An L-Attributed Grammar for Adjoint Code

An L-Attributed Grammar for Adjoint Code
复制标题

DOI:
--
复制
发表时间:
2007
期刊:
--
影响因子:
--
通讯作者:
U. Naumann;J. Riehme
U. Naumann;J. Riehme
中科院分区:
其他
文献类型:
--
作者:
U. Naumann;J. Riehme

文献摘要

相似文献

LuFG Informatik 12,Department of Computer Science,RWTH亚琛大学52056亚琛,Germanynaumann@stce. rwth-aachen. deAbstract.用所谓的伴随码可以有效地计算高维函数的导数,并具有机器精度。我们提出了一个L-属性文法的一个简单的命令式语言(C语言的子集)的过程内伴随代码的单遍生成。我们的想法可以很容易地应用到任何适合语法导向翻译的编程语言。此外,概念上的见解在伴随代码的多通道生成的上下文中是有用的。我们的重点是正确性。必要的特定于域的代码优化超出了本文的范围。我们在这方面的相应工作提供参考。
LuFG Informatik 12, Department of Computer Science, RWTH Aachen University52056 Aachen, Germanynaumann@stce.rwth-aachen.deAbstract. Gradients of high-dimensional functions can be computed efficientlyand with machine accuracy by so-called adjoint codes. We present an L-attributedgrammar for the single-pass generation of intraprocedural adjoint code for a sim-ple imperative language (a subset of C). Our ideas can easily be applied to anyprogramming language that is suitable for syntax-directed translation. Moreoverthe conceptual insights are useful in the context of multi-passgeneration of adjointcode. Our focus is on correctness. The necessary domain-specific code optimiza-tions are beyond the scope of this paper. We give references to correspondingwork in this area.