f-lawless

Fear not the machine, nor view it abstract,
for yours is the hand that directs every act

[f-lawless @izflawless]

In the back-end, its just nbfc-linux

whirr is only a TUI for more intuition and interaction

[f-lawless @izflawless]

worthy of getting added to your Linux ricing or nah?

[f-lawless @izflawless]

repo is https://t.co/Apyy1cM8uF

[f-lawless @izflawless]

Yes, I know the standard library already ships this idea: std::pmr::monotonic_buffer_resource (C++17, <memory_resource>). Same bump-pointer approach, and it plugs into pmr::vector and pmr::string.

I wanted to build the bare version first to see where the time goes.

[f-lawless @izflawless]

Why can't I just use std::vector of C++ as an arena allocator and call it a day?

I'd never tried it, so I built one and benchmarked it against naive new/delete. Here's what perf says

[f-lawless @izflawless]

This is also the best case for an arena: allocate lots, free all at once. malloc is doing far more general work (any size, any order, thread safety).

Great for parsers, per-frame game data, and request-scoped memory. Not a malloc replacement.

[f-lawless @izflawless]

@brsy_10 For doing my 1+1=2 operations