Migrating CUDA to oneAPI: A Smith-Waterman Case Study

Costanzo, Manuel; Rucci, Enzo; Sánchez, Carlos García; Naiouf, Marcelo; Prieto-Matías, Manuel

doi:10.1007/978-3-031-07802-6_9

Cited by 8 publications

(6 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Last, we discuss the obtained results considering related works. 11 It is important to note that this change implied a significant reduction in the number of lines of code 12 Fortunately, several are available from an increasing number of vendors https://www.khronos.org/sycl/ 13 https://intel.github.io/llvm-docs/GetStartedGuide.html 14 Swiss-Prot: https://www.uniprot.org/downloads 15 ENV NR: https://ftp.ncbi.nlm.nih.gov/blast/db/…”

Section: Resultsmentioning

confidence: 99%

“…On the other hand, DPC++-specific functions were replaced by pure SYCL calls to query the device information. As a result, this final version of the code can be compiled with any of the SYCL-compatible compilers 12 .…”

Section: Sycl Standardization (Optional)mentioning

confidence: 99%

“…In particular, we have selected SW# [11]: a CUDA-based, memory-efficient implementation for biological sequence alignment, which can be used either as a stand-alone application or a library. This paper is an extended and thoroughly revised version of [12]. The work has been extended by providing:…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Assessing Opportunities of SYCL and Intel oneAPI for Biological Sequence Alignment

Costanzo¹,

Rucci²,

García-Sánchez

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Bioinformatics and Computational Biology are two fields that have been exploiting GPUs for more than two decades, with being CUDA the most used programming language for them. However, as CUDA is an NVIDIA proprietary language, it implies a strong portability restriction to a wide range of heterogeneous architectures, like AMD or Intel GPUs. To face this issue, the Khronos Group has recently proposed the SYCL standard, which is an open, royalty-free, cross-platform abstraction layer, that enables the programming of a heterogeneous system to be written using standard, single-source C++ code. Over the past few years, several implementations of this SYCL standard have emerged, being oneAPI the one from Intel. This paper presents the migration process of the SW# suite, a biological sequence alignment tool developed in CUDA, to SYCL using Intel’s oneAPI ecosystem. The experimental results show that SW# was completely migrated with a small programmer intervention in terms of hand-coding. Moreover, it was possible to port the migrated code between different architectures (considering multiple vendor GPUs and also CPUs), with no noticeable performance degradation on 5 different NVIDIA GPUs. As a consequence, SYCL and Intel oneAPI can offer attractive opportunities for the Bioinformatics community, especially considering the vast existence of CUDA-based legacy codes.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Sycl Standardization (Optional)mentioning

confidence: 99%

See 1 more Smart Citation

Assessing Opportunities of SYCL and Intel oneAPI for Biological Sequence Alignment

Costanzo¹,

Rucci²,

García-Sánchez

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the last decades, academia and companies have been working on developing a unified language to program heterogeneous hardware, capable of improving productivity and portability. One of the most recent and promising proposals is the SYCL standard from the Khronos Group 3 . SYCL is an open, royalty-free, crossplatform standard, that enables the programming of a heterogeneous system to be written using standard, single-source C++ code.…”

Section: Introductionmentioning

confidence: 99%

“…In particular, we have selected SW# [15]: a CUDA-based, memory-efficient implementation for biological sequence alignment, which can be used either as a stand-alone application or a library. This paper is an extended and thoroughly revised version of [3]. The work has been extended by providing:…”

Section: Introductionmentioning

confidence: 99%