[Development] Using SSE/NEON in Qt 6

[Lisandro Damián Nicanor Pérez Meyer]

Hi!

On Thu, 6 Feb 2020 at 14:14, Kevin Kofler <kevin.kofler at chello.at> wrote:
>
> Lisandro Damián Nicanor Pérez Meyer wrote:
> > At least in Debian we do this for qtbase on i386, and have different
> > versions of corelib and gui (the only ones which where directly affected
> > by this).
>
> FYI, QtWebEngine is pretty much unfixably screwed on non-SSE2. The last
> version that was reasonably workable (still with heavy patching, mind you)
> was 5.9:
> https://src.fedoraproject.org/rpms/qt5-qtwebengine/blob/f25/f/qtwebengine-opensource-src-5.9.1-no-sse2.patch
>
> After that, Chromium dropped the V8 x87 backend, and since V8 has no
> interpreter fallback, the SSE2 one is the only option left. I was able to
> forward-port the x87 backend to 5.10, with minimal, emulator-only testing
> (WARNING: huge patch):
> https://src.fedoraproject.org/rpms/qt5-qtwebengine/blob/f27/f/qtwebengine-everywhere-src-5.10.1-no-sse2.patch
>
> Forward-porting that patch to current Chromium (anything ≥ QtWebEngine 5.11)
> is a lot of work, you basically have to track all the changes to the
> x86/SSE2 V8 backend and port them to the x87 backend included in the patch.
> I stopped working on that when it became too much work for me and when
> Fedora decided to drop support for non-SSE2 x86 anyway (though it officially
> was only for Fedora ≥ 29, but QtWebEngine already stopped supporting it in
> Fedora 28 updates).
>
> Blame Google for that.

Thanks, and yes, we are aware of that. But people tend to understand
that for huge beasts like web engines, not for the whole Qt stuff.
Anyways I think SSE2 should be now a good default. But that's my
personal opinion.


-- 
Lisandro Damián Nicanor Pérez Meyer
http://perezmeyer.com.ar/
http://perezmeyer.blogspot.com/