Buffering and optimization

programming
operatingsystem
linux
cprogramming
#1

One of the best article on buffering (buffer cache) / optimization I ever have read.

How is GNU yes so fast?

#2

Hello there,

Thank you for sharing it! I’d like to request from now on not to post URL shorteners because this is an InfoSec community and some folks might be intimidated/skeptical about clicking it.

I hope you understand. Take care and enjoy our community!

#3

Hello. The link is now fixed.

(xchg rbp, rsp) #4

Very good article indeed. I stumbled across this this morning. I found it quite interesting.

yes source code comparison

DCMqZt5UAAEJ0uz.jpg large.jpg2012x2048 389 KB

#5

Great work @sagittarius-a

just for info, GNU coreutils is still the fastest because it works on buffer cache directly.

I have an old ibook G3 (late 2001, 256MB RAM, 750fx PPC cpu) with OpenBSD -current on it.

The result is: 3.08MiB/s, really really slow.

#6

but OpenBSD does have something cool, pledge(). :wink:

#7

You’re right @kowalski, It’s a new risk mitigation feature for OpenBSD, however pledge() was born for security purposes, it is not for optimization.

It works really well :slight_smile:

#8

This article is such interesting, if you have got others like this please post them :wink: