Optimized FPGA Implementation for Accelerating Local Sequence Alignment

Software Implementation of LSAL

SW directory contains all the software implementations of the lsal algorithm:

• lsal.c -> base lsal algorithm
• lsal_opt.c -> first optimization (optimal algorithm)
• lsal_opt_pad.c -> second optimization (second optimal algorithm, simpler code)

In the same directory there is also a run_all.sh file that you can use to compile, execute and create graphs for different values of (N,M).

In order to get the graphs for the different optimizations execute the bash script:

bash ./run_all.sh

The script will create graphs with comparisons of all three implementation for different values of N,M. You can modify the python script in order to add or remove N,M values.

Hardware Implementation of LSAL

Under hw directory there are two hardware implementation of the lsal algorithm:

• lsal_base: the source code of the base algorithm with no optimizations (containing both the kernel and host code for executing on an FPGA)
• lsal_optimized: the source code of the optimized algorithm (kernel+host code)

Comparison of LSAL on x86 vs FPGA

The FPGA implementation demonstrates a significant performance improvement, achieving approximately a 5x speedup compared to the x86 implementation.