期刊文献+

32位微处理器总线接口部件的设计 被引量:2

On Designing for Chinese Use a BIU (Bus Interface Unit) of a 32-bit RISC
下载PDF
导出
摘要 由于微处理器和存储器两者之间速度的差异性 ,存储系统已经成为提高微处理器性能的一个瓶颈。同时 ,系统总线的开销在整个访存延迟中占有相当大的比重。因而 ,设计一个高效的总线接口对于提高微处理器的性能是非常重要的。文中在 32位微处理器 ARS0 3总线接口部件的设计中 ,使用 Load/ Store缓冲模型和流水、乱序执行的地址、数据总线等方法来提高其效率 ,采用 M/M/ 1 / K排队论模型确定了缓存队列的长度。实际应用程序仿真结果表明 ,总线接口的设计是高效的 ,去掉使用的优化方法会使 ARS0 3的执行时间平均增加 2 1 .6%。 The Aviation Microelectronic Center of NPU(Northwestern Polytechnical University) has recently completed the development of a 32-bit super-scalar RISC microprocessor, which we call ARS03(Advanced RISC System 03), for filling the need of Chinese military aviation. In this paper, we present the design of the BIU of ARS03, which we deem to be successful because it helps ARS03 to meet performance requirements. Fig.1 gives the schematical diagram of the architecture of ARS03. Section 2 explains in detail the design of its BIU. Subsection 2.1 explains the BIU's processing logic. Subsection 2.2 explains the two parts of BIU's buffer separately: output FIFO(First In First Out) in subsubsection 2.2.1 and input FIFO in subsubsection 2.2.2. In connection with output FIFO, we stress that, from the buffer full probabilities calculated by us for different queue lengths and presented in Table 1, we select 8 as the suitable value for queue length K. Subsections 2.3 and 2.4 explain respectively the BIU's snoop logic and control path. After embedding our BIU into ARS03, we evaluated ARS03's performance with five application programs of avionics. Simulation results, as shown in Fig.4, indicate that our design of BIU, as compared with traditional design of BIU, can reduce ARS03's execution time by 17.8% as an average.
出处 《西北工业大学学报》 EI CAS CSCD 北大核心 2004年第3期370-374,共5页 Journal of Northwestern Polytechnical University
关键词 总线接口部件 微处理器 缓冲队列 Bus Interface Unit (BIU), 32-bit RISC microprocessor, buffer queue
  • 相关文献

参考文献6

  • 1Hennessy J L, Patterson D A. Computer Architecture: A Quantitative Approach. 3 San Mateo: Morgan Kaufmann Publishers, 2002
  • 2Vinodh C, Bruce J. Concurrency Latency or System Overhead: Which has the Largest Impact on Uniprocessor DRAMSystem Performance? In: Proceedings of 28th International Symposium on Computer Architecture, Goteborg Sweden, 2001: 62~71
  • 3PowerPC Microprocessor Family: The Programming Environments for 32-Bit Microprocessors. Motorola Inc, 1997
  • 4PowerPC 603e and EM603e RISC Microprocessor Family User's Manual. IBM, 1998
  • 5Clifford E C. Synthesis and Scripting Techniques for Designing Multi-Asynchronous Clock Designs. SNUG(Synopsys Users Group) Conference, San Jose, CA, 2001
  • 6马婉良,高德远,张盛兵.微处理器设计中提高访存效率的一种方法[J].西北工业大学学报,1999,17(3):338-343. 被引量:5

共引文献4

同被引文献9

  • 1M Hennessy, Patterson. Computer Architecture:A Quantitative Approach (3Edition)[M].Tsinghua University Press,2001.
  • 2J Yeh, et al. Increasing the Instruction Fetch Rate via Multiple Branch Prediction and aBranch Address Cache[C]. Tokyo:Confe-rence Proceedings '93 International Conference on Supercomputing, 1993.67-76.
  • 3Motorola.PowerPC Microprocessor Family: The Programming Environments for 32-Bit Microprocessors[R].Motorola Inc., 1997.
  • 4Motorola.MPC750 RISC Microprocessor Family User's Manual[R].Motorola Inc., 2001.
  • 5夏宇闻.Verilog数字系统设计教程,2001
  • 6Clifford E.Simulation and synthesis techniques for asynchronous FIFO design.SNUG San Jose,2001
  • 7William J Dally,John W Poulton.Digital systems engineering,Cambridge University Press,1998:468
  • 8马婉良,高德远,张盛兵.微处理器设计中提高访存效率的一种方法[J].西北工业大学学报,1999,17(3):338-343. 被引量:5
  • 9吴自信,张嗣忠.异步FIFO结构及FPGA设计[J].单片机与嵌入式系统应用,2003,3(8):24-26. 被引量:23

引证文献2

二级引证文献49

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部