國立高雄大學圖資館 |

Language: English

Back

Enhancing software reliability with ...

Oplinger, Jeffrey Thomas.

Enhancing software reliability with speculative threads.

Record Type:	Electronic resources : Monograph/item
Title/Author:	Enhancing software reliability with speculative threads.
Author:	Oplinger, Jeffrey Thomas.
Description:	111 p.
Notes:	Adviser: Monica S. Lam.
Notes:	Source: Dissertation Abstracts International, Volume: 65-09, Section: B, page: 4745.
Contained By:	Dissertation Abstracts International65-09B.
Subject:	Engineering, Electronics and Electrical.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3145591
ISBN:	0496045539

Enhancing software reliability with speculative threads.
Oplinger, Jeffrey Thomas.

Enhancing software reliability with speculative threads. - 111 p.

Adviser: Monica S. Lam.

Thesis (Ph.D.)--Stanford University, 2004.

As high-end microprocessors continue to provide more and more computing power, non-performance metrics such as security, availability, reliability, and usability have become much more important. Errors and vulnerabilities in software programs have caused significant losses of data and productivity throughout the world. Software tools are available to help identify and prevent these problems, but they are often not used in practice because of large runtime overheads and limited applicability. We believe that hardware support can make these existing tools faster and more useful, as well as provide new functionality that helps programmers write safer code.

ISBN: 0496045539Subjects--Topical Terms:

226981
Engineering, Electronics and Electrical.

Enhancing software reliability with speculative threads.
LDR:02918nmm _2200289 _450 001 162898
005 20051017073534.5
008 090528s2004 eng d
020 $a 0496045539
035 $a 00149399
040 $a UnM $c UnM
100 0 $a Oplinger, Jeffrey Thomas. $3 228044
245 1 0 $a Enhancing software reliability with speculative threads.
300 $a 111 p.
500 $a Adviser: Monica S. Lam.
500 $a Source: Dissertation Abstracts International, Volume: 65-09, Section: B, page: 4745.
502 $a Thesis (Ph.D.)--Stanford University, 2004.
520 # $a As high-end microprocessors continue to provide more and more computing power, non-performance metrics such as security, availability, reliability, and usability have become much more important. Errors and vulnerabilities in software programs have caused significant losses of data and productivity throughout the world. Software tools are available to help identify and prevent these problems, but they are often not used in practice because of large runtime overheads and limited applicability. We believe that hardware support can make these existing tools faster and more useful, as well as provide new functionality that helps programmers write safer code.
520 # $a Our experimental results suggest that monitored execution is more amenable to this parallelization model than general-purpose execution. Applying thread-level speculation improves the running time of monitored code by a factor of 1.7. Together with a 1.6-times speedup from exploiting additional single-thread instruction-level parallelism (ILP), an overall speedup of 2.6 is obtained---effectively 5.8 instructions per cycle in performance. Through a number of real-life examples, we also show how fine-grain transactional programming can be used to detect and recover from buffer overflow exploits.
520 # $a We suggest using a monitor-and-recover programming paradigm to enhance software reliability and propose an architectural design based on thread-level speculation (TLS) that makes this paradigm more efficient and easier to program. Programmers can add monitoring code, with normal sequential semantics, to examine the execution of a program. Our architecture reduces the resulting slowdown by speculatively executing the monitoring code in parallel with the main computation. To recover from errors, programmers can define fine-grain transactions whose side effects are either committed or aborted via program control. These transactions are implemented efficiently through TLS hardware support.
590 $a School code: 0212.
650 # 0 $a Engineering, Electronics and Electrical. $3 226981
650 # 0 $a Computer Science. $3 212513
690 $a 0544
690 $a 0984
710 0 # $a Stanford University. $3 212607
773 0 # $g 65-09B. $t Dissertation Abstracts International
790 $a 0212
790 1 0 $a Lam, Monica S., $e advisor
791 $a Ph.D.
792 $a 2004
856 4 0 $u http://libsw.nuk.edu.tw:81/login?url=http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3145591 $z http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3145591