[prev in list] [next in list] [prev in thread] [next in thread]
List: ubuntu-devel
Subject: Re: Proposal: Let's drop i386
From: Lyn Perrine <walterorlin () gmail ! com>
Date: 2018-05-14 4:43:52
Message-ID: CAC5bvTy2Wi_e7SNc5KMvgg_QzbwV7KVq7Urv+fipeUDTK0n-JQ () mail ! gmail ! com
[Download RAW message or body]
[Attachment #2 (multipart/alternative)]
The other question is does anyone test ubuntu on non SSE2 hardware anymore?
On Sun, May 13, 2018 at 11:48 AM, Henri Sivonen <hsivonen@hsivonen.fi>
wrote:
> On Sun, May 13, 2018 at 9:34 PM, Matthias Klose <doko@ubuntu.com> wrote:
> > On 13.05.2018 05:00, Henri Sivonen wrote:
> >> On Thu, May 10, 2018 at 11:25 PM, Thomas Ward <teward@thomas-ward.net>
> wrote:
> >>> However, killing i386 support globally could introduce issues,
> including
> >>> but not limited to certain upstream softwares having to go away
> >>> entirely, due to the interdependency or issues with how certain apps
> >>> work (read; Wine, 32-bit support, 64-bit support being flaky, and
> >>> Windows apps being general pains in that they work on 32bit but not
> >>> always on 64-bit).
> >>
> >> If 32-bit x86 support becomes mainly a thing that's run on x86_64
> >> hardware as a compatibility measure for things like Wine, it would
> >> make sense to bring the instruction set baseline to the x86_64 level.
> >> Specifically, it would make sense to compile the 32-bit x86 packages
> >> with SSE2 unconditionally enabled.
> >>
> >> This would mean dropping support for Pentium Pro and earlier or Athlon
> >> XP and earlier, but it's pretty sad to leave all that performance on
> >> the table in order to support the few computers still in use that have
> >> Pentium Pro or earlier or Athlon XP or earlier.
> >>
> >> As upstream software assumes SSE2 as the baseline, it will be less and
> >> less a run-time check and compiling software without SSE2 will mean
> >> shipping it in a damaged form performance-wise.
> >
> > I disagree, until you provide data how many packages fail to build, at
> least in
> > the testsuites, when run without the extra x87 precision bits.
>
> I don't have this data, but considering that SSE2 is a mandatory part
> of x86_64, it seems implausible that packages would be
> SSE2-intolerant. Considering that x86_64 defaults to SSE2
> floating-point math (or does Ubuntu override this?) and considering
> that ARM doesn't have x87 available, it seems implausible that
> packages would rely on x87. (On the contrary, since e.g. Firefox and
> Chromium upstreams don't do non-SSE2 x86 builds anymore, it seems more
> plausible that there exist packages whose upstream doesn't test x87
> floating-point math anymore.)
>
> --
> Henri Sivonen
> hsivonen@hsivonen.fi
> https://hsivonen.fi/
>
> --
> Ubuntu-devel-discuss mailing list
> Ubuntu-devel-discuss@lists.ubuntu.com
> Modify settings or unsubscribe at: https://lists.ubuntu.com/
> mailman/listinfo/ubuntu-devel-discuss
>
[Attachment #5 (text/html)]
<div dir="ltr">The other question is does anyone test ubuntu on non SSE2 hardware \
anymore? <br></div><div class="gmail_extra"><br><div class="gmail_quote">On Sun, May \
13, 2018 at 11:48 AM, Henri Sivonen <span dir="ltr"><<a \
href="mailto:hsivonen@hsivonen.fi" \
target="_blank">hsivonen@hsivonen.fi</a>></span> wrote:<br><blockquote \
class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc \
solid;padding-left:1ex"><span class="">On Sun, May 13, 2018 at 9:34 PM, Matthias \
Klose <<a href="mailto:doko@ubuntu.com">doko@ubuntu.com</a>> wrote:<br> > On \
13.05.2018 05:00, Henri Sivonen wrote:<br> >> On Thu, May 10, 2018 at 11:25 PM, \
Thomas Ward <<a href="mailto:teward@thomas-ward.net">teward@thomas-ward.net</a>> \
wrote:<br> >>> However, killing i386 support globally could introduce \
issues, including<br> >>> but not limited to certain upstream softwares \
having to go away<br> >>> entirely, due to the interdependency or issues \
with how certain apps<br> >>> work (read; Wine, 32-bit support, 64-bit \
support being flaky, and<br> >>> Windows apps being general pains in that \
they work on 32bit but not<br> >>> always on 64-bit).<br>
>><br>
>> If 32-bit x86 support becomes mainly a thing that's run on x86_64<br>
>> hardware as a compatibility measure for things like Wine, it would<br>
>> make sense to bring the instruction set baseline to the x86_64 level.<br>
>> Specifically, it would make sense to compile the 32-bit x86 packages<br>
>> with SSE2 unconditionally enabled.<br>
>><br>
>> This would mean dropping support for Pentium Pro and earlier or Athlon<br>
>> XP and earlier, but it's pretty sad to leave all that performance on<br>
>> the table in order to support the few computers still in use that have<br>
>> Pentium Pro or earlier or Athlon XP or earlier.<br>
>><br>
>> As upstream software assumes SSE2 as the baseline, it will be less and<br>
>> less a run-time check and compiling software without SSE2 will mean<br>
>> shipping it in a damaged form performance-wise.<br>
><br>
> I disagree, until you provide data how many packages fail to build, at least \
in<br> > the testsuites, when run without the extra x87 precision bits.<br>
<br>
</span>I don't have this data, but considering that SSE2 is a mandatory part<br>
of x86_64, it seems implausible that packages would be<br>
SSE2-intolerant. Considering that x86_64 defaults to SSE2<br>
floating-point math (or does Ubuntu override this?) and considering<br>
that ARM doesn't have x87 available, it seems implausible that<br>
packages would rely on x87. (On the contrary, since e.g. Firefox and<br>
Chromium upstreams don't do non-SSE2 x86 builds anymore, it seems more<br>
plausible that there exist packages whose upstream doesn't test x87<br>
floating-point math anymore.)<br>
<span class="im HOEnZb"><br>
-- <br>
Henri Sivonen<br>
<a href="mailto:hsivonen@hsivonen.fi">hsivonen@hsivonen.fi</a><br>
<a href="https://hsivonen.fi/" rel="noreferrer" \
target="_blank">https://hsivonen.fi/</a><br> <br>
</span><div class="HOEnZb"><div class="h5">-- <br>
Ubuntu-devel-discuss mailing list<br>
<a href="mailto:Ubuntu-devel-discuss@lists.ubuntu.com">Ubuntu-devel-discuss@lists.<wbr>ubuntu.com</a><br>
Modify settings or unsubscribe at: <a \
href="https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel-discuss" \
rel="noreferrer" target="_blank">https://lists.ubuntu.com/<wbr>mailman/listinfo/ubuntu-devel-<wbr>discuss</a><br>
</div></div></blockquote></div><br></div>
[Attachment #6 (text/plain)]
--
ubuntu-devel mailing list
ubuntu-devel@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic