Regression in Zig/C comparison

[vegerot]

The website mentions this comparison between C and Zig.

In more recent versions, while Zig still uses faster instructions, it actually uses one more instruction which makes the comparison look less cool.

Of course it's just doing some stack management which is way faster than an imul, but just thought it was worth pointing out :)

https://godbolt.org/z/qWGvEo9o6

[alexrp]

What page is this on?

[vegerot]

This used to be on the overview, but looks like it was moved to old

www.ziglang.org/old-content/learn/overview.en.md

Line 60 in eb96ef0

- Carefully chosen undefined behavior. For example, in Zig both signed and unsigned integers have undefined behavior on overflow, contrasted to only signed integers in C. This [facilitates optimizations that are not available in C](https://godbolt.org/z/n_nLEU).

[bronze1man]

https://godbolt.org/z/fab14rvM9
zig 0.13.0 only have three lines:

half:
        mov     eax, edi
        shr     eax
        ret