On Mon, Mar 16, 2026 at 6:41 PM Pádraig Brady <[email protected]> wrote:
> Newer hardware for example won't have to worry about the JCC issue,
> so that might be left as a distro level / user level setting.

Right. One of data sources to support decisions of that kind might be
https://data.firefox.com/dashboard/hardware

They have a nice per-microarch percentage. However, the Firefox user
base (mostly desktop) probably differs from coreutils (all sorts of
machines).


> align-loops is fairly generic, but also may increase size.
> Ideally we could [...] keep this as focused as possible.

I think, size increase is capped with 124 bytes with four loops in
four hot functions, unless I miss something. That's exactly why I
place lookup functions in a separate translation module.

I'll add --*able-cdc-align-loops and cdc-jcc-erratum, it's the right
thing to do, indeed. Default values might be tuned later.


> It would be worth discussing this with compiler folks though,
> as they'd be very interested in significant wins like that

I agree. I think it's the right thing to do when the code lands in the
trunk. The code might serve as yet another baseline then to test
performance regressions against.


--
WBRBW, Leonid Evdokimov, https://darkk.net.ru tel:+79816800702
PGP: 6691 DE6B 4CCD C1C1 76A0  0D4A E1F2 A980 7F50 FAB2

Reply via email to