[cudaaligner] banded myers with fixed bandwidth and low memory requirements

NVIDIA-Genomics-Research / GenomeWorks

SDK for GPU accelerated genome assembly and analysis

https://clara-parabricks.github.io/GenomeWorks/

Apache License 2.0

286 stars 76 forks source link

[cudaaligner] banded myers with fixed bandwidth and low memory requirements #513

Closed ahehn-nv closed 4 years ago

ahehn-nv commented 4 years ago

Banded Myers takes an upper bound for the bandwidth now
Banded Myers has a dynamic memory structure now - it doesn't allocate buffers of a fixed size for an alignment (based on the max_query_length and max_target_length given at construction) anymore.
Added a new API to create a cudaaligner with a fixed bandwidth
batched_device_matrices can accommodate matrices of different sizes now

ahehn-nv commented 4 years ago

Functionality to distinguish optimal/approximate alignments and add the number of edit of the alignments to the alignment object will follow in a separate PR once this is merged.