Folklore Confirmed: Compiling for Speed $$=$$ Compiling for Energy

Yuki, Tomofumi; Rajopadhye, Sanjay

doi:10.1007/978-3-319-09967-5_10

Cited by 29 publications

(36 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Runtime costs corresponds closely to energy costs (consistent with findings in, e.g. [50] for CPU systems). We comment on extreme cases for runtime comparisons (energy comparisons are similar) and give the median for both runtime and energy costs.…”

Section: The Cost Of Fencessupporting

confidence: 87%

Exposing errors related to weak memory in GPU applications

Sorensen

Donaldson

2016

SIGPLAN Not.

View full text Add to dashboard Cite

We present the systematic design of a testing environment that uses stressing and fuzzing to reveal errors in GPU applications that arise due to weak memory effects. We evaluate our approach on seven GPUs spanning three Nvidia architectures, across ten CUDA applications that use fine-grained concurrency. Our results show that applications that rarely or never exhibit errors related to weak memory when executed natively can readily exhibit these errors when executed in our testing environment. Our testing environment also provides a means to help identify the root causes of such errors, and automatically suggests how to insert fences that harden an application against weak memory bugs. To understand the cost of GPU fences, we benchmark applications with fences provided by the hardening strategy as well as a more conservative, sound fencing strategy.

show abstract

Section: The Cost Of Fencessupporting

confidence: 87%

Exposing errors related to weak memory in GPU applications

Sorensen

Donaldson

2016

SIGPLAN Not.

View full text Add to dashboard Cite

show abstract

“…Our solution gracefully blends DVFS [11,25,31,38], employing the decoupled access-execute model for memory-bound applications, and race-to-sleep [44] methods, running the original version under maximum frequency for compute-bound applications.…”

Section: Discussionmentioning

confidence: 99%

“…While the behavior of complex applications is challenging for static analysis and for hardware predictors, software multi-versioning DAE (SMV-DAE) improves energy delay product (EDP), by over 20% on average (over 70% peak), and provides energy benefits of over 30% for memory-bound general-purpose applications. Furthermore, SMV-DAE is a self-adapting technique, automatically selecting the best performing access version at runtime, and a self-healing technique, resorting to the original code version for applications which are not amenable to decoupling (compute-bound applications), hence the race-to-sleep technique [44] is applied instead.…”

Section: Contributionsmentioning

confidence: 99%

Multiversioned decoupled access-execute: the key to energy-efficient compilation of general-purpose programs

Koukos

Ekemark

Zacharopoulos

et al. 2016

Proceedings of the 25th International Conference on Compiler Construction

View full text Add to dashboard Cite

Computer architecture design faces an era of great challenges in an attempt to simultaneously improve performance and energy efficiency. Previous hardware techniques for energy management become severely limited, and thus, compilers play an essential role in matching the software to the more restricted hardware capabilities. One promising approach is software decoupled access-execute (DAE), in which the compiler transforms the code into coarsegrain phases that are well-matched to the Dynamic Voltage and Frequency Scaling (DVFS) capabilities of the hardware. While this method is proved efficient for statically analyzable codes, generalpurpose applications pose significant challenges due to pointer aliasing, complex control flow and unknown runtime events. We propose a universal compile-time method to decouple generalpurpose applications, using simple but efficient heuristics. Our solutions overcome the challenges of complex code and show that automatic decoupled execution significantly reduces the energy expenditure of irregular or memory-bound applications and even yields slight performance boosts. Overall, our technique achieves over 20% on average energy-delay-product (EDP) improvements (energy over 15% and performance over 5%) across 14 benchmarks from SPEC CPU 2006 and Parboil benchmark suites, with peak EDP improvements surpassing 70%.

show abstract

“…Jimborean et al [9] propose compile-time code transformations that decouple memory accesses and computation in order to adapt the code for more efficient DVFS. Yet, as proved by Yuki et al [17], frequency scaling is mostly suitable for memory bound codes, whereas compute-bound codes exhibit significant performance degradations when run at lower frequencies. Compute-bound applications are generally optimized for performance, which is commonly known as "race to sleep" [17], yielding energy savings as positive side-effects.…”

Section: Related Workmentioning

confidence: 99%

“…Yet, as proved by Yuki et al [17], frequency scaling is mostly suitable for memory bound codes, whereas compute-bound codes exhibit significant performance degradations when run at lower frequencies. Compute-bound applications are generally optimized for performance, which is commonly known as "race to sleep" [17], yielding energy savings as positive side-effects. In contrast, Saputra et al [14] apply traditional compiler optimizations (loop fusion, tiling, etc) and scale down frequency to a lower value, which reduces the energy, while maintaining the performance of the original (unoptimized) code under maximum frequency.…”

Section: Related Workmentioning

confidence: 99%

Software-controlled processor stalls for time and energy efficient data locality optimization

Clauss

Fassi

Jimborean

2014

2014 International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation (SAMOS XIV)

View full text Add to dashboard Cite

Abstract-Data locality optimization is a well-known goal when handling programs that must run as fast as possible or use a minimum amount of energy. However, usual techniques never address the significant impact of numerous stalled processor cycles that may occur when consecutive load and store instructions are accessing the same memory location. We show that two versions of the same program may exhibit similar memory performance, while performing very differently regarding their execution times because of the stalled processor cycles generated by many pipeline hazards. We propose a new programming structure called "xfor", enabling the explicit control of the way data locality is optimized in a program and thus, to control the amount of stalled processor cycles. We show the benefits of xfor regarding execution time and energy saving.

show abstract

Folklore Confirmed: Compiling for Speed $$=$$ Compiling for Energy

Cited by 29 publications

References 20 publications

Exposing errors related to weak memory in GPU applications

Exposing errors related to weak memory in GPU applications

Multiversioned decoupled access-execute: the key to energy-efficient compilation of general-purpose programs

Software-controlled processor stalls for time and energy efficient data locality optimization

Contact Info

Product

Resources

About