首页> 外文会议>IEEE East-West Design and Test Symposium >Fault Tolerance Properties of Systems Generated with the Use of High-Level Synthesis
【24h】

Fault Tolerance Properties of Systems Generated with the Use of High-Level Synthesis

机译:使用高级综合生成的系统的容错特性

获取原文

摘要

During the last decades, electronic systems became an important matter of controlling many critical processes. However, those critical processes often require increased reliability. This requirement places pressure on system developers to make systems reliable. Be-cause of ever growing chip-level integration, capabilities of electronic systems are expanding, and, thus, leading to more complex system architectures, the number of man-hours needed to develop such systems is significantly increasing. Many people believe the solution is to move the development to a higher level of abstraction (e.g. an algorithm level) and use the so-called High-Level Synthesis (HLS) for this purpose. In this research, we aimed towards a decision, whether the usage of HLS impacts the resulting reliability properties of the system, and, thus, whether the HLS-generated system matches reliability properties of its corresponding VHDL-implemented version. We found out that, for the selected set of circuits, HLS performs better in terms of resource consumption, but, also, which we consider surprising, in terms of reliability. For the selected set, HLS achieved better reliability by 3.03 percentage points in contrast to the classical approach utilizing a traditional Hardware Description Language (HDL). In these experiments, no redundancy was intentionally inserted into benchmarking circuits.
机译:在过去的几十年中,电子系统成为控制许多关键过程的重要问题。但是,这些关键过程通常需要提高可靠性。该要求给系统开发人员带来了使系统可靠的压力。由于芯片级集成度的不断提高,电子系统的功能正在扩展,因此,导致更复杂的系统架构,开发此类系统所需的工时数量显着增加。许多人认为解决方案是将开发移至更高的抽象级别(例如算法级别),并为此目的使用所谓的高级综合(HLS)。在这项研究中,我们的目标是决定是否使用HLS会影响系统的最终可靠性,因此,由HLS生成的系统是否与其相应的VHDL实现版本的可靠性相匹配。我们发现,对于所选的一组电路,HLS在资源消耗方面表现更好,但在可靠性方面我们也感到惊讶。对于选定的集合,与使用传统硬件描述语言(HDL)的经典方法相比,HLS的可靠性提高了3.03个百分点。在这些实验中,没有有意地将冗余插入基准电路。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

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