I'm passionate about writing small, close to OS utilities for myself and writing high performance C++ programs. Particularly, extracting almost demoscene level of performance from computers seemingly lacking said performance.
My Ph.D. thesis puts this passion to practical use, by developing a scientific application which uses every ooze of performance it can get from a system.
The end result is "You shouldn't be able to compute with this accuracy at this speed. What have you done?", which is extremely satisfying for me.
The utilities are for solving my daily problems, mostly automation. Tools reaching a certain maturity are open sourced with copyleft licenses at source hut: https://git.sr.ht/~bayindirh/
The other tool is a material simulation code, for scientific applications. It will be open sourced, again, when it reaches maturity. It’s single machine, multithreaded C++, but we may move to multi-machine if necessary in the future.
My Ph.D. thesis puts this passion to practical use, by developing a scientific application which uses every ooze of performance it can get from a system.
The end result is "You shouldn't be able to compute with this accuracy at this speed. What have you done?", which is extremely satisfying for me.
Also, I'm learning Go in these days.