臺灣博碩士論文加值系統

對於像是科學計算或者模擬之類的計算密集程式，開發及除錯是個累人且耗時的工作，這是因為這類程式通常花很多時間執行，且開發及除錯時又需要反覆執行的原因，雖然有一些檢查點/恢復的機制被提出來加快開發程式，但是開發人員在修改及編譯程式後依然需要重頭執行，這些不必要的重複執行造成軟體開發的生產力低落.

這篇論文中我們提出了一個函式導向的檢查點/恢復機制(FOCR)來減少Linux上的計算密集程式的開發時間，FOCR讓程式開發人員可以讓程式在修改過後從修改過的函式開始執行，也就是不需要每次都重頭執行沒修改過的部份，這機制不同於以往的方法，它讓計算密集程式的開發時間縮短，對增進軟體開發的生產力有很大幫助。

For computation-intensive programs such as those used in scientific computation and those designed for simulation, developing and debugging is an exhausting and time-consuming job because of repeated program execution. Although there were some proposed checkpoint/restore mechanisms to facilitate developing computation-intensive programs, developers still suffer from time-consuming execution after slightly modifying source codes of programs. During the phase of repeated execution, it provides no contribution to the developers' productivity.

In this paper, we propose a function-oriented checkpoint/restoration mechanism (FOCR) to accelerate execution of computation-intensive programs in Linux. FOCR allows developers to continue the program execution from the modified point even though source codes are slightly modified, which is totally different to other proposals. With the help of this mechanism, valuable contribution to the developers' productivity is provided by shortening edit-compile-debug cycle.

Chapter 1 Introduction 1
Chapter 2 Background and Related work 3
2.1 Background 3
2.1.1 Ptrace library 3
2.1.2 Process Memory Layout in Linux 3
2.2 Related work 5
2.2.1 Process checkpointing 5
Chapter 3 System design 8
3.1 Approach 8
3.2 System architecture 11
3.3 System Component Description 11
Chapter 4 Implementation 15
4.1 Function Pre-processor(FP) 15
4.2 Function Information Recorder(FIR) 16
4.3 Data Recorder(DR) 17
4.4 Stack Recorder(SR) 17
4.5 Heap Recorder(HR) 19
4.6 Modified Function Analyzer(MFA) 19
4.7 Program Restorer(PR) 21
Chapter 5 Experiment 24
5.1 Environment 24
5.2 Applications 24
5.2.1 Description 25
5.3 Results 26
Chapter 6 Discussion 32
Chapter 7 Conclusion 34
References 35

[1]J. S. Plank, M. Beck, G. Kingsley, and K. Li, Libckpt: Transparent checkpointing under unix: University of Tennessee], Computer Science Department, 1994.
[2]S. Kraemer, R. Leupers, D. Petras, and T. Philipp, A checkpoint/restore framework for systemc-based virtual platforms, in System-on-Chip, 2009. SOC 2009. International Symposium on, 2009, pp. 161-167.
[3]D. Subhraveti and J. Nieh, Record and transplay: partial checkpointing for replay debugging across heterogeneous systems, in Proceedings of the ACM SIGMETRICS joint international conference on Measurement and modeling of computer systems, 2011, pp. 109-120.
[4]J. Engblom, Transporting bugs with checkpoints, in Syst., Softw., SoC, Silicon Debug Conf.(S4D), Southampton, UK, 2010.
[5]Z. Guo, X. Wang, J. Tang, X. Liu, Z. Xu, M. Wu, et al., R2: An application-level kernel for record and replay, in Proceedings of the 8th USENIX conference on Operating systems design and implementation, 2008, pp. 193-208.
[6]B. Lewis, Debugging backwards in time, arXiv preprint cs/0310016, 2003.
[7]S. Narayanasamy, G. Pokam, and B. Calder, Bugnet: Continuously recording program execution for deterministic replay debugging, in ACM SIGARCH Computer Architecture News, 2005, pp. 284-295.
[8]A. Slowinska, T. Stancescu, and H. Bos, Howard: A Dynamic Excavator for Reverse Engineering Data Structures, in NDSS, 2011.
[9]G. Vallée, R. Lottiaux, D. Margery, C. Morin, and J.-Y. Berthou, Ghost process: a sound basis to implement process duplication, migration and checkpoint/restart in linux clusters, in Parallel and Distributed Computing, 2005. ISPDC 2005. The 4th International Symposium on, 2005, pp. 97-104.
[10]M. Bozyigit and M. Wasiq, User-level process checkpoint and restore for migration, ACM SIGOPS Operating Systems Review, vol. 35, pp. 86-96, 2001.
[11]E. Roman, A survey of checkpoint/restart implementations, in Lawrence Berkeley National Laboratory, Tech, 2002.
[12]J. Duell, P. H. Hargrove, and E. S. Roman, Requirements for linux checkpoint/restart, 2002.