if we store the output of running tools to memory, we can consume a lot of it. Storing it into files and using those files will have more advantages:
1) memory consumption
2) if we do not know the benchmark, we can just move the file to the directory with unknown benchmarks
3) we can run more threads - one that will run tools and the other that will process these files (of course, that would be possible with the in-memory objects too, but this way it is less bug-prone, since it is more separated)
if we store the output of running tools to memory, we can consume a lot of it. Storing it into files and using those files will have more advantages: 1) memory consumption 2) if we do not know the benchmark, we can just move the file to the directory with unknown benchmarks 3) we can run more threads - one that will run tools and the other that will process these files (of course, that would be possible with the in-memory objects too, but this way it is less bug-prone, since it is more separated)