Characterizing the challenges and evaluating the efficacy of a CUDA-to-OpenCL translator

Gardner, Mark K.; Sathre, Paul; Feng, Wu-chun; Martínez, Gabriel Bravo

doi:10.1016/j.parco.2013.09.003

Cited by 8 publications

(6 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CU2CL Gardner et al summarize the challenges of translating CUDA code to its OpenCL equivalence (Gardner et al 2013). They develop an automated CUDA-to-OpenCL source-to-source translator (CU2CL), to automatically translate three medium-to-large, CUDA-optimized codes to OpenCL, thus enabling the codes to run on other GPU-accelerated systems (Martinez et al 2011;Sathre et al 2019).…”

Section: Cuda-to-openclmentioning

confidence: 99%

Parallel programming models for heterogeneous many-cores: a comprehensive survey

et al. 2020

View full text Add to dashboard Cite

Heterogeneous many-cores are now an integral part of modern computing systems ranging from embedding systems to supercomputers. While heterogeneous many-core design offers the potential for energy-efficient high-performance, such potential can only be unlocked if the application programs are suitably parallel and can be made to match the underlying heterogeneous platform. In this article, we provide a comprehensive survey for parallel programming models for heterogeneous many-core architectures and review the compiling techniques of improving programmability and portability. We examine various software optimization techniques for minimizing the communicating overhead between heterogeneous computing devices. We provide a road map for a wide variety of different research areas. We conclude with a discussion on open issues in the area and potential research directions. This article provides both an accessible introduction to the fast-moving area of heterogeneous programming and a detailed bibliography of its main achievements.

show abstract

Section: Cuda-to-openclmentioning

confidence: 99%

Parallel programming models for heterogeneous many-cores: a comprehensive survey

et al. 2020

View full text Add to dashboard Cite

show abstract

“…A most intuitive way of the code translation is to let programmers manually translate the code of each platform by substituting similar API functions. There are also many efforts to achieve automatic translations, as shown in [22][23][24]. Researches show that the automatically translated code also performs well, at least for the case of converting CUDA codes to OpenCL codes.…”

Section: Translations Between Cuda and Openclmentioning

confidence: 99%

Large-Scale Data Computing Performance Comparisons on SYCL Heterogeneous Parallel Processing Layer Implementations

2020

View full text Add to dashboard Cite

Today, many big data applications require massively parallel tasks to compute complicated mathematical operations. To perform parallel tasks, platforms like CUDA (Compute Unified Device Architecture) and OpenCL (Open Computing Language) are widely used and developed to enhance the throughput of massively parallel tasks. There is also a need for high-level abstractions and platform-independence over those massively parallel computing platforms. Recently, Khronos group announced SYCL (C++ Single-source Heterogeneous Programming for OpenCL), a new cross-platform abstraction layer, to provide an efficient way for single-source heterogeneous computing, with C++-template-level abstractions. However, since there has been no official implementation of SYCL, we currently have several different implementations from various vendors. In this paper, we analyse the characteristics of those SYCL implementations. We also show performance measures of those SYCL implementations, especially for well-known massively parallel tasks. We show that each implementation has its own strength in computing different types of mathematical operations, along with different sizes of data. Our analysis is available for fundamental measurements of the abstract-level cost-effective use of massively parallel computations, especially for big-data applications.

show abstract

“…The CU2CL tool [12] is an automated CUDA-to-OpenCL source-tosource translator (CU2CL), enabling portability of CUDA software to platforms lacking a CUDA implementation. We chose OpenCL as the target language for the following reasons: (1) its broad-based adoption and support, e.g., it currently supports not only GPUs but also CPUs, Intel Xeon Phi, and even FPGAs, (2) a similar programming model to CUDA, and (3) a "write-once, run-anywhere" open standard.…”

Section: Cu2cl Translatormentioning

confidence: 99%

“…• A unified runtime layer implemented on top of CUDA and one or more other models that support other devices • Automatic source-level translation of kernel and/or host code • Automatic conversion of PTX IR to another device's IR • Runtime compatibility or wrapper layers between CUDA and one or more other models Here we leveraged CU2CL [12], a compiler-based CUDA-to-OpenCL auto-translator for both kernel and host code, that uses a small amount of on-the-fly generated runtime compatability code to construct functionally-equivalent analogs to CUDA functions. Closely related is the work of Kim et al [16], which provides hand-written bi-directional runtime wrapper layers between CUDA and OpenCL host code and a kernel translator; this work is distinguished from CU2CL by providing an OpenCL-to-CUDA compatibility layer but not aiming to provide a static translation of the host runtime API and providing a compatibility layer instead.…”

Section: Related Workmentioning

confidence: 99%

“…Additionally, non-GPU accelerator devices such as the Intel Xeon Phi and FPGAs have captured a significant market and performance share as viable alternative devices. In this paper, we explore the potential to achieve functional and performance portability across these devices by leveraging the CU2CL translator by Gardner et al [12] to translate and evaluate three scientific applications, as noted earlier: (1) lid-driven cavity (LDC), a computational fluid dynamics (CFD) application, (2) Fen Zi, a molecular dynamics application, and (3) GEM, a molecular modeling application.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the Portability of CPU-Accelerated Applications via Automated Source-to-Source Translation

Sathre

Gardner

Feng

2019

Proceedings of the International Conference on High Performance Computing in Asia-Pacific Region

Self Cite

View full text Add to dashboard Cite

Over the past decade, accelerator-based supercomputers have grown from 0% to 42% performance share on the TOP500. Ideally, GPUaccelerated code on such systems should be "write once, run anywhere, " regardless of the GPU device (or for that matter, any parallel device, e.g., CPU or FPGA). In practice, however, portability can be significantly more limited due to the sheer volume of code implemented in non-portable languages. For example, the tremendous success of CUDA, as evidenced by the vast cornucopia of CUDAaccelerated applications, makes it infeasible to manually rewrite all these applications to achieve portability. Consequently, we achieve portability by using our automated CUDA-to-OpenCL source-tosource translator called CU2CL. To demonstrate the state of the practice, we use CU2CL to automatically translate three medium-tolarge, CUDA-optimized codes to OpenCL, thus enabling the codes to run on other GPU-accelerated systems (as well as CPU-or FPGAbased systems). These automatically translated codes deliver performance portability, including as much as threefold performance improvement, on a GPU device not supported by CUDA.

show abstract

Characterizing the challenges and evaluating the efficacy of a CUDA-to-OpenCL translator

Cited by 8 publications

References 13 publications

Parallel programming models for heterogeneous many-cores: a comprehensive survey

Parallel programming models for heterogeneous many-cores: a comprehensive survey

Large-Scale Data Computing Performance Comparisons on SYCL Heterogeneous Parallel Processing Layer Implementations

On the Portability of CPU-Accelerated Applications via Automated Source-to-Source Translation

Contact Info

Product

Resources

About