The Effect of Combining Compiler Optimizations on Code Size
The Effect of Combining Compiler Optimizations on Code Size
复制标题
DOI：
10.1109/sccc.2011.25
复制
发表时间：
2011-11
期刊：
影响因子：
--
通讯作者：
J. Foleiss;A. F. Silva;L. B. Ruiz
中科院分区：
文献类型：
--
作者：
J. Foleiss;A. F. Silva;L. B. Ruiz
最近，我们的嵌入式计算研究小组一直致力于以小占用和低能耗为目标的代码生成。具体来说，我们一直在研究优化对代码大小的影响。之前，我们已经得出结论，评估单个优化不会产生有趣的结果。这是因为减少代码大小的优化效率受到所使用的优化技术组合的极大影响。因此，有必要确定哪些优化技术与代码大小减少密切相关。在本文中，我们提出了一些实验评估结果，揭示了优化技术之间的关系，产量小二进制文件。
Lately, our embedded computing research group has been working on code generation targeting small footprint and low energy consumption. In specific, we've been researching the effect of optimizations on code size. Previously, we've concluded that evaluating single optimizations does not lead to interesting results. That's due to the fact that optimization efficiency towards reducing code size is greatly affected by the combinations of optimization techniques being used. Therefore, it is necessary to identify which optimization techniques are heavily related, in respect to code size reduction. In this paper, we present some experimental evaluation results that reveal relationships among optimization techniques that yield to small binaries.