Speedup Prediction for Selective Compilation of Embedded Java Programs

In this paper; we putforward an approach to optimize Java applications for embedded systems. Its principle is to use a specific profiler dedicated to optimization. The profiler takes as an input a Java Application and provides the best possible sequence of optimization techniques. Strategy resolution goes through the study ofthe fundamental optimization techniques and the exploration of the research space that they compose.

show abstract

“…We can also sort constant according to their type to increase redundancy and consequently improve compression ratio. 4 Overview of our profiler…”

Section: Techniques Included In the Profilermentioning

confidence: 99%

“…A compiled code is faster but its size is greater than the same interpreted code. Colin [4] proposes a compilation framework for AOT compilers. His profiler records execution time and calls frequency of each method then selects code portions to be compiled.…”

Section: Introductionmentioning

confidence: 99%

Profile-Based Optimization for Embedded Java Applications

Ammous

Benameur

Abed

2006

2006 International Symposium on Industrial Embedded Systems

View full text Add to dashboard Cite

show abstract

“…The performance of the virtual machine is the main drawback of this approach, which also requires that the virtual machine is ported to the target platform. Some of the performance issues are addressed with JIT [13] and customized embedded virtual machines [14].…”

Section: Related Workmentioning

confidence: 99%

Phantom: a serializing compiler for multitasking embedded software

Nácul¹,

Givargis²

2006

2006 American Control Conference

View full text Add to dashboard Cite

In an era of powerful general-purpose embedded compute platforms, the migration of system functionality from application specific integrated circuits (ASICs) to software has become a promising trend toward addressing the system complexity and shrinking timeto-market window challenges. Hence, in modern embedded systems, software development plays an increasingly vital role. On the other hand, the real-time concurrent programming model provides the high level abstractions necessary to effectively design complex software. Support for real-time concurrent programming is typically provided by a realtime operating system. We propose an alternate solution, the Phantom serializing compiler, to support the real-time concurrent programming model. The Phantom serializing compiler generates a single-threaded monolithic executable, from the multi-threaded application software, capable of directly executing on the underlying embedded compute platform. Further, the generated executable is tuned for maximum performance and efficiency, yielding an application-specific solution. In this paper, we give an overview of the Phantom serializing compiler and demonstrate its feasibility with some experimental results.

show abstract

“…The performance of the virtual machine is the main drawback of this approach, which also requires that the virtual machine is ported to the target platform. Some of the performance issues are addressed with JIT [Aycock 2003] and customized embedded virtual machines [Verdiere et al 2002].…”

Section: Related Workmentioning

confidence: 99%

Synthesis of time-constrained multitasking embedded software

Nácul

Givargis

2006

ACM Trans. Des. Autom. Electron. Syst.

View full text Add to dashboard Cite

In modern embedded systems, software development plays a vital role. Many key functions are being migrated to software, aiming at a shorter time to market and easier upgrades. Multitasking is increasingly common in embedded software, and many of these tasks incorporate real-time constraints. Although multitasking simplifies coding, it demands an operating system and imposes significant overhead on the system. The use of serializing compilers, such as the Phantom compiler, allows the synthesis of a monolithic code from a multitasking C application, eliminating the need for an operating system. In this article, we introduce the synthesis of multitasking applications that execute in a timely manner. We incorporate the notion of timing constraints into the Phantom compiler, and show that our approach is effective in meeting such constraints, allowing fine-grained concurrency among the tasks. As an additional case study, we present the implementation of a software-based modem and show that real-time applications such as the modem have guaranteed performance in the serialized code generated by the Phantom compiler.

show abstract

Speedup Prediction for Selective Compilation of Embedded Java Programs

Cited by 7 publications

References 5 publications

Profile-Based Optimization for Embedded Java Applications

Profile-Based Optimization for Embedded Java Applications

Phantom: a serializing compiler for multitasking embedded software

Synthesis of time-constrained multitasking embedded software

Contact Info

Product

Resources

About