跳到主要內容

臺灣博碩士論文加值系統

(18.97.14.87) 您好!臺灣時間:2025/02/13 05:17
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:張季青
研究生(外文):Ji-Ching Chang
論文名稱:精簡指令集電腦架構之合併數值預測與資料重複使用
論文名稱(外文):Combine the Value Prediction and Data Reuse in RISC Architecture
指導教授:謝忠健謝忠健引用關係
指導教授(外文):Jong-Jiann Shieh
學位類別:碩士
校院名稱:大同工學院
系所名稱:資訊工程研究所
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:1999
畢業學年度:87
語文別:英文
論文頁數:88
中文關鍵詞:數值預測資料重複使用指令階層平行度數值預測表重複使用緩衝器
外文關鍵詞:VPDRILPVPTRB
相關次數:
  • 被引用被引用:0
  • 點閱點閱:100
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
中文摘要
最近兩種硬體的技術─數值預測和資料重複使用─被提出來,以開發利用程式中相同的部份並嘗試消除資料相依關係。
數值預測是一種預測執行的方法,能夠開發程式中可預測的資訊去預測數值。我們從數值預測表中擷取預測的值當作下一個指令的輸入。與數值預測方法不同;資料重複使用是一種非預測執行的方法,經由擷取指令的輸入運算元及計算結果,利用重複使用緩衝器去記錄程式中可以重複使用的部份。因此;如果輸入是相同的,指令執行單元可以不經由運算而直接得到結果。
現在,我們不只企圖去瞭解數值預測和資料重複使用彼此間的差異性,更進一步的;我們把這兩個架構整合在一起。我們提出一個新的方法,經由合併數值預測和資料重複使用這兩個方法的優點,可以提昇預測來源運算元的比率並且增加指令階層平行度。
Abstract
Recently two hardware techniques --- Value Prediction (VP) and Data Reuse (DR) --- have been proposed for exploiting redundancy in programs to collapse data dependences .
VP is a speculative technique that exploits predictable information in program to predict values. The predicted values are obtained from VPT (Value Prediction Table) which are used as inputs by instructions. Unlike VP, DR is a non-speculative technique that use the Reuse Buffer(RB) to record reusable part in programs by obtaining input operands and results of instructions. The instruction execution unit get the the result without execute if the instruction have the same input.
Here, we attempt to not only understand the difference between the VP and DR , but also try to combine them together. We propose a new method that predict source operands from combining the advantage of VP and DR that can increase predictable rate and improve ILP.

Acknoledgement.......................................Ⅵ
Abstract in Chinese..................................Ⅶ
Abstract in English..................................Ⅷ
Table of Contents....................................Ⅸ
List of Figures....................................ⅩⅡ
List of Tables.....................................ⅩⅣ
Chapter 1 Introduction...............................1
Chapter 2 Related work...............................7
Chapter 3 Value Prediction and Data Reuse...........12
3.1 VALUE PREDICTION.................................12
3.1.1 Classification of Value Sequences..............12
3.1.2 Data Value Prediction Models...................15
3.1.2.1 Computational Predictors.....................15
3.1.2.2 Context Based Predictors.....................18
3.2 DATA REUSE.......................................22
3.2.1 The Concept of Data Reuse......................22
3.2.2 Schemes for Data Reuse.........................25
3.2.2.1 Scheme Sv: Reuse based upon operand values...28
3.2.2.2 Scheme Sn: Reuse based upon register names...31
3.3 THE DIFFERENCE BETWEEN VP & DR...................33
3.3.1 Value Prediction and Instruction Reuse.........33
3.3.2 Impact of Differences..........................36
3.3.2.1 Validation ...................................36
3.3.2.2 Amount of Redundancy Captured................36
3.3.2.3 Misprediction................................37
3.3.2.4 Resource Contention ..........................38
3.3.2.5 Execution Latency............................39
Chapter 4 The Design of Combining VP & DR...........41
4.1 THE ARCHITECTURE OVERVIEW........................42
4.1.1 The combining Model............................42
4.1.2 Interface Signal...............................44
4.2 COMBINE THE VP AND DR............................46
4.2.1 The Information in RAM Module..................46
4.2.1.1 The RB Entry.................................46
4.2.1.2 The VPT Entry................................47
4.2.2 The Relation of Combining Architecture.........48
4.2.3 Algorithm of C.C.+RT...........................50
4.2.4 Algorithm of RD_update.........................53
4.3 THE DESIGN OF COMBINING ARCHITECTURE.............55
4.3.1 The design of C.C.+RT..........................55
4.3.1.1 RB hit ( actual result ).....................55
4.3.1.2 VPT hit  RB hit ( speculative result ).....56
4.3.1.3 VPT hit  RB miss ( speculative operand )...57
4.3.1.4 VPT miss  RB miss..........................57
4.3.2 The design of RD_update........................58
Chapter 5 Experimental Result........................59
Chapter 6 Conclusion.................................63
Reference ............................................65
Appendix A...........................................68

Reference
[1]Y Sazeides and J. E. Smith, “The Predictability of Data Values,” in Proc. of 30th Annual intemational Symposium on Microarchitecture (MICRO-30), pages 248-258, Dec. 1997.
[2]A. Sodani and G. S. Sohi, “Dynamic Instruction Reuse,” in Proc. of 24th Annual International Symposium on Computer Architecture, pages 194-205, July 1997.
[3]D. Burger, T. M. Austin, and S. Bennett, “Evaluating Future Microprocessors: The SimpleScalar Tool Set.” Technical Report CS-TR-961308, University of Wisconsin-Madison, July 1996. (URL: http://www.cs.wisc.edu/-mscalar/simplescalar.html)
[4]S. P. Harbison. “A Computer Architecture for the Dynamic Optimization of High-Level Linguage Programs.” Ph.D. thesis, Carnegie Mellon University, Sept. 1980.
[5]M. H. Lipasti and J. P. Shen, “Exceeding the dataflow limit via value prediction,” in Proceedings of the 29th Annual ACM/IEEE International Symposium and Workshop on Microarchitecture, pp. 226-237, December 1996.
[6]M. H. Lipasti, C. B. Wilkerson, and J. P. Shen, “Value locality and data speculation,” in Proceedings of the 7th International Conference on Architectural Support for Programming Languages and Operating Systems, pp. 138-147, October 1996.
[7]A. Sodani and G. S. Sohi, “Understanding the Difference Between Value Prediction and Instruction Reuse,” in Proc. of 24th Annual International Symposium on Computer Architecture, pages 194-205, July 1997.
[8]S. Mehrotra and L. Harrison, “Examination of a memory access classification scheme for pointer intensive and numeric programs,” in Proceedings of the l0th International Conference on Supercomputing, May 1996.
[9]D. Joseph and D. Grunwald, “Prefetching using markov predictors,” in Proceedings of the 24th International Symposium on Computer Architecture, pp. 252-263, June 1997.
[10]R. J. Eickerneyer and S. Vassiliadis, “A load instruction unit for pipelined processors,” IBM Journal of Research and Development., vol. 37, pp. 547-564, July 1993.
[11]J. Gonzalez and A. Gonzalez, “Speculative execution via address prediction and data prefetching,” in Proceedings of the 11th International Conference on Supercomputing, pp. 196-203, July 1997.
[12]A. Mendelson and F. Gabbay, “Speculative execution based on value prediction.” Tech. Rep. (Available from http://www-ee.technion.ac.il/fredg), Technion, 1997.
[13]Yichen Liu, “Study on Instruction Reuse in RISC Architecture.” Master thesis, Tatung Institute of Technology , Juan 1999.
[14]T.-Y. Ych and Y. N. Patt, “Alternative implementations of two-level adaptive branch prediction,” in Proceedings of the 19th International Symposium on Computer Architecture, pp. 124-134, May 1992.
[15]P.-Y Chang, E. Hao, and Y N. Patt, “Target prediction for indirect jumps,” in Proceedings of the 24th International Symposium on Computer Architecture, pp. 274-283, June 1997
[16]C. Young and M. D. Smith, “Improving the accuracy of static branch prediction using branch correlation,” in Proceedings of the 6th International Conference on Architectural Supportfor Programming Languages and Operating Systems, pp. 232-241, October 1994.
[17]R. Nair, “Dynamic path-based branch correlation,” in Proceedings of the 28th Annual ACM/IEEE International Syrnposium and Workshop on Microarchitecture, pp. 15-23, December 1995.
[18]S. Mahlke and B. Natarajan, “Compiler synthesized dynamic branch prediction,” in Proceedings of the 29th Annual ACM/IEEE International Symposium and Workshop on Microarchitecture, pp. 153-164, Decernber 1996.
[19]L-C. K. Cheng, J. T. Coffey, and T. N. Mudge, “Analysis of branch prediction via data compression,” in Proceedings of the 7th International Conference on Architectural Support for Prograrnming Languages and Operating Systems, October 1996.
[20]S. P. Harbison, “An architectural alternative to optimizing compilers,” in Proc. International Conference on Architectural Support for Programming Languages and Operating Systems (ASPLOS), page 57 65, Mar. 1982.
[21]S. E. Richardson, “Caching function results: Faster arithmetic by avoiding unnecessary computation.” Technical Report SMLI TR-92- 1, Sun Microsystems Laboratories, Sept. 1992.
[22]S. F. Oberman and M. J. Flynn,”On Division and Reciprocal Caches.” Technical Report CSL-TR-95-666, Stanford University, Apr. 1995.
[23]F. Gabbay and A. Mendelson, “Using Value Prediction to Increase the Power of Speculative Execution Hardware.” ACM Transaction on Computer Systems (TOCS), Aug. 1998.
[24]J.L. Hennessy and D.A. Patterson. Computer Architecture a Quantitative Approach. Morgan Kaufman Publishers, Inc., Second edition, 1995

QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top