Program Phase Detection based on Critical Basic Block Transitions

Ratanaworabhan, Paruj; Burtscher, Martin

doi:10.1109/ispass.2008.4510734

Cited by 13 publications

(12 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Thus, every basic block ends either with a branch instruction, or before an instruction targeted by a branch. This enables mechanisms based on basic blocks to automatically adapt to the program phase, as a program's phase is characterized by the blocks being used (RATANAWORABHAN;BURTSCHER, 2008). However, our low overhead hardware implementation allows multiple entry points, as it is not possible to efficiently detect the beginning of a block which was not targeted by a branch without changing the instruction set functionality.…”

Section: Basic Block and Relaxed Blocksmentioning

confidence: 99%

See 1 more Smart Citation

Profiling and Reducing Micro-Architecture Bottlenecks at the Hardware Level

Moreira

Alves

Koren³

2014

2014 IEEE 26th International Symposium on Computer Architecture and High Performance Computing

View full text Add to dashboard Cite

Section: Basic Block and Relaxed Blocksmentioning

confidence: 99%

“…Although several research papers used basic block analysis, most did so using a software approach, even for hardware adaptations (PANAIT; SASTURKAR;WONG, 2004;RATANAWORABHAN;BURTSCHER, 2008). One of the few techniques that actually performed basic block analysis at the hardware level was the rePlay framework (PATEL; LUMETTA, 2001).…”

Section: Introductionmentioning

confidence: 99%

Profiling and Reducing Micro-Architecture Bottlenecks at the Hardware Level

Moreira

Alves

Koren³

2014

2014 IEEE 26th International Symposium on Computer Architecture and High Performance Computing

View full text Add to dashboard Cite

“…A complete fragment cache flush is triggered whenever dynamo recognises a sharp increase in the fragment creation rate. Ratanaworabhan and Burtscher (2008) proposed a phase detection method that is based on high frequency misses of basic blocks. To detect phases, an application is profiled on some input to generate a trace of basic blocks identifiers.…”

Section: Related Workmentioning

confidence: 99%

Efficient memory control for avionics and embedded systems

Weisberg

Wiseman

2013

IJES

View full text Add to dashboard Cite

Abstract:Modern aircrafts have seen a significant growth of avionics systems. Most of new aircrafts have systems for navigation, automatic flight control, collision avoidance systems, flight data recorder, weather radar system as well as communication and monitoring systems. All of these systems use embedded computer systems with increasing memory requirements. In order to reduce the electricity consumed by these systems, we suggest detecting the portion of the computer memory that is actually used. The other portion can be temporarily shut down and turned on again when needed. Such a shut down can notably reduce the electricity consumption of the avionics systems.

show abstract

“…Sherwood et al [14], [15] and Ratanaworabhan et al [13] use basic bloc vector (BBVs) to detect program phase changes. A basic bloc vector is a list of all basic blocs entered during program execution, and a count of how many times each basic bloc was run.…”

Section: Related Workmentioning

confidence: 99%

“…Detecting program phase changes has been widely addressed in the past [7], [10], [11], [12], [13], [14]. These schemes all share in common the fact that they either require strong knowledge of the program and/or the system's architecture or that the program be compiled with specific options or instrumented.…”

Section: Introductionmentioning

confidence: 99%

A User Friendly Phase Detection Methodology for HPC Systems' Analysis

Chetsa

Lefèvre²,

Pierson³

et al. 2013

2013 IEEE International Conference on Green Computing and Communications and IEEE Internet of Things and IEEE Cyber, Physical A

View full text Add to dashboard Cite

Abstract-A wide array of today's high performance computing (HPC) applications exhibits recurring behaviours or execution phases throughout their runtime. Accurate detection of program phases allows reconfiguring the system for a better power/performance trade off; and can reduce the simulation time of programs by identifying regions of code whose performance is critical to the entire program. Program phases are also reflected in different behaviours the system goes through or system phases, which can be used as an alternative means of program phase detection for users lacking expertise.In this paper, we present an execution vector based (EVbased) phase detection, which is an on-line methodology for detecting phases in the behaviour of a HPC system and determining execution points that correspond to these phases. We also present a methodology for defining a small set of EVs representative of the system's behaviour over a fixed period of time and show that EV-based phase detection identifies recurring phases. Our methodology is illustrated with benchmarks and a real life application.

show abstract

Program Phase Detection based on Critical Basic Block Transitions

Cited by 13 publications

References 15 publications

Profiling and Reducing Micro-Architecture Bottlenecks at the Hardware Level

Profiling and Reducing Micro-Architecture Bottlenecks at the Hardware Level

Efficient memory control for avionics and embedded systems

A User Friendly Phase Detection Methodology for HPC Systems' Analysis

Contact Info

Product

Resources

About