Vacuum packing: extracting hardware-detected program phases for post-link optimization

Barnes, R. D.; Nyström, Erik; Merten, Matthew C.; Hwu, Wen mei

doi:10.1109/micro.2002.1176253

Cited by 18 publications

(21 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…These data allows the program behaviour to be more directly understood from the viewpoint of the underlying hardware platform, and although low level, this information can be used for guiding higher level adaptive behaviour. Barnes et al use hardware profiling to detect hot code regions and apply code optimizations efficiently [6]. Schneider and Gross present a runtime compiler framework using instruction level information provided by hardware counters to detect hot spots and bottlenecks [33].…”

Section: Related Workmentioning

confidence: 99%

Phase-based adaptive recompilation in a JVM

Verbrugge

2008

Proceedings of the 6th Annual IEEE/ACM International Symposium on Code Generation and Optimization

View full text Add to dashboard Cite

Modern JIT compilers often employ multi-level recompilation strategies as a means of ensuring the most used code is also the most highly optimized, balancing optimization costs and expected future performance. Accurate selection of code to compile and level of optimization to apply is thus important to performance. In this paper we investigate the effect of an improved recompilation strategy for a Java virtual machine. Our design makes use of a lightweight, low-level profiling mechanism to detect high-level, variable length phases in program execution. Phases are then used to guide adaptive recompilation choices, improving performance. We develop both an offline implementation based on trace data and a self-contained online version. Our offline study shows an average speedup of 8.7% and up to 21%, and our online system achieves an average speedup of 4.4%, up to 18%. We subject our results to extensive analysis and show that our design achieves good overall performance with high consistency despite the existence of many complex and interacting factors in such an environment.

show abstract

Section: Related Workmentioning

confidence: 99%

Phase-based adaptive recompilation in a JVM

Verbrugge

2008

Proceedings of the 6th Annual IEEE/ACM International Symposium on Code Generation and Optimization

View full text Add to dashboard Cite

show abstract

“…Merten et al [5] use the phase information to develop a run-time system for dynamically optimizing frequently executed code. Barnes et al [6] extend this idea to perform phase-directed compiler optimizations. Biesbrouck et al [7] use phase behavior to guide simulation for Simultaneous Multithreading [8].…”

Section: Introductionmentioning

confidence: 99%

“…The phase analysis for serial programs is already a mature area [1,6,[10][11][12][13][14][15][16][17]. Architectural metrics (for example, CPI, cache miss rate or hit rate, branch frequency) and code signatures have been examined to extract phase information.…”

Section: Introductionmentioning

confidence: 99%

Analyzing the impact of on-chip network traffic on program phases for CMPs

Ozisikyilmaz

Memik

Kim

et al. 2009

2009 IEEE International Symposium on Performance Analysis of Systems and Software

View full text Add to dashboard Cite

Abstract-It is known that the execution of programs exhibits repetitive phases; in other words, the execution of programs can be partitioned into segments of execution, during which the application exhibits unique architectural properties. This property has been used for various optimization goals. In addition, phase information is utilized to reduce the run time of the architectural simulation. Conventionally, an application is examined in an architecture-independent manner (such as the number of times a basic block is executed) to extract information about the phases and then only the representative execution intervals are executed to analyze architectural choices. We claim that such approaches are becoming inadequate in the many-core era as application execution is not dominated by the instructions only, but instead the communication structure of the application is becoming as important as the instruction behavior. Hence, we propose to utilize communication behavior to determine the phases of an application. Our results reveal that the inclusion of the communication information can increase the accuracy of the phase detection significantly. Specifically, for SPLASH2 and MineBench applications, the average (geometric mean) CPI error rate with the instruction-based phase detection is 11.01%, while our phase detection scheme has an average error rate of 3.41% when compared to the simulations that run the applications to completion.

show abstract

“…Phase behavior can be exploited for accurate architecture simulation [27,28], to save energy by dynamically reconfiguring caches and processor width [1,29,8,7], to guide compiler optimizations [20,2], to guide remote profiling [21], and to choose which core to run a process on in a multi-core architecture [15]. All of these techniques take advantage of the phase behavior that exists in programs, and most of them focus on the phase behavior seen at a specific granularity (fixed length interval length) of execution.…”

Section: Introductionmentioning

confidence: 99%