In this repository is hosted a Python and C++ implementation of a simple procedure to find the complement of a gene nucleotide sequence in biochemistry.
In both the python and cpp directories is an optimised implementation (in optimised) and a less optimised implementation (in original), the interactive component of this presentation will require the audience to profile the implementation they prefer.
First clone the repository and move to the directory created:
git clone https://github.com/Hitham2496/profiling-presentation.git
cd profiling-presentation
To profile the C++ code, we use Valgrind, move to the original implementation in cpp/original and follow the instructions in the slides.
You will first need to compile the code with make.
You may then move to the cpp/optimised directory and repeat the profiling process.
Take time to look at the code as well and understand what could be causing performance issues, some of the comments may be helpful
Similarly we use the in-built cProfile and pstats packages to profile the Python implementations.
Follow the instructions on the slides in both the python/original and python/optimised directories and visualise the output with e.g. snakeviz.
To install snakeviz you may use pip e.g.
pip install snakeviz (--user)
Yannick posed an interesting bonus puzzle, why does C/C++ process sorted arrays so much faster than unsorted arrays?
For details, see Yannick's snippet.
Solution in computing club next week!