Assessing Opportunities of SYCL for Biological Sequence Alignment on GPU-based Systems

TL;DR

This paper tackles the portability problem of CUDA-based bioinformatics kernels by migrating SW# to SYCL via Intel's oneAPI and SYCLomatic. It reports a complete migration with about 95% automatic translation and modest hand-tuning, achieving functional, cross-architecture, and cross-SYCL-implementation portability across NVIDIA, AMD, and Intel platforms. Performance remains comparable between CUDA and SYCL across diverse workloads, with small variances (often within 2–10%), and portability is demonstrated even when changing CPU/GPU vendors and SYCL backends. The work highlights SYCL as a viable path for sustaining CUDA-based legacy codes in Bioinformatics, and outlines future improvements including optimization for SWIPE-like techniques, FPGA deployment, and broader performance-portability analyses.

Abstract

Bioinformatics and Computational Biology are two fields that have been exploiting GPUs for more than two decades, 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. In addition, 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. Moreover, performance remained stable when switching to another SYCL implementation. As a consequence, SYCL and its implementations can offer attractive opportunities for the Bioinformatics community, especially considering the vast existence of CUDA-based legacy codes.

Figures (9)

• Figure 1: Distribution of the warnings generated by SYCLomatic.
• Figure 2: Performance comparison when varying work-group size.
• Figure 3: Performance comparison when varying protein databases.
• Figure 4: Performance comparison when varying the query length.
• Figure 5: Performance comparison when varying the alignment algorithm.