Intel Further Speeds Up strnlen() In The GNU C Library For Recent Intel/AMD CPUs

[phoronix]

Phoronix: Intel Further Speeds Up strnlen() In The GNU C Library For Recent Intel/AMD CPUs

Intel software engineers are responsible for many of the great x86_64-related optimizations to the GNU C Library "glibc" over the years. While they've extensively tuned many Glibc functions for achieving peak performance on their modern CPUs, it's a never-ending quest. Merged this week was another optimization to strnlen(), the function for determining the number of bytes in a fixed-size string...

https://www.phoronix.com/news/Intel-EVEX-Faster-strnlen

[patrick1946]

They should make it really slow. Null terminated strings are one of the worst C decisions to save some bytes in the '70.

[elbar]

There are systems and corporations where glibc is forbidden. And C string where character zero is sign of the end of string and is also a mark of well designed system according C lang designer where compilation of 10-100 MLOC app(100k-500k is typical scholar project ) took at least half a day so it is also a wormsign

[uxmkt]

They should make it really slow. Null terminated strings are one of the worst C decisions to save some bytes in the '70.

WP: """Null-terminated strings were produced by the .ASCIZ directive of the PDP-11 assembly languages and the ASCIZ directive of the MACRO-10 macro assembly language for the PDP-10. These predate the development of the C programming language [..]"""

[Volta]

They should make it really slow. Null terminated strings are one of the worst C decisions to save some bytes in the '70.

It doesn't have to be NULL terminated when you're using strnlen(), so that's probably why it can be optimized.

[Volta]

There are systems and corporations where glibc is forbidden. And C string where character zero is sign of the end of string and is also a mark of well designed system according C lang designer where compilation of 10-100 MLOC app(100k-500k is typical scholar project ) took at least half a day so it is also a wormsign

What?

[Netch]

I'd really like to understand where this optimization is useful enough. strnlen(), while being a vitally safe in a context when you aren't certain in input, is "critical path" for too few applications. Who asked glibc developers, or this is their own play?

[patrick1946]

WP: """Null-terminated strings were produced by the .ASCIZ directive of the PDP-11 assembly languages and the ASCIZ directive of the MACRO-10 macro assembly language for the PDP-10. These predate the development of the C programming language [..]"""

Yes, but so was string with a size to this time. So they chose it to save maybe one to 3 bytes. AFAIK to this time strings had commonly a 2byte size.

[patrick1946]

It doesn't have to be NULL terminated when you're using strnlen(), so that's probably why it can be optimized.

Yes, if it can get it for a string literal, it can optimize it. Still, null-terminated strings are an anti-pattern.

[markg85]

They should make it really slow. Null terminated strings are one of the worst C decisions to save some bytes in the '70.

Go ahead and do that on your computer.
Write a custom library that has the same glibc signature for the functions you don't like. Make it like a million times slower for your own pleasure.
LD_PRELOAD that library system wide for everything you run.
And enjoy running - or rather snail pace crawling - your pc.