Skip to content

GPU_Smith_Waterman.md

Latest commit

 

History

History
47 lines (33 loc) · 1.24 KB

GPU_Smith_Waterman.md

File metadata and controls

47 lines (33 loc) · 1.24 KB

GPU-Accelerated Smith-Waterman

This module provides an optional GPU implementation of the Smith–Waterman local alignment algorithm using PyOpenCL. When a compatible GPU or OpenCL runtime is unavailable, the code falls back to a pure Python implementation.

Usage

python -m src.python.gpu_smith_waterman SEQ1 SEQ2

To call the module from Java, the CLI flag --sw-align triggers the Python implementation when used alongside --align.

# Analyse sample.fa and align against reference.fa
java -jar dnanalyzer.jar sample.fa --align reference.fa --sw-align

# Or specify both query and reference explicitly
java -jar dnanalyzer.jar --align query.fa reference.fa --sw-align

Or import the class in your own scripts:

from src.python.gpu_smith_waterman import SmithWatermanGPU
sw = SmithWatermanGPU()
score, matrix = sw.align("ACACACTA", "AGCACACA")

The score variable will contain the maximum alignment score and matrix contains the dynamic programming matrix.

Dependencies

  • Python 3.8+
  • pyopencl (optional – falls back to CPU if not installed)

Install dependencies via:

pip install pyopencl

GPU execution has been tested on CUDA and OpenCL compatible devices.