首页> 外文期刊>IEEE transactions on very large scale integration (VLSI) systems >Ravel-XL: a hardware accelerator for assigned-delay compiled-codelogic gate simulation
【24h】

Ravel-XL: a hardware accelerator for assigned-delay compiled-codelogic gate simulation

机译:Ravel-XL:用于分配延迟的编译代码逻辑门仿真的硬件加速器

获取原文
获取原文并翻译 | 示例

摘要

Ravel-XL is a single-board hardware accelerator for gate-levelndigital logic simulation. It uses a standard levelized-code approach tonstatically schedule gate evaluations. However, unlike previousnapproaches based on levelized-code scheduling, it is not limited tonzero- or unit-delay gate models and can provide timing accuracyncomparable to that obtained from event-driven methods. We review thensynchronous waveform algebra that forms the basis of the Ravel-XLnsimulation algorithm, present an architecture for its hardwarenrealization, and describe an implementation of this architecture as ansingle VLSI chip. The chip has about 900000 transistors on a die that isnapproximately 1.4 cm2, requires a 256 pin package and isndesigned to run at 33 MHz. A Ravel-XL board consisting of the processornchip and local instruction and data memory can simulate up to onenbillion gates at a rate of approximately 6.6 million gate evaluationsnper second. To better appreciate the tradeoffs made in designingnRavel-XL, we compare its capabilities to those of other commercial andnresearch software simulators and hardware accelerators
机译:Ravel-XL是用于门级数字逻辑仿真的单板硬件加速器。它使用标准的分级代码方法静态地调度门评估。但是,与基于分层代码调度的先前方法不同,它不受限于零延迟或单位延迟门模型,并且可以提供与从事件驱动方法获得的时序精度相当的时序精度。然后,我们回顾构成Ravel-XLn仿真算法基础的同步波形代数,介绍其硬件实现的架构,并将该架构的实现描述为单个VLSI芯片。该芯片的管芯上约有900000个晶体管,面积约为1.4 cm2,需要256引脚封装,并且设计为以33 MHz运行。由处理器芯片,本地指令和数据存储器组成的Ravel-XL板可以以每秒660万次门评估的速率模拟多达数十亿门。为了更好地了解在设计nRavel-XL中所做出的权衡,我们将其功能与其他商用和非研究软件模拟器和硬件加速器的功能进行了比较。

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