner loop iteration by an amount whichguarantees that dependencies will be enforced.
机译:抽象生成用于启用多个代码的代码的编译器处理器并行处理程序。该代码使多处理器系统可在以下操作方式:一组嵌套中外循环的一个交互循环分配给每个处理器。如果是外循环包含的迭代次数多于系统中的处理器,处理器最初被分配了较早的迭代,并且其余的迭代被分配给处理器一个完成他们的早期迭代。每个处理器运行内部循环循环迭代。为了强制依赖在循环中,每个处理器报告其进度内部循环的迭代到处理器执行随后的外循环迭代,并等待直到计算前面的外循环的处理器在前面还是后面在处理其内循环迭代时保证将强制执行依赖项。

著录项

  • 公开/公告号DE68923666T2

    专利类型

  • 公开/公告日1996-03-21

    原文格式PDF

  • 申请/专利权人 DIGITAL EQUIPMENT CORP US;

    申请/专利号DE1989623666T

  • 发明设计人 HARRIS KEVIN W US;NOYCE WILLIAM B US;

    申请日1989-06-22

  • 分类号G06F9/44;

  • 国家 DE

  • 入库时间 2022-08-22 03:41:55

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

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