parBuffer takes longer than just using a single core.

[noahmartinwilliams]

I made this repo to demonstrate how the parallel library works, and it takes longer to use parallelism than it does to use a single core. It doesn't seem to run all threads at the same time very often according to threadscope, and I'm pretty sure it's a problem with the parallel library.

Edit: It seems if I split the password list into chunks then it runs faster with parallelism, but it still doesn't run using all cores very efficiently. This might just be a problem with my code.

[konsumlamm]

This is likely because the work the sparks (the threads that do the parallel evaluation) are doing is too small, so the overhead of sparking computations is greater than the gain through parallelism. Splitting the list into chunks increases the work a single spark is doing and reduces the total sparking overhead.