钱恺宸, 尹琛, 景乃锋. 一种基于数据流解耦的可重构阵列映射方法[J]. 微电子学与计算机, 2021, 38(8): 28-33.
引用本文: 钱恺宸, 尹琛, 景乃锋. 一种基于数据流解耦的可重构阵列映射方法[J]. 微电子学与计算机, 2021, 38(8): 28-33.
QIAN Kaichen, YIN Chen, JING Naifeng. A mapping method for reconfigurable array based on data flow decoupling[J]. Microelectronics & Computer, 2021, 38(8): 28-33.
Citation: QIAN Kaichen, YIN Chen, JING Naifeng. A mapping method for reconfigurable array based on data flow decoupling[J]. Microelectronics & Computer, 2021, 38(8): 28-33.

一种基于数据流解耦的可重构阵列映射方法

A mapping method for reconfigurable array based on data flow decoupling

  • 摘要: 可重构阵列依靠数据流驱动带来的能效优势被作为加速器广泛运用在特定领域之中.随着应用范围的增大,当应用中存在不同执行速率的区域时,采用传统的空间映射方案将整个数据流图进行直接映射会造成严重的性能损失.提出了一种基于数据流解耦的映射方法,通过在执行速率不同的区域之间加入解耦单元以解决执行速率失配的问题.同时利用一种分布式多阶段的布局算法,将解耦后的数据流图映射在分簇式的互连结构中.与传统映射方案相比,对数据流进行解耦后的映射方案在保持较高的互连布通率的基础上可以平均提升57.68%的执行性能,同时降低32%~43%的互连开销.

     

    Abstract: Reconfigurable arrays are widely used as accelerators in specific fields relying on the energy efficiency advantages brought by the driving of data flow. With the increase of application range, when there are regions with different execution rates in the application, it will cause serious performance loss by adopting the traditional spatial mapping scheme to map the whole data flow graph directly. A mapping method based on data flow decouplingisproposed to solve the mismatch of execution rate by adding decoupling units between regions with different execution rates. At the same time, a distributed multi-stage placement algorithm is used to map the decoupled dataflow graph into a clustered interconnection structure. Compared with the traditional mapping scheme, the mapping scheme after decoupling the data flow can improve the execution performance by 57.68% on average while maintaining a high interconnection routabilityand reduce interconnection costs by 32% to 43%.

     

/

返回文章
返回