search

Martinsos / edlib

Lightweight, super fast C/C++ (& Python) library for sequence alignment using edit (Levenshtein) distance.
http://martinsos.github.io/edlib
MIT License
506 stars 165 forks source link

Adding some level of parallelization to Edlib #113

Open Martinsos opened 6 years ago

Martinsos commented 6 years ago

Edlib was made not made with special care for parallelization and does not parallelize over threads or anything. The thing is, I am pretty sure that in practice there is often a need to run many of edit distance calculations in parallel as fast as possible, and in that case you want to completely utilize the computing power you have available: all cores. So the idea is to learn more about these use cases and then to offer support for this through Edlib. I am not sure what this exactly means, as I am not very well acquainted with this level of paralelization, so the next step would be to learn more about the use cases and possibilities, and then to decide on the next steps.

Martinsos commented 6 years ago

118 is somewhat connected to this.

bpanda-dev commented 2 years ago

Hey @Martinsos! I am interested in parallelization of Edlib on GPUs supporting OpenCL/CUDA framework. I would like to extend this work onto the GPUs. Please let me know if you are okay with it. It would be specifically useful for read mapping pipeline wherein we have a large number of Candidate reference segments and we need filtering techniques.