Deferring accelerator offloading decisions to application runtime

Vaz, Gavin; Riebler, Heinrich; Kenter, Tobias; Plessl, Christian

doi:10.1109/reconfig.2014.7032509

Cited by 1 publication

(2 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For the evaluated benchmark applications, a minimum score of 16 is required to include them all. More details on the decision process can be found in our previous work [10,11,61]. After the CGD, HTrOP requests the Parallelization Engine to generate the OpenCL kernel .…”

Section: Code Generation Decisionmentioning

confidence: 99%

“…Chain of convolutions to enhance image quality heat2D [47] 25 Simulation of heat transfer in 2D nbody [31] 20 N-body particle simulation (kernel) motion [50] 82 Motion detection; application contains five hotspots raytrace [64] 68 Calculate rays from point of light (kernel) sha256 [15] 37 Cryptographic hash function with 256-bit digest stereo3D [61] 23 3D stereo matching Lines of code (LoCs) are related to the hotspot size.…”

Section: Application Loc Hotspotmentioning

confidence: 99%

See 1 more Smart Citation

Transparent Acceleration for Heterogeneous Platforms With Compilation to OpenCL

Riebler

Vaz

Kenter

et al. 2019

ACM Trans. Archit. Code Optim.

Self Cite

View full text Add to dashboard Cite

Multi-accelerator platforms combine CPUs and different accelerator architectures within a single compute node. Such systems are capable of processing parallel workloads very efficiently while being more energy efficient than regular systems consisting of CPUs only. However, the architectures of such systems are diverse, forcing developers to port applications to each accelerator using different programming languages, models, tools, and compilers. Developers not only require domain-specific knowledge but also need to understand the low-level accelerator details, leading to an increase in the design effort and costs.To tackle this challenge, we propose a compilation approach and a practical realization called HTrOP that is completely transparent to the user. HTrOP is able to automatically analyze a sequential CPU application, detect computational hotspots, and generate parallel OpenCL host and kernel code. The potential of HTrOP is demonstrated by offloading hotspots to different OpenCL-enabled resources (currently the CPU, the generalpurpose GPU, and the manycore Intel Xeon Phi) for a broad set of benchmark applications. We present an in-depth evaluation of our approach in terms of performance gains and energy savings, taking into account all static and dynamic overheads. We are able to achieve speedups and energy savings of up to two orders of magnitude, if an application has sufficient computational intensity, when compared to a natively compiled application.This article is an extension of a conference paper/poster at PPoPP'18 [49]. Prior work only outlined the tool flow and the parallelization mechanism. Evaluation is entirely new, including energy measurements and trends over several data points per application.

show abstract

Section: Code Generation Decisionmentioning

confidence: 99%

Section: Application Loc Hotspotmentioning

confidence: 99%