TBO I had to read up on what the io_uring even is. Sounds pretty exciting. So do you think it's gonna be like what swiss tables are to hashmaps?
-
-
-
This is way bigger. It's going to make things that need to use disk 3x faster. It addresses one of the biggest complaints about the linux kernel for my entire career: friction around real async file IO. It lets you dramatically reduce / eliminate syscalls for file & network IO.
- 3 more replies
New conversation -
-
-
I have to try it, there's definitely some fun code to write
-
I am not exaggerating when I say it is going to change everything. Already for storage it's so easy to hit line speeds. I'm about to get linux 5.5 setup and post numbers about networking accept rates vs the other popular approaches, but I fully expect it to be a bloodbath.
- 6 more replies
New conversation -
-
-
Very interesting. What are the short pros/cons of this vs IOU?
-
iou is basically just a low-level unsafe binding to the C library. rio is pure rust, no need to compile and link against C, and it aims to prevent a lot of misuse by leveraging rust's lifetimes and RAII to make use-after-free of the buffers you feed io_uring impossible
- 5 more replies
New conversation -
-
-
Is it possible to use with mio?
-
Completion implements Future, so you can use it with any async runtime. mio is just a thing that smacks epoll/kqueue/iocp as rio smacks io_uring. you can use rio easily without an async runtime as well, using wait() instead of .await. I donno if mio 0.7 can still be used alone
- 2 more replies
New conversation -
-
-
what is up with the last line of that README? I assume/hope it's a forgotten leftover joke, but maybe not something to publicise officially in its present format,
@rustlang -
narc repellent working as intended
- 4 more replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.