The Transmeta Code Morphing/spl trade/ Software: using speculation, recovery, and adaptive retranslation to address real-life challenges

Dehnert, James C.; Grant, B.K.; Banning, John; Johnson, Roger D.; Kistler, Thomas; Klaiber, Alexander C.; Mattson, Jim

doi:10.1109/cgo.2003.1191529

Cited by 174 publications

(235 citation statements)

References 17 publications

Supporting

Mentioning

228

Contrasting

Unclassified

Order By: Relevance

“…It is used to recover from branch misprediction and out-of-order execution [33,38], hardware faults [32,34], speculative memory-reordering in VLIW machines [13,18], optimistic dynamic binary translation and code optimization [11,14], and transactional memory [17,28]. In each of these cases, recovery is used Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page.…”

Section: Introductionmentioning

confidence: 99%

Static analysis and compiler design for idempotent processing

Kruijf

Sankaralingam

Jha

2012

Proceedings of the 33rd ACM SIGPLAN Conference on Programming Language Design and Implementation

View full text Add to dashboard Cite

Recovery functionality has many applications in computing systems, from speculation recovery in modern microprocessors to fault recovery in high-reliability systems. Modern systems commonly recover using checkpoints. However, checkpoints introduce overheads, add complexity, and often save more state than necessary.This paper develops a novel compiler technique to recover program state without the overheads of explicit checkpoints. The technique breaks programs into idempotent regions-regions that can be freely re-executed-which allows recovery without checkpointed state. Leveraging the property of idempotence, recovery can be obtained by simple re-execution. We develop static analysis techniques to construct these regions and demonstrate low overheads and large region sizes for an LLVM-based implementation. Across a set of diverse benchmark suites, we construct idempotent regions close in size to those that could be obtained with perfect runtime information. Although the resulting code runs more slowly, typical performance overheads are in the range of just 2-12%.The paradigm of executing entire programs as a series of idempotent regions we call idempotent processing, and it has many applications in computer systems. As a concrete example, we demonstrate it applied to the problem of compiler-automated hardware fault recovery. In comparison to two other state-of-the-art techniques, redundant execution and checkpoint-logging, our idempotent processing technique outperforms both by over 15%.

show abstract

Section: Introductionmentioning

confidence: 99%

Static analysis and compiler design for idempotent processing

Kruijf

Sankaralingam

Jha

2012

Proceedings of the 33rd ACM SIGPLAN Conference on Programming Language Design and Implementation

View full text Add to dashboard Cite

show abstract

“…While in theory these run-time annotations hold only for a given input, in practice the dependency information tends to be fairly stable across inputs. We are not the first to make this observation: compiler researchers have proposed using runtime dependency information for profile-driven speculative optimizations [24] and run-time system builders have implemented some of these optimizations in production systems [8]. COMPASS, which does not have the same strong correctness criteria as compilers, will also stand to benefit from this kind of representation.…”

Section: Program Representationmentioning

confidence: 99%

COMPASS: A Community-driven Parallelization Advisor for Sequential Software

Sethumadhavan

Arora

Ganapathi

et al. 2009

2009 ICSE Workshop on Multicore Software Engineering

View full text Add to dashboard Cite

show abstract

“…Each micro-operation places its result in one resource, r ko , at most. Now, we are able to model W k , the operations of each instruction, i k over the interval T k , as a sequence of triplets shown in (14). Now,…”

Section: Formulating the Simple Modelmentioning

confidence: 99%

“…In a co-designed virtual machine the source architecture, that is the one visible to the binary applications running on the machine, is emulated on a target architecture. One of the most well-known co-designed virtual machines is the Transmeta Crusoe processor [13], which uses a "code morphing" (CMS) layer [14] to transparently run Intel IA-32-based software (source architecture) on an underlying VLIW (Very Long Instruction Word) architecture (target architecture). Using our model, this CMS layer can be implemented in 3 n I , in order to maximize performance or it can be implemented in 3 m R , in order to maximize design flexibility and postproduction maintenance.…”

Section: A Microprogrammingmentioning

confidence: 99%

A Model of Computer Architecture with Applications

Mutigwe¹,

Kinyua²,

Aghdasi³

2013

IJCTE

View full text Add to dashboard Cite

I. INTRODUCTIONThe design of general-purpose computers has, until recently, been largely a qualitative exercise [1]. The work by Flynn [2] and Hennessy and Patterson [3] has helped start a reversal of that trend by putting the design of computers on a quantitative footing. Along those lines, in this work we propose to formalize the description of computer architectures. We develop a generic model for computing architecture that has the following three features. 1) Fidelity: The model accurately represents the structural parts of a given architecture and how they are related, while at the same time it allows for abstract computer models to represent the operations of these parts and their interactions. 2) Accessibility: The model is an intuitive, algebraic model that can be used by computer architects and other designers of digital systems with little training in formal mathematical methods. 3) Extensibility: The model can be used to model entire systems or the individual components of such systems. Our development of the model presented in this paper looks at both the static and dynamic aspects of a computers architecture. The static aspects are the associations between the components. While the dynamic aspects describe how the connected components interact in order to perform the computations.The rest of the paper is organized as follows. Section II describes some related work and defines some key terms.Manuscript

show abstract

The Transmeta Code Morphing/spl trade/ Software: using speculation, recovery, and adaptive retranslation to address real-life challenges

Cited by 174 publications

References 17 publications

Static analysis and compiler design for idempotent processing

Static analysis and compiler design for idempotent processing

COMPASS: A Community-driven Parallelization Advisor for Sequential Software

A Model of Computer Architecture with Applications

Contact Info

Product

Resources

About