Re: [boost] [hash2][review] Early review (due to holidays)

Yes, this is the challenge with optimisations, caching, etc, is: in many cases you have to make a trade-off between size/performance. I think it is also important to mention that if you ultimately want to tag a block as unsafe buffers usage and just do pointer arithmetics on top of the memory you received through a span, you are free to do that. That's your prerogative. That's an aspect of the abstract machine you now have a much finer control over because you spanified your interfaces. The other way around is not true though. You can opt-in to spanify your interfaces and retain the ability to deep dive in a lower level type of code, and use pointers as you see fit, However, you don't benefit from higher abstractions once all you have is just offer pointer/length pairs. I think Peter has made this case in a much more educated manner that I have. Claudio. On Fri, Dec 6, 2024 at 11:06 PM Ivan Matek via Boost wrote:

On Fri, Dec 6, 2024 at 7:19 PM Vinnie Falco wrote:

...

...

How?

maybe we are not talking about same situation, but this is what I meant,

godbolt https://godbolt.org/z/K8xEjEoMK link [image: image.png] Compiler will generate "specialized" functions for each different size of span with non dynamic extent. Here you can see how he implemented summation for 3 and 4 integers in different ways.

This is great for performance and makes checking easier since as Peter explained compiler knows more, but it creates larger binaries(generally speaking, I know compilers are smart, can inline, for 2 instantiations does not really matter, etc). Function taking dynamic span or runtime specified n will probably be slower because it does normal loop, but there is only one copy of it in resulting assembly.

There is a real life example of this here in fmt. note link gives certificate error and I did not manage to find another link https://vitaut.net/posts/2020/reducing-library-size/

_______________________________________________ Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost